5,036 671 4MB
Pages 537 Page size 612 x 782.9 pts Year 2011
Handbook of Traffic Psychology
This page intentionally left blank
Handbook of Traffic Psychology
Bryan E. Porter Old Dominion University Norfolk VA, USA
AMSTERDAM l BOSTON l HEIDELBERG l LONDON l NEW YORK l OXFORD l PARIS SAN DIEGO l SAN FRANCISCO l SINGAPORE l SYDNEY l TOKYO Academic Press is an imprint of Elsevier
Academic Press is an imprint of Elsevier 32 Jamestown Road, London NW1 7BY, UK 225 Wyman Street, Waltham, MA 02451, USA 525 B Street, Suite 1800, San Diego, CA 92101-4495, USA First edition 2011 Copyright Ó 2011 Elsevier Inc. All rights reserved with the exception of Chapter 32 which is in the Public Domain No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without the prior written permission of the publisher Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in Oxford, UK: phone (+ 44) (0) 1865 843830; fax (+44) (0) 1865 853333; email: [email protected]. Alternatively, visit the Science and Technology Books website at www.elsevierdirect.com/rights for further information Notice No responsibility is assumed by the publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging-in-Publication Data A catalog record for this book is available from the Library of Congress ISBN : 978-0-12-381984-0 For information on all Academic Press publications visit our website at www.elsevierdirect.com Typeset by TNQ Books and Journals Printed and bound in United States of America 11 12 13 14
10 9 8 7 6 5 4 3 2 1
Contents
Preface List of Contributors
vii ix
9. Neuroscience and Young Drivers
Part I Theories, Concepts, and Methods 1. How Many E’s in Road Safety?
3 13
27
Martha Hı´jar, Ricardo Pe´rez-Nu´n˜ez and Cristina Incla´n-Valadez
4. Self-Report Instruments and Methods
127
Maria T. Schultheis and Kevin J. Manning
11. Visual Attention While Driving
137
12. Social, Personality, and Affective Constructs in Driving
149
Dwight Hennessy
13. Mental Health and Driving 43
Timo Lajunen and Tu¨rker O¨zkan
5. Naturalistic Observational Field Techniques for Traffic Psychology Research
10. Neuroscience and Older Drivers
David Crundall and Geoffrey Underwood
Ray Fuller
3. CaseeControl Studies in Traffic Psychology
109
A. Ian Glendon
John A. Groeger
2. Driver Control Theory
Part II Key Variables to Understand in Traffic Psychology
165
Joanne E. Taylor
14. Person and Environment: Traffic Culture
179
Tu¨rker O¨zkan and Timo Lajunen
61
David W. Eby
15. Human Factors and Ergonomics
193
Ilit Oppenheim and David Shinar
6. Naturalistic Driving Studies and Data Coding and Analysis Techniques
73
Sheila G. Klauer, Miguel Perez and Julie McClafferty
7. Driving Simulators as Research Tools in Traffic Psychology
16. Factors Influencing Safety Belt Use 215 87
Oliver Carsten and A. Hamish Jamson
8. Crash Data Sets and Analysis Young-Jun Kweon
Part III Key Problem Behaviors Jonathon M. Vivoda and David W. Eby
17. Alcohol-Impaired Driving 97
231
Krystall Dunaway, Kelli England Will and Cynthia Shier Sabo v
vi
Contents
18. Speed(ing)
249
Part V Major Countermeasures to Reduce Risk
267
29. Driver Education and Training
Thomas D. Berry, Kristie L. Johnson and Bryan E. Porter
19. Running Traffic Controls Richard Retting
20. Driver Distraction
275
Michael A. Regan and Charlene Hallett
21. Driver Fatigue
403
Esko Keskinen and Kati Hernetkoski
30. Persuasion and Motivational Messaging
423
David S. Anderson
287
Jennifer F. May
31. Enforcement
Part IV Vulnerable and Problem Road Users 22. Young Children and “Tweens” 23. Young Drivers
Part VI Interdisciplinary Issues
315
32. The Intersection of Road Traffic Safety and Public Health
Patty Huang and Flaura Koplin Winston
24. Older Drivers
457
David A. Sleet, Ann M. Dellinger and Rebecca B. Naumann
339
Barbara Freund and Paula Smith
25. Pedestrians
Bryan E. Porter
301
Kelli England Will
441
33. Public Policy 353
471
Rune Elvik
Ron Van Houten
26. Bicyclists
367
34. Travel Mode Choice
Ian Walker
27. Motorcyclists
375
David J. Houston
485
Stephen G. Stradling
35. Road Use Behavior in Sub-Saharan Africa 503 Karl Peltzer
28. Professional Drivers Tova Rosenbloom
389 Index
519
Preface
In compiling the Handbook, I had a vision to place into one work the latest research findings and future questions to be pursued in the field. I wanted the work to reach multiple audiences, including advanced undergraduates learning about applications and methods, graduate students needing the latest reviews and suggestions for research questions, and scholars in the field who benefit from one resource representing the field at-large for ease of reference and background. The final result, I believe, completes the true meaning of “handbook”da “how to” resource to know, and do work in, the field. It can even be adopted as a textbook for courses in traffic psychology. The book’s chapters are organized into six main sections: (1) Theories, Concepts, and Methods; (2) Key Variables to Understand in Traffic Psychology; (3) Key Problem Behaviors; (4) Vulnerable and Problem Road Users; (5) Major Countermeasures to Reduce Risk; and (6) Interdisciplinary Issues. Each chapter is a stand-alone resource for readers who want to start with a particular issue or topic. The chapters within each section also have different purposes and, at times, will attract different audiences whose needs vary depending on experience in traffic psychology. The material within is global, coming as it does from contributors representing 12 countries on five continents. There is also a breadth of interdisciplinary perspective, with experts from psychology, engineering, medicine, political science, and public health. The first section, Theories, Concepts, and Methods, gives readers an overview of traffic psychology as a field (Groeger), theoretical contributions (Fuller), and “how to” chapters to practice common methods. Caseecontrols (Hı´jar, Pe´rez-Nun˜ez, and Incla´n-Valadez), self-report ¨ zkan), direct observation (Eby), in-vehicle (Lajunen and O instrumentation (Klauer, Perez, and McCafferty), simulation (Carsten and Jamson), and crash data set methods (Kweon) are discussed. New students in traffic psychology, or experienced scholars wishing to consider different methods, will particularly benefit. In the second section, Key Variables, a wide range of variables are explored that providedliterallydthe “set” of those thought to be among the most important to understand. Authors explore neuroscience contributions to driving (Glendon for young drivers; Schultheis and
Manning for older drivers), which are becoming very important to the field and its future potential. Visual search patterns (Crundall and Underwood), social, personality, and affect (Hennessy), and mental health impacts (Taylor) are explored. Finally, the person, environment, and culture ¨ zkan and Lajunen) and human factors (Oppenheim and (O Shinar) impacts are reviewed. In these chapters, readers can review the latest information and research questions from within a person through to that person’s interactions with the larger social system. The third section will be very popular with readers interested in particular behaviors. Here, chapters provide what is the latest knowndand unknowndabout major problem behaviors leading to crashes, injuries, and fatalities. These behaviors are critical for traffic safety at-large, not just traffic psychology. These are safety-restraint use (Vivoda and Eby), impaired driving (Dunaway, Will, and Sabo), speeding (Berry, Johnson, and Porter), running traffic controls (Retting), distracted driving (Regan and Hallett), and fatigued driving (May). Vulnerable road users are the focus of the fourth section. Traffic psychology and related fields have a significant interest in reducing harm to subgroups of people who are disproportionately harmed on the roadways or who need particular protections that they cannot provide themselves. The field also focuses on those subgroups that disproportionally create roadway problems. This section’s chapters review young children and “tweens” (Will), young drivers (Huang and Winston), older drivers (Freund and Smith), pedestrians (Van Houten), bicyclists (Walker), motorcyclists (Houston), and professional drivers (Rosenbloom). Traffic psychologists and their colleagues are often called upon to assist in the development and evaluation of countermeasures to reduce roadway risks. The fifth section reviews major countermeasures that have received the most attention to date. Specifically, driver education and training (Keskinen and Hernetkoski), persuasion and motivational messaging (Anderson), and enforcement (Porter) are discussed. Readers in the field, or those practicing in general transportation sciences and policy, will find these chapters useful in their discussions about what questions and countermeasures may or may not be appropriate to address their needs.
vii
viii
Finally, the sixth section provides interdisciplinary perspectives. Readers will find how traffic psychology intersects with public health (Sleet, Dellinger, and Naumann) and public policy (Elvik). Environmental protection by reducing personal vehicle use in favor of public transport or other mode choices has a growing research base (Stradling). Also, traffic psychology’s role to assist worldwide injury prevention, with Africa as an important and critical example, is outlined (Peltzer). Given the ambitious nature of the work, I thank my family, Debbie, Amanda, and Sadie, and my students, whose patience and support I much appreciate. Old Dominion University’s support has also been substantial to my work on the Handbook, including a semester’s research leave to help organize the project. I thank my publisher at Elsevier, Nikki Levy, for her support of this work, and Barbara Makinster, who was my development editor. Finally, I thank my
Preface
colleagues who kindly offered advice on early drafts of the handbook material: David W. Eby, Ian Glendon, Raphael Huguenin, Geoffrey Underwood, and Kelli England Will. I am delighted to share the Handbookdfinally after so much planning and executiondwith readers interested in traffic psychology. I am excited to share how my field can make important contributions to reducing crashes, injuries, and fatalities on our roadways. I am honored to provide a forum for my colleagues to share their tremendous experience with those wanting to know who we are as a discipline. I am also proud to provide this resource to the field to celebrate its accomplishments. On behalf of the Handbook’s authors, I hope you both enjoy the book and find it useful to your own pursuits in our exciting discipline. Bryan E. Porter Old Dominion University
List of Contributors
David S. Anderson Center for the Advancement of Public Health, College of Education and Human Development, George Mason University, Fairfax, VA, USA Thomas D. Berry Christopher Newport University, Newport News, VA, USA Oliver Carsten Institute for University of Leeds, Leeds, UK David Crundall UK
Transport
Studies,
University of Nottingham, Nottingham,
David J. Houston Department of Political Science, University of Tennessee, Knoxville, TN, USA Patty Huang Center for Injury Research and Prevention and Division of Child Development and Rehabilitation Medicine, The Children’s Hospital of Philadelphia, Philadelphia, PA, USA Cristina Incla´n-Valadez Department of Geography and Environment, London School of Economics and Political Science, London, UK
Ann M. Dellinger Centers for Disease Control and Prevention, Atlanta, GA, USA
A. Hamish Jamson Institute for Transport Studies, University of Leeds, Leeds, UK
Krystall Dunaway Department of Pediatrics, Eastern Virginia Medical School, Norfolk, VA, USA
Kristie L. Johnson VA, USA
David W. Eby Michigan Center for Advancing Safe Transportation throughout the Lifespan and University of Michigan Transportation Research Institute, Ann Arbor, MI, USA
Esko Keskinen Department of Behavioral Sciences and Philosophy, University of Turku, Turku, Finland
Rune Elvik Norway
Institute of Transport Economics, Oslo,
Young-Jun Kweon Virginia Department of Transportation, Charlottesville, VA, USA
Barbara Freund Health Sciences Division, Pasadena City College, Pasadena, CA, USA
Timo Lajunen Department of Psychology, Middle East Technical University, Ankara, Turkey
Ray Fuller School of Psychology, Trinity College Dublin, Dublin, Ireland
Kevin J. Manning Department of Psychology, Drexel University, Philadelphia, PA, USA
A. Ian Glendon School of Psychology, University, Gold Coast, Queensland, Australia
Griffith
Jennifer F. May Department of Psychology, Old Dominion University, Norfolk, VA, USA
John A. Groeger School of Applied Psychology, University College Cork, Cork, Ireland
Julie McClafferty Center for Automotive Safety Research, Virginia Tech Transportation Institute, Blacksburg, VA, USA
Charlene Hallett French Institute of Science and Technology for Transport, Development and Networks, Lyon, France Dwight Hennessy Department of Psychology, Buffalo State College, Buffalo, NY, USA
Old Dominion University, Norfolk,
Sheila G. Klauer Center for Automotive Safety Research, Virginia Tech Transportation Institute, Blacksburg, VA, USA
Rebecca B. Naumann Centers for Disease Control and Prevention, Atlanta, GA, USA
Kati Hernetkoski Department of Behavioral Sciences and Philosophy, University of Turku, Turku, Finland
Ilit Oppenheim Ben-Gurion University of the Negev, Beer Sheva, Israel ¨ zkan Department of Psychology, Middle East Tu¨rker O Technical University, Ankara, Turkey
Martha Hı´jar Center of Research in Population Health, National Institute of Public Health, Cuernavaca, Morelos, Mexico
Karl Peltzer Human Sciences Research Council, Pretoria, South Africa, and University of the Free State, Bloemfontein, South Africa
ix
x
List of Contributors
Miguel Perez Center for Automotive Safety Research, Virginia Tech Transportation Institute, Blacksburg, VA, USA Ricardo Pe´rez-Nu´n˜ez Center for Health Systems Research, National Institute of Public Health, Cuernavaca, Morelos, Mexico
David A. Sleet Centers for Disease Control and Prevention, Atlanta, GA, USA Paula Smith Health Sciences Division, Pasadena City College, Pasadena, CA, USA Stephen G. Stradling Transport Research Institute, Edinburgh Napier University, Edinburgh, UK
Bryan E. Porter Department of Psychology, Old Dominion University, Norfolk, VA, USA
Joanne E. Taylor School of Psychology, Massey University, Palmerston North, New Zealand
Michael A. Regan French Institute of Science and Technology for Transport, Development and Networks, Lyon, France
Geoffrey Underwood Nottingham, UK
Richard Retting Sam Schwartz Engineering, Arlington, VA, USA
University
of
Nottingham,
Ron Van Houten Department of Psychology, Western Michigan University, Kalamazoo, MI, USA
Tova Rosenbloom Phoenix Road Safety Studies and Department of Management, Bar-Ilan University, Ramat Gan, Israel
Jonathon M. Vivoda Michigan Center for Advancing Safe Transportation throughout the Lifespan and University of Michigan Transportation Research Institute, Ann Arbor, MI, USA
Cynthia Shier Sabo Department of Biostatistics, Virginia Commonwealth University, Richmond, VA, USA
Ian Walker Department of Psychology, University of Bath, Bath, UK
Maria T. Schultheis Department of Psychology, Drexel University, Philadelphia, PA, USA David Shinar Sheva, Israel
Ben-Gurion University of the Negev, Beer
Kelli England Will Department of Pediatrics, Eastern Virginia Medical School, Norfolk, VA, USA Flaura Koplin Winston Center for Injury Research and Prevention, The Children’s Hospital of Philadelphia and University of Pennsylvania, Philadelphia, PA, USA
Full author biographies available online on ScienceDirectÒ , www.sciencedirect.com
Part I
Theories, Concepts, and Methods 1. How Many E’s in Road Safety? 2. Driver Control Theory: From Task Difficulty Homeostasis to Risk Allostasis 3. Case–Control Studies in Traffic Psychology 4. Self-Report Instruments and Methods 5. Naturalistic Observational Field Techniques for Traffic Psychology Research
3 13 27 43 61
6. Naturalistic Driving Studies and Data Coding and Analysis Techniques 7. Driving Simulators as Research Tools in Traffic Psychology 8. Crash Data Sets and Analysis
73 87 97
This page intentionally left blank
Chapter 1
How Many E’s in Road Safety? John A. Groeger University College Cork, Cork, Ireland
1. INTRODUCTION I was introduced to driver behavior research by Ivan Brown, with whom I went to work at the Medical Research Council’s Applied Psychology Unit in 1985. Ivan’s knowledge of the field was voluminous, and his proselytizing on behalf of a psychological dimension to road safety was both tireless and remarkably successful in shaping decades of research in the area in the United Kingdom. If Ivan shaped the UK agenda, Talib Rothengatter (d. 2009), with whom I first began to collaborate in 1987 as part of the remarkably foresighted European Union-funded GIDS project (Michon, 1993), gave form and substance to the behavioral aspects of traffic psychology throughout Europe and beyond. Both would have written this overview chapter far better than I can hope to do. Although I was, and remain, more interested in the cognitive underpinnings of complex skilled activity, road safety was much more central to Ivan’s concerns. He was the first person I encountered who invoked the “three E’s” mantra of road safety. It is only recently that I found reference to what is, I believe, the original coining of the phrase “education, enforcement, engineering.” According to Damon (1958), Julien H. Harvey, who was then director of the Kansas City Safety Council, gave a presentation in Topeka in 1923 during which he presented a drawing of a triangle with sides labeled “Education,” “Enforcement,” and “Engineering.” Since then, the three E’s have dominated perspectives on road safety, with occasional forays into the literature by safety experts advocating increasing the number of E’s in road safety. I, too, am going to travel this path in an attempt to overview what I consider some of the most important contributions to the literature in recent years.
2. EDUCATION One of the virtues of the three E’s is the succinct summary they offer of what remain the primary parameters of safety. However, in each case, drawing the remit of each “E” narrowly limits not only the scope but also the extent of the Handbook of Traffic Psychology. DOI: 10.1016/B978-0-12-381984-0.10001-3 Copyright Ó 2011 Elsevier Inc. All rights reserved.
potential to contribute to safety. This is demonstrably so with respect to “education.” Education has come to mean the transmission of an established body of knowledge and skills to those who lack these. In the road safety context, it has less to do with the developing of individual potential, implicated in wider use of the term “education,” and typically refers to “driver education” and “public education.” Driver education is a term used more widely in North America to cover the preparation of intending drivers for independent driving. It comprises, depending on the jurisdiction, classroom or electronic dissemination of the declarative knowledge base on which driving relies, as well as what is typically referred to as “driver training” (i.e., practical instruction on the operations the driver is required to perform when driving, including the rules that pertain to vehicle operation (Lonero, 2008)). Despite the evident face validity of driver education, the evidence of a direct safety benefit from driver education is scant and equivocal, as a succession of reviews during the past few decades have shown (Brown, Groeger, & Biehl, 1987; Christie, 2001; Ker et al., 2005; Mayhew & Simpson, 2002; Roberts & Kwan, 2001). Evidence with regard to the effectiveness of the skill and declarative knowledge components of driver education is, to some extent, more compelling. For example, there is very good evidence that the driving performance of drivers improves as they gain behind-the-wheel experience with professional driving instructors or accompanying adults (Groeger, 2000; Groeger & Clegg, 2007; Hall & West, 1996). However, there is surprisingly little evidence that the classroom or individual education leads to an increase in knowledge about, and attitudes toward, driving. One study showed that those who were pseudo-randomly assigned to classroom or individual CD-ROM- or Internet-supported study performed similarly on a post-course test of drivingrelated knowledge (Masten & Chapman, 2004). Unfortunately, the study did not include a pre-course assessment of driving knowledge, and thus the comparability of groups before undertaking courses and the relative improvement in knowledge of driving by virtue of course participation are 3
4
unclear. This suggests that the classroom setting per se does not lead to better outcomes than home study, although the educational value overall is difficult to ascertain. Some studies, which are considered later in relation to exposure, are more encouraging with regard to the contribution of driver education to safety. Mass media campaigns are also a means by which education might make a contribution to road safety. In discussing their effectiveness, I separate campaigns that seek to change behavior by emphasizing that the unwanted behavior is antisocial, or where there are safety-related consequences of some unwanted behavior, from campaigns that implicate enforcement. Two related meta-analyses of the effects of carefully conducted, substantial, wellcontrolled media campaigns on alcohol-related accidents (e.g., single-vehicle nighttime crashes) or blood alcohol content levels reveal impressively large reductions in alcohol-involved driving of approximating 13% (Elder et al., 2004; Tay, 2005a). Although impressive, the fact that no more than approximately a dozen studies, worldwide, over several decades met the rigorous standards for inclusion in these meta-analyses is very revealing of the dearth of peer-reviewed studies that demonstrate convincing reductions on relevant outcome measures. Differences between the effectiveness of campaigns against speeding or drunk driving (Tay, 2005b) both show the inherent complexity of evaluating public education campaigns and emphasize the very important point that even carefully constructed and targeted campaigns may not be equally effective as a means of reducing all unsafe/ illegal behaviors, regardless of what these are. Tay’s study also serves to emphasize the importance of message content, in that different types of unsafe/illegal behaviors may not equally support “response efficacy” (i.e., provide useful and effective avoidance strategies). The importance of this and other aspects of message content, delivery, pretesting, as well as audience effects and target offenses, has been more formally investigated in a number of other studies. These experimental studies typically use behavioral intentions, rather than measured change in specified actual behaviors, as outcome measures, but they have allowed investigation of the subtle interplay between the threat implied in campaign messages and consequent fear induced and the likely acceptance or rejection of the message among various groups (Cauberghe, De Pelsmacker, Janssens, & Dens, 2009; Lewis, Watson, & Tay, 2007; Lewis, Watson, & White, 2008, 2010). These and other studies have considerable potential to shape message content and delivery, and they provide a coherent account of how and why messages may have the potential to be effective. However, quantification of the actual safety benefits of these and other variables will be a considerable challenge, just as it has been for driver training.
PART | I
Theories, Concepts, and Methods
It would be remiss not to acknowledge a final sense in which education can make a contribution to road safety. Many who are engaged in this area have benefited from, and seek to pass on, the expertise and experience of others. As such, those of us in educational roles have the ultimate responsibility for maintaining and enhancing the knowledge base of current theory, methods, and research findings available to policymakers, other safety professionals, and society at large.
3. ENFORCEMENT Few studies demonstrate the centrality of enforcement to road safety as well as that by Tay (2005b), in which it was shown that the number of breath tests performed per month and the percentage of drivers arrested were associated with a statistically significant reduction in the number of serious crashes per month. Enforcement is likely to be by far the most important determinant of the likelihood of apprehension for a criminal act, and as such it is critical to deterrence. Thus, for example, classical deterrence theory proposed that criminal acts are less likely to be committed where the certainty of punishment is high and the punishment is both severe and swift (Taxman & Piquero, 1998). Classical deterrence theory emphasizes the importance of direct punishment of the individual. In doing so, classical deterrence theory neglects the potential to increase further offending of offenders’ experience of avoiding punishment, as well as their more vicarious experience of both punishment and punishment avoidance (Stafford & Warr, 1993). Piquero and Paternoster (1998) provide empirical support for this reconceptualization of classical deterrence theory, showing that expressed intentions to drink and drive were affected by personal and vicarious experiences as well as by both punishment and punishment avoidance. Furthermore, very strong deterrent effects were observed where the certainty of punishment for the respondent was high. Interestingly, Piquero and Paternoster also show that “moral beliefs that prohibit drunk driving are an effective source of inhibition” (p. 3), and impulsivity among individuals is associated with whether vicarious experience of punishment and punishment avoidance influences offending (Piquero & Pogarsky, 2002). Watling, Palk, Freeman, and Davey (2010) attempted to extend this analysis to “drug” rather than “drunk” driving. They showed that punishment avoidance and vicarious punishment avoidance were predictors of the propensity to drug drive in the future but note that knowing of others apprehended for drug driving was not a sufficient deterrent. It may be that particular types or patterns of drug use are more prevalent among individuals high in impulsivity, and if so, increased vicarious knowledge of punishment through publicity or media reports of detection might not be effective for specific driving-under-influence
Chapter | 1
How Many E’s in Road Safety?
offenses (in addition to the difficulty of using media appropriate for such offenders). In part because of the influence of vicarious knowledge of detection and punishment on deterrence, publicity campaigns enhance the effect of rigorous enforcement. Miller, Blewden, and Zhang (2004), during the introduction of a zero alcohol tolerance regime for drivers younger than age 20 years, investigated the effects of compulsory roadside breath testing (CBT), CBT twinned with a media campaign, and a subsequent period of greater police presence during CBT (i.e., “booze buses”). They reported a reduction in expected nighttime crashes of 22.1%, with a further reduction of 13.9% due to enhanced media, and that booze buses yielded a further 27.4% reduction where implemented. This almost halving of expected nighttime accidents persisted for several years beyond the life of the intervention. Whereas Miller and colleagues (2004) showed that increasing the apparent seriousness with which offenses are treated by enforcement agencies serves to increase deterrence, diminishing the seriousness of offenses appears to have the opposite effect and has a more general effect on safety. McCarthy (1993) reported that an increase in rural interstate speed limits significantly increased overall the incidence of alcohol-related accidents, and that alcoholrelated accidents became more prevalent in lower speed environments. Blais and Dupont (2005) emphasized the pervasiveness of safety effects resulting from strict police enforcement. Reviewing the international literature on enforcement programs focused on a broad range of offenses (including random breath testing, sobriety checkpoints, random road watch, photo radar, mixed programs, and red light cameras), they concluded that interventions resulted in an average decrease, ranging between 23 and 31%, of injury accidents. On the other hand, the consequences of lax enforcement are demonstrated in a case-crossover study of traffic law enforcement and risk of death from motor vehicle crashes (Redelmeier, Tibshirani, & Evans, 2003). These authors showed that there was a protective effect from recent convictions on individual drivers, such that the risk of a fatal crash in the month after a conviction was approximately 35% lower than that in a comparable month with no conviction for the same driver, and that this protective effect declined rapidly a few months after conviction. This protective benefit was consistent across ages, incidence of previous convictions, and other personal characteristics, and it was greater for speeding violations with penalty points than for speeding violations without points. The authors concluded that enforcement “effectively reduces the frequency of fatal motor vehicle crashes in countries with high rates of motor vehicle use. Inconsistent enforcement, therefore, may contribute to thousands of deaths each year worldwide” (p. 2177). As technologies have developed, the opportunities for enforcement other than by traditional policing have also
5
increased with automated detection of speeding offenses, red light violations, etc. In addition to the roadside reminders regarding potential violations, and the increase in implied and actual surveillance, the increased likelihood of punishment, and reduced opportunity for punishment avoidance, automated detection systems greatly enhance the potential for deterrence. There is substantial evidence that speed cameras not only reduce speeding but also reduce collisions and speed-related collisions (Pilkington & Kinra, 2005), with a meta-analysis suggesting that “injury crash reductions in the range of 20 to 25% appear to be a reasonable estimate of site-specific safety benefit from conspicuous, fixed-camera, automated speed enforcement programs” (Thomas, Srinivasan, Decina, & Staplin, 2008, p. 117). Retting, Ferguson, and Hakkert (2003) reported, on the basis of a meta-analysis of international studies of red light camera effectiveness, that injury crashes overall were substantially reduced at signalized intersections, particularly right-angle injury crashes, although the incidence of rear-end collisions increased. Other studies suggest that although violations are reduced, the overall safety benefit of red light cameras is at least questionable (Erke, 2009; Wahl et al., 2010). One of the difficulties for automated enforcement is that some who drive are not licensed to do so. A number of attempts have been made to quantify the number of unlicensed drivers in the United Kingdom during the past decade, and two very different approaches to tackling this difficult problem have yielded remarkably similar estimates. In a survey-based approach, samples of those holding provisional licenses (in the United Kingdom, such drivers must not drive without being accompanied by a qualified driver) were written to and asked, anonymously, whether or not they had driven while unlicensed and the extent of that driving (Knox, Turner, Silcock, Beuret, & Metha, 2003). The proportion of drivers admitting to driving illegally was then weighted by the number of drivers known to hold provisional licenses or who were disqualified by the courts. This approach yielded an estimate of approximately 476,300 unlicensed drivers. On March 31, 2006, the UK police randomly stopped 5793 vehicles and checked whether the drivers held current driving licenses, finding that 1.6% did not. Extrapolated to the UK driving population of approximately 31 million, this gives an estimate of approximately 480,000 unlicensed drivers. Various methodological weaknesses underlie these studies. The survey evidence is based on understandably low response rates (10e20%), despite the anonymity of responding, and gives no indication of the number of those who may be driving who have never held a license. The police random survey, by using the number of licensed drivers to estimate the total number of unlicensed drivers, obviously leads to underestimation. In addition to the problem that unlicensed drivers pose for automated
6
enforcement, there is evidence that unlicensed drivers are between three and nine times more likely to be involved in accidents that result in injury or death (Knox et al., 2003). These risk ratios for the general population are similar to those reported for a Californian sample (DeYoung, Peck, & Helander, 1997), although they are likely to be higher for particular groups of drivers (Blows et al., 2005). Just as studies of more traditional police-based enforcement demonstrate an enhancing effect of education, in the form of concurrent mass media campaigns, these studies of the effects of automated enforcement emphasize an increasing interaction between the effects of enforcement and engineering. This third element of Harvey’s three E’s safety mantra is considered next.
4. ENGINEERING Traditionally, safety benefits from engineering would have been anticipated from improvements to vehicle build quality, reliability, improved braking performance, and the protection offered to vehicle occupants. Others arise from improvements to road design, surface quality, reduced deterioration during adverse conditions (Elvik & Greibe, 2005), safer roadside furniture (Elvik, 1995), less confusing signage, etc. During the past decade, impressive additional safety benefits have accrued from the improvements to occupant protection. Studies of the safety benefits of child safety seats and booster seats for older/heavier children have shown that compared with restraint by seat belts alone, restraint by belts positioned more correctly by the concomitant use of boosters resulted in 59% fewer injuries to children aged 4e7 years who were involved in motor vehicle crashes (Durbin, Elliott, & Winston, 2003), whereas the U.S. National Highway Traffic Safety Administration (NHTSA) reported that child safety seats are 71% effective in reducing fatalities among infants and 54% effective among toddlers (NHTSA, 2009a, 2009b). Safety has also increased because of a reduction in the number of younger children traveling in the front seats of vehicles (Durbin, Elliot, Arbogast, Anderko, & Winston, 2005); children in rear seats are 50e66% less likely to suffer injury (Arbogast, Kallan, & Durbin, 2009). Air bags, although potentially problematic for children (Durbin, Kallan, et al., 2003), have been shown to reduce fatalities in frontal crashes by 14% when no seat belt is used and by 11% when used in conjunction with a seat belt (Braver, Ferguson, Greene, & Lund, 1997; NHTSA, 2009a, 2009b). Seat belt use is also associated with substantial reductions in mortality following head-on collisions (Crandall, Olson, & Sklar, 2001; Evans, 1986). However, although the reductions are robust, it should be acknowledged that both seat belt use and air bag deployment are associated with what are largely less serious specific injuries that would
PART | I
Theories, Concepts, and Methods
otherwise have been unlikely to occur (Hutt & Wallis, 2004; Smith & Hall, 2005). These enhancements to occupant protection have wellestablished safety benefits and have few, if any, direct implication for the way the driving task is carried out. In contrast, developments in in-vehicle and roadside telematics may have profound consequences for how the driving task is carried out and, in some cases, even what the driving task is. Even the most developed systems are still some way from having real market penetration, including systems that are commercially available, and the majority of the systems envisaged are barely at the stage at which live trials in real traffic are possible. Whereas the vehicle modifications presented previously have clear safety benefits, those that will arise from the advance of telematics rely largely on experts’ expectations (Kulmala, 2010). Nevertheless, the anticipated safety benefits, assuming systems are fitted in all European Union vehicles, are impressive. Those with the highest fatality reduction potential are electronic stability control (fatality reduction of approximately 17%), lane keeping support systems (~15% reduction in fatalities), and systems that warn the driver when the speed limit is exceeded and when locations with higher incidences of accidents are being approached (~13%). Other systems considered might induce speed adaptation depending on weather conditions, obstacles, or congestion; warn drivers of imminent collisions and apply brakes if necessary; advise drivers regarding accidents, obstructions, and poor visibility or road conditions in the locale; warn of upcoming red light obligations; or assist the driver at nighttime or during other poor visibility conditions by warning of obstructions beyond the range of headlights. Assessing the safety impacts of future developments is a profound methodological and theoretical challenge that relies on assumptions about how people will adapt to what might be very dramatic differences in what is required of them as “drivers,” the road environments in which they will drive, and the likelihood of other vehicles they may interact with having similar technological enhancements. Although I do not doubt that reliable relative assessments of potential safety benefits can be made, the treating of these estimates as likely to result in absolute casualty reductions is premature, at best. As I have attempted to show, Harvey’s three E’s have had, and continue to have, considerable relevance for how we conceptualize potential contributors to road safety. In my view, however, other E’s, as reviewed in the following sections, also merit consideration.
5. EXPOSURE Although it is inherent in the concept of risk that negative outcomes are weighted in some way by some index of the possible outcomes, road safety statistics typically consider
Chapter | 1
How Many E’s in Road Safety?
injury and fatality as head counts, or head counts weighted by the size of the population, numbers of vehicles, or per unit of distance driven, in order to take into account at least to some extent what the relative opportunities are for collisions to occur. This more detailed consideration of such accident data allows us to identify, for example, that young, inexperienced drivers, relative to other motorists, have more of their accidents during weekends, at night or the early morning, or the accidents are more likely to take place when young drivers are accompanied by several similarly aged passengers rather than when traveling alone or when traveling with one other. Discovering particular patterns of “proneness” can be critical to understanding why such events take place, as well as for designing potential countermeasures. Although in principle this proposition might justify including exposure as an additional E in safety, two examples may make its case particularly compelling. I have been interested in the links between genetics, cognition, and sleep for many years, but it was only when re-reading some literature on young drivers that a particulardas yet untesteddhypothesis regarding the overinvolvement of young drivers in accidents occurred to me (Groeger, 2006). A variation on a particular gene (period 3) is associated with people’s self-declared preferences for being active in mornings or evenings. When sleep deprived, people with these particular genetic variations are especially poor at performing tasks in the early morning (Groeger et al., 2008). During the school/work week, teenagers report obtaining far less sleep than they would wish (Groeger, Zijlstra, & Dijk, 2004); thus, I hypothesized that teenagers with a particular configuration of the period 3 gene may actually be more likely to have accidents in the early morning when they had not slept earlier that night (Groeger, 2006). Epidemiological evidence considered in Groeger (2006) showed that when the numbers of reported trips made by young drivers at particular times of the day are taken into account, teenage drivers are far more likely to be involved in road traffic accidents in the early hours of the morning than are drivers in their early twenties. Without the perspicacity of Sweeney, Giesbrecht, and Bose (2004), who reported both accident and trip frequencies by age and time of day, this very detailed and specific prediction would not have occurred to me. If research currently under way in our laboratory provides support for this hypothesis, we may well have a new and quite distinct way of accounting for at least some of the higher crash involvement of inexperienced drivers. The notion that there are particular patterns of exposure associated with young, inexperienced drivers’ accidents is a major part of the rationale underlying graduated driver licensing (GDL) interventions. Although there are a wide variety of GDLs in operation throughout the world, the underlying principles are the same: Higher risk activities,
7
such as driving at night, with alcohol, and with teenage passengers, are delayed until the driver is older. Evidence that GDL is associated with a substantial casualty reduction has been accumulating during approximately the past decade. Although some have suggested that the effects may arise partly through reduced exposure as a result of decreased or delayed licensing, Masten and Foss’s (2010) survival analysis demonstrated that 16-year-old drivers experienced lower first-crash incidence during the first 5 years of unsupervised driving than did those licensed under the previous system, with greater benefits for young male drivers. A meta-analysis of the effects of GDL systems goes further and helps to identify which components of GDL are associated with accident reductions (Vanlaar et al., 2009). In the learner stage, these components include the length of night restriction and driver education. In the intermediate stage, the influential components include driver education, whether night restrictions are lifted (for work purposes), passenger restrictions and whether these are lifted if passengers are family members, and, importantly, whether there is an exit test. Two of these effective components are particularly worthy of comment. First, it is noteworthy that the inclusion of a test prior to reducing restrictions is associated with greater reductions in accident risk. Because youthfulness and inexperience both contribute to overinvolvement in accidents, an intervention that reduces risk exposure solely on the basis of age, as I have speculated elsewhere, will be less effective in reducing accidents (Groeger & Banks, 2007). Second, the meta-analysis by Vanlaar and colleagues presents far more convincing evidence for the safety benefits of driver education than has hitherto been found, and it is striking that these emerge when some of the huge variability in exposure of young, inexperienced drivers is controlled. Although not featured in the studies considered previously, the role of parentally imposed restrictions on driving, through parenteteen agreements, etc., is also likely to contribute substantially to increased safety arising from graduated licensing (SimonsMorton, Ouimet, & Catalano, 2008).
6. EXAMINATION OF COMPETENCE AND FITNESS Perhaps because of the political and societal issues that testing raises, the potential road safety contribution of driver assessment has been largely ignored. Most discussions of driver education and testing readily concede that the competency standards drivers must demonstrate to gain a driver’s license are, at best, rudimentary (Baughan, Sexton, Maycock, Chin, & Quimby, 2005; Lonero, 2008; Mayhew, 2007). Increasing the requirement to demonstrate knowledge of the more theoretical aspects of driving seems unlikely to increase driving ability. In the United Kingdom,
8
for example, the pass rate for the on-road driving test changed little after the introduction of extensive theory testing (Wells, Tong, Sexton, Grayson, & Jones, 2008). Wells et al. performed a thorough analysis of the relationship between demonstrating theoretical knowledge of driving and practical driving competence, which controlled for respondent age as well as hours of tuition and practice. The researchers found that no effect of passing a hazard perception theory test was observed among male or female respondents, even when the analysis was conducted separately for those taking their first, second, third, fourth, etc. driving test. Because it does not appear to improve driving ability, theory testing seems unlikely to have a safety benefit, other than by slightly delaying licensing. Another study, also conducted in the United Kingdom, raises serious questions about the reliability of driving test outcomes. Baughan and Simpson (1999) asked candidates taking the practical on-road driving test to voluntarily undergo a second test within days of their first test. The passefail designation was the same for 64% of those re-tested, suggesting that substantial numbers of drivers or examiners were unable to perform their respective roles consistently. These findings suggest that there is considerable scope for improving the reliability and validity of the practical driving test. Repeated, rather than one-off, testing and the introduction of more objective, electronic measurement of driver performance would not only identify those drivers capable of performing more consistently but also probably serve to delay licensingdwith a consequent increase in safety. Furthermore, as previously mentioned, incorporating testing within a GDL regime has been shown to increase safety; thus, requiring that drivers are competent to cope with the demands of the circumstances in which they will drive when restrictions are reduced has much to recommend it. In my view, improving our ability to examine drivers’ competence is perhaps the least wellexplored opportunity for enhancing road safety. Arguably, our rather unsophisticated approach to assessing driver competence has made it particularly difficult to develop rigorous, reliable, and especially valid assessments of fitness to continue driving among those who, through age or infirmity, are believed to pose a greater safety risk. Epidemiological studies, based on different data sets, indicate that older drivers with heart disease or stroke are more likely to be involved in at-fault traffic accidents (McGwin, Sims, Pulley, & Roseman, 2000). Sagberg (2006), using a different methodology, showed that similar risk factors pertain in Europe, adding that nonmedicated diabetes and depression are also associated with greater crash involvement. Although many jurisdictions require that drivers with such conditions notify the licensing authority and withdraw from driving, evidence suggests that compliance with this requirement is low. McCarron, Loftus, and McCarron (2008) reported that among
PART | I
Theories, Concepts, and Methods
consecutive hospital referrals, approximately 40% of drivers who suffered a heart attack or minor stroke were driving 1 month after the event. There is thus good evidence to support the suggestion that certain medical conditions are associated with increased involvement in traffic accidents, and that voluntary compliance with advice to cease driving is not adhered to. Given the many factors that are associated with lower crash risk, such as not being young, having driving experience, reduced exposure in terms of distance driven, and the timing and circumstances of trips, simply denying those who suffer from heart disease, untreated diabetes, stroke, etc., does not seem justified or fair. Regrettably, our ability to assess the driving skills and competences associated with greater safety is, at best, limited. A variety of approaches to this admittedly difficult issue have been adopted, and the relative successes of these approaches have been reported (Hunter et al, 2009; Schultheis, DeLuca, & Chute, 2009). Identifying those who pose greater threats to safety than is acceptable, however, is only one part of this complex, and often tragic, problem. Denying people the right to independent mobility, which the motor car has conferred so lavishly on so many of us, cannot be the end of our involvement with this issue as traffic professionals. Involvement might take several forms: counseling or supporting the former driver and family members in order to enable all concerned to cope with the decision, increasing access to other forms of transport, or developing remedial programs that might offer a reasonable possibility for former drivers to return to driving. The scientific challenge of the latter is substantial due to the limited theoretical understanding and practical progress with regard to improving driver education and training and driver assessment.
7. EMERGENCY RESPONSE For decades, the widely accepted view among medical professionals was that the prognosis for a seriously injured individual was in part determined by whether the patient reached the hospital within 1 h of the trauma onset. A search for historical sources and empirical support for this notion of a “golden hour” proved fruitless (Brooke Lerner & Moscati, 2001; Newgard et al., 2010). Despite this, there is ample evidence that in the case of life-threatening injuries, delaying treatment until the patient reaches a trauma center increases the likelihood of death (Hoffman, 1976). Studies show that compared to patients treated in the field or first hospital destination, patients whose first treatment is delayed until the trauma center is reached are 3.3 times more likely to die (Gomes et al., 2010). Sampalis, Lavoie, Williams, Mulder, and Kalina (1993) reported that a total prehospital time of more than 60 min was associated with a statistically significant adjusted relative odds of
Chapter | 1
How Many E’s in Road Safety?
dying (OR ¼ 3.0). Based on an analysis of a sample of more than 1400 traffic accidents on motorways and other roads, Sa´nchez-Mangas, Garcı´a-Ferrrer, De Juan, and Arroyo (2010) reported that a 10-min reduction in the typical medical response time of approximately 1 h was associated with an approximately one-third decrease in mortality probability. The circumstances in which certain types of accidents occur, such as in isolated rural areas or at times of the day or night when the crash site is less likely to be encountered by others who might contact emergency services, can leave those involved at greater risk of death. This is particularly the case for accidents involving younger drivers. Work carried out as part of the development of telematic systems estimated that an automated accident warning system linked to emergency services would reduce motor vehicle fatalities in Finland by 5e10%. In 95% of these cases, the consequence would be injuries requiring further hospital and other treatment, and in the remaining 5% the consequence would be injuries requiring no further treatment at all (Sihvola, Luoma, Schirokoff, Karkola, & Salo, 2009). Until such systems are widely available, and even when they are, in the case of severely injured road users, the rubric of (1) getting to the patient quickly, (2) treating what can be treated on site, and (3) getting the patient to an appropriate trauma treatment center as quickly as possible will remain key to ensuring that severe injuries do not unnecessarily result in death.
8. EVALUATION The seventh, and perhaps most important, “E” in road safety is evaluation. Throughout the previous discussion of the contributions to safety by education, engineering, enforcement, exposure, emergency response, and examining competence and fitness, I illustrated a range of methods and techniques that are critical to making informed, unbiased assessments of the effectiveness of safety interventions. In doing so, I glossed over very substantial challenges of the methods that road safety researchers have used to evaluate the effectiveness of countermeasures during the past century. Because safety is the desired outcome of interventions in this area, the “gold standard” for those seeking to evaluate effectiveness is casualty reduction, particularly fatalities. Counting crashes and relating these to some “explanatory” variables is the basis of much evaluation in road safety. These explanatory variables may reflect some purposeful “treatment” (e.g., increased penalties, mandatory driver education, and seat belt use) or some change consequent on other developments in society (e.g., economic activity, migration, and fuel shortages). The statistical challenges arising from crash- or casualtyfrequency data are very lucidly described by Lord and
9
Mannering (2010). Because crashes are intrinsically rare events, Poisson distributions are generally assumed. As Lord and Mannering note, the Poisson approach requires that the mean crash frequency and variance in crash frequency will be equal, but these dispersion requirements are frequently violated, resulting in biased efficacy estimates. Such count-data methods require that counts are made with respect to some temporal or spatial context. However, information on variations in the explanatory variables of interest is rarely available at the level of detail required to adequately assess their explanatory power. Lord and Mannering use the example of attempting to model the effects of precipitation on monthly crash data, when in reality it is the distribution of precipitation in far smaller time units that is likely to influence whether or not crashes occur. Temporal and spatial contexts may also be intercorrelated, leading to inaccurate assessments of the explanatory power of other variables. Other problems addressed include the difficulties that arise from correlations between injury severity and crash type, underreporting of certain types of crashes, low mean and sample sizes, omission of other potential explanatory variables, and what Lord and Mannering refer to as “endogenous” variables, where the explanatory variable changes as a function of the dependent variable (e.g., evaluating the effectiveness of ice warning signs in preventing ice-related crashes, where ice warning signs are more likely to be placed at locations at which such accidents have occurred). These problems are intrinsic to the dependent variable in which we have most interest and on which we seek to exert a causal influence. Detecting and inferring causality is, to say the least, problematic. An insightful paper by Ezra Hauer (2010) considers “causality” in two other approaches to evaluation in road safety: cross-sectional and beforeeafter studies. He demonstrates the difficulty, perhaps even the impossibility, of inferring causality in both cases. Hauer’s case is that cross-sectional regression studies, in which crash frequencies are contrasted across sites with different types of intervention, rarely capture, and then poorly account for, what Lord and Mannering (2010) might refer to as the spatial and temporal context and also other circumstantial aspects of individual interventions. By doing so, crosssectional studies can neither negate nor corroborate each other’s findings. Beforeeafter studies, although more closely having the experimental control Hauer advocates as essential for determining causation, yield outcomes in which the efficacy of treatments depends on the particular circumstances of the treatment sites. Thus, the safety effects of replacing a stop sign with a traffic signal will depend on the amount and nature of traffic using the intersection, the number of lanes over which control is sought, the proximity of treated intersections to untreated ones, their relative conspicuity, and a host of other factors.
10
Without a very large number of treated sites, for which the characteristics of each are carefully recorded, it is impossible to adequately estimate the extent to which any effect observed depends on particular circumstances in which the treatment is realized. Hauer implies that without being able to do so, predicting the effect an intervention will have in a given circumstance is impossible (see also Hauer, 1997). Hauer (2010) makes a further point that is sometimes neglected when road safety researchers engage in evaluation: Theory in road safety is weak but indispensable. In addition to citing a discussion between the eminent statisticians William G. Cochran and Sir Ronald Fisher in which the latter argued that elaborate theories, in which as many consequences of causal hypotheses are envisioned as possible, were among “the most potent weapons in observational studies” (Cochran, 1965, p. 252), Hauer asserts that “to do applied research without providing a theory is like attempting to build the roof of a house with no foundation” (p. 1130). Hauer’s comment testifies to what is in my view the essential step we must make to move from mere description to predictiondthe building and testing of theory. Fisher’s comment regarding the required elaborateness of theory, and the need to envision as many consequences of a theory as possible, is particularly apposite for road safety. As discussed previously, there are many difficulties in using crash or casualty counts, but alternatives are available to our research generation as to no other. The increasing sophistication of onboard vehicle and roadside monitoring systems, largely developed for nonsafety purposes, affords a broader range of dependent variables for assessing performance than ever before. Such nonintrusively collected data are almost routinely available, potentially from truly representative samples of drivers, and genuinely reflective of what these drivers actually dodnot just those whose misfortune it is to crash. Shankar, Jovanis, Aguerde, and Gross (2008) offer a very helpful primer on the use of such naturalistic data in road safety settings. Theories that elaborate the link between these variables downstream to safety and upstream to models of intended and unintended driver behavior can meet the requirements of Fisher’s dictum and offer the possibility of more cost-effective, reliable, and extensible evaluation. Although I have been more critical of this final “E” than of any other, it is probably the most important contributor to future safety. Without evaluation, there would be no measurable contribution to safety of any intervention, no opportunity for researchers to test their predictions concerning how and why certain interventions may be effective, no opportunity for policymakers to implement those interventions most likely to prove effective, and no opportunity for road safety experts to optimize their implementation.
PART | I
Theories, Concepts, and Methods
9. CONCLUSION In this chapter, I considered Harvey’s three E’s and their contribution to road safety, and I discussed how they remain a very useful way of considering safety interventions, despite the fact that almost a century has passed since they were first proposed. I also identified other potential contributors to safety: exposure, emergency response, examining for competence, and evaluation. I doubt that these additional E’s will endure as long as those Harvey identified, nor would I encourage others along this acrostic path. The purpose of exploring this highly influential mnemonic was to acknowledge both the contribution its elements have made and their continuing relevance. Each element, and indeed the offspring of each element, is more complex than the original formulation acknowledges. This methodological and theoretical complexity arises largely from the increasing inter- and intradisciplinarity on which our field, and ultimately “safety,” relies.
ACKNOWLEDGMENTS This work was supported by the Science Foundation Ireland (09/RFP/ NES2520) and Ireland’s Road Safety Authority. A version of this chapter was presented at the Sultanate of Oman’s Traffic Safety Summit.
REFERENCES Arbogast, K. B., Kallan, M. J., & Durbin, D. R. (2009). Front versus rear seat injury risk for child passengers: Evaluation of newer model year vehicles. Traffic Injury Prevention, 10(3), 297e301. Baughan, C. J., Sexton, B., Maycock, G., Chin, L., & Quimby, A. (2005). Novice driver safety and the British practical driving test. (Report No. RR652). Crowthorne, UK: Transport Research Laboratory. Baughan, C. J., & Simpson, H. (1999). Consistency of driving performance at the time of the L-test, and implications for driver testing. In G. B. Grayson (Ed.), Behavioural research in road safety IX. Crowthorne, UK: Transport Research Laboratory. Blais, E., & Dupont, B. (2005). Assessing the capability of intensive police programmes to prevent severe road accidents: A systematic review. British Journal of Criminology, 45, 914e937. Blows, S., Ivers, R. Q., Connor, J., Ameratunga, S., Woodward, M., & Norton, R. (2005). Unlicensed drivers and car crash injury. Traffic Injury Prevention, 6(3), 230e234. Braver, E. R., Ferguson, S. A., Greene, M. A., & Lund, A. K. (1997). Reductions in deaths in frontal crashes among right front passengers in vehicles equipped with passenger air bags. Journal of the American Medical Association, 278(17), 1437e1439. Brooke Lerner, E., & Moscati, R. M. (2001). The golden hour: Scientific fact or medical “urban legend”? Academic Emergency Medicine, 8(7), 758e760. Brown, I. D., Groeger, J. A., & Biehl, B. (1987). Is driver training contributing enough towards road safety? In J. A. Rothengatter, & R. A. de Bruin (Eds.), Road users and traffic safety (pp. 135e156) Assen, The Netherlands: Van Gorcum.
Chapter | 1
How Many E’s in Road Safety?
Cauberghe, V., De Pelsmacker, P., Janssens, W., & Dens, N. (2009). Fear, threat and efficacy in threat appeals: Message involvement as a key mediator to message acceptance. Accident Analysis and Prevention, 41, 276e285. Christie, R. (2001). The effectiveness of driver training as a road safety measure: A review of the literature. Melbourne: Australia: Royal Automobile Club of Victoria. Cochran, W. G. (1965). The planning of observational studies of human populations. Journal of the Royal Statistical Society: Series A (General), 128, 234e266. Crandall, C. S., Olson, L. M., & Sklar, D. P. (2001). Mortality reduction with air bag and seat belt use in head-on passenger car collisions. American Journal of Epidemiology, 153(3), 219e224. Damon, N. (1958). The Action Program for Highway Safety. Annals of the American Academy of Political and Social Science, 320(1), 15e26. DeYoung, D. J., Peck, R. C., & Helander, C. J. (1997). Estimating the exposure and fatal crash rates of suspended/revoked and unlicensed drivers in California. Accident Analysis and Prevention, 29(1), 17e23. Durbin, D. R., Elliott, M., Arbogast, K. B., Anderko, R. L., & Winston, F. K. (2005). The effect of seating position on risk of injury for children in side impact collisions. Pediatrics, 115, 305e309. Durbin, D. R., Elliott, M. R., & Winston, F. K. (2003). Belt positioning booster seats and reduction in risk of injury among children in vehicle crashes. JAMA, 289, 2835e2840. Durbin, D. R., Kallan, M., Elliott, M., Cornejo, R. A., Arbogast, K. B., & Winston, F. K. (2003). Risk of injury to restrained children from passenger air bags. Traffic Injury Prevention, 4, 58e63. Elder, R. W., Shults, R. A., Sleet, D. A., et al. (2004). Effectiveness of mass media campaigns for reducing drinking and driving and alcohol-involved crashes: A systematic review. American Journal of Preventive Medicine, 27, 57e65. Elvik, R. (1995). The safety value of guardrails and crash cushions: A meta-analysis of evidence from evaluation studies. Accident Analysis and Prevention, 27(4), 523e549. Elvik, R., & Greibe, P. (2005). Road safety effects of porous asphalt: A systematic review of evaluation studies. Accident Analysis and Prevention, 37(3), 515e522. Erke, A. (2009). Red light for red-light cameras? A meta-analysis of the effects of red-light cameras on crashes. Accident Analysis and Prevention, 41(5), 897e905. Evans, L. (1986). The effectiveness of safety belts in preventing fatalities. Accident Analysis and Prevention, 18(3), 229e241. Gomes, E., Arau´jo, R., Carneiro, A., Dias, C., Costa-Pereira, A., & Lecky, F. E. (2010). The importance of pre-trauma centre treatment of life-threatening events on the mortality of patients transferred with severe trauma. Resuscitation, 81(4), 440e445. Groeger, J. A. (2000). Understanding driving: Applying cognitive psychology to a complex everyday task. Hove, UK: Psychology Press. Groeger, J. A. (2006). Youthfulness, inexperience and sleep loss: The problems young drivers face and those they pose for us. Injury Prevention, 12, i19ei24. Groeger, J. A., & Banks, A. P. (2007). Anticipating the content and circumstances of skill transfer: Unrealistic expectations of driver training and graduated licensing? Ergonomics, 50, 1250e1263. Groeger, J. A., & Clegg, B. A. (2007). Changes in the rate of driving instruction: Learning to drive and the power law of practice. Applied Cognitive Psychology, 21(9), 1229e1244.
11
Groeger, J. A., Viola, A. U., Lo, J. C. Y., Archer., S. N., von Schantz, M., & Dijk, D. J. (2008). Early morning executive functioning during sleep deprivation is compromised by a PERIOD3 polymorphism. Sleep, 31(8), 1159e1167. Groeger, J. A., Zijlstra, F. R. H., & Dijk, D. J. (2004). Sleep quantity, sleep difficulties and their perceived consequences in a representative sample of some 2000 British adults. Journal of Sleep Research, 13, 359e371. Hall, J., & West, R. (1996). Role of formal instruction and informal practice in learning to drive. Ergonomics, 39(4), 693e706. Hauer, E. (1997). Observational beforeeafter studies in road safety. Oxford: Pergamon. Hauer, E. (2010). Cause, effect and regression in road safety: A case study. Accident Analysis and Prevention, 42, 1128e1135. Hoffman, E. (1976). Mortality and morbidity following road accidents. Annals of the Royal College of Surgeons of England, 58, 233e240. Hutt, J. R. B., & Wallis, L. A. (2004). Injuries caused by airbags: A review. Trauma, 6(4), 271e278. Ker, K., Roberts, I., Collier, T., Beyer, F., Bunn, F., & Frost, C. (2005). Post-licence driver education for the prevention of road traffic crashes: A systematic review of randomised controlled trials. Accident Analysis and Prevention, 37(2), 305e313. Knox, D., Turner, B., Silcock, D., Beuret, K., & Metha, J. (2003). Research into unlicensed driving: Final report. (Road Safety Research Report No. 48). London: Department for Transport. Kulmala, R. (2010). Ex-ante assessment of the safety effects of intelligent transport system. Accident Analysis and Prevention, 42, 1359e1369. Lewis, I., Watson, B., & Tay, R. (2007). Examining the effectiveness of physical threats in road safety advertising: The role of the thirdperson effect, gender, and age. Transportation Research Part F: Traffic Psychology and Behaviour, 10, 48e60. Lewis, I., Watson, B., & White, K. M. (2008). An examination of message-relevant affect in road safety messages: Should road safety advertisements aim to make us feel good or bad? Transportation Research Part F: Traffic Psychology and Behaviour, 11, 403e417. Lewis, I., Watson, B., & White, K. M. (2010). Response efficacy: The key to minimizing rejection and maximizing acceptance of emotionbased anti-speeding messages. Accident Analysis and Prevention, 42(2), 459e467. Lonero, L. P. (2008). Trends in driver education and training. American Journal of Preventive Medicine, 35, 316e323. Lord, D., & Mannering, F. (2010). The statistical analysis of crashfrequency data: A review and assessment of methodological alternatives. Transportation Research Part A: Policy and Practice, 44(5), 291e305. Masten, S. V., & Chapman, E. A. (2004). The effectiveness of home-study driver education compared to classroom instruction: The impact on student knowledge and attitudes. Traffic Injury Prevention, 5(2), 117e121. Masten, S. V., & Foss, R. D. (2010). Long-term effect of the North Carolina graduated driver licensing system on licensed driver crash incidence: A 5-year survival analysis. Accident Analysis and Prevention, 42(6), 1647e1652. Mayhew, D. R. (2007). Driver education and graduated licensing in North America: Past, present, and future. Journal of Safety Research, 38(2), 229e235. Mayhew, D. R., & Simpson, H. M. (2002). The safety value of driver education and training. Injury Prevention, 8(2), i3ei8.
12
McCarron, M. O., Loftus, A. M., & McCarron, P. (2008). Driving after a transient ischaemic attack or minor stroke. Emergency Medicine Journal, 25(6), 358e359. McCarthy, P. S. (1993). The effect of higher rural interstate speed limits in alcohol-related accidents. Journal of Health Economics, 12(3), 281e299. McGwin, G., Jr., Sims, R. V., Pulley, L. E., & Roseman, J. M. (2000). Relations among chronic medical conditions, medications, and automobile crashes in the elderly: A population-based caseecontrol study. American Journal of Epidemiology, 152(5), 424e431. Michon, J. A. (Ed.). (1993). Generic intelligent driver support. London: Taylor & Francis. Miller, T., Blewden, M., & Zhang, J. F. (2004). Cost savings from a sustained compulsory breath testing and media campaign in New Zealand. Accident Analysis and Prevention, 36(5), 783e794. National Highway Traffic Safety Administration. (2009a). Children: Traffic safety facts. (Report No. DOT HS 810-987). Washington, DC: U.S. Department of Transportation. National Highway Traffic Safety Administration. (2009b). Occupant protection: Traffic safety facts. (Report No. DOT HS 810-991). Washington, DC: U.S. Department of Transportation. Newgard, C. D., Schmicker, R. H., Hedges, J. R., Trickett, J. P., Davis, D. P., Bulger, E. M., Aufderheide, T. P., Minei, J. P., Hata, J. S., Gubler, K. D., Brown, T. B., Yelle, J.-D., Bardarson, B., & Nichol, G. (2010). Emergency medical services intervals and survival in trauma: Assessment of the “golden hour” in a North American prospective cohort. Annals of Emergency Medicine, 55(3), 235e246. Pilkington, P., & Kinra, S. (2005). Effectiveness of speed cameras in preventing road traffic collisions and related casualties: Systematic review. British Medical Journal, 330, 331e334. Piquero, A. R., & Paternoster, R. (1998). An application of Stafford and Warr’s reconceptualization of deterrence to drinking and driving. Journal of Research in Crime and Delinquency, 35(1), 3e39. Piquero, A. R., & Pogarsky, G. (2002). Beyond Stafford and Warr’s reconceptualization of deterrence: Personal and vicarious experiences, impulsivity, and offending behavior. Journal of Research in Crime and Delinquency, 39(2), 153e186. Hunter, J., de Vries, J., Brown, Y., & Hekstra, A. (2009). Handbook of disabled driver assessment. Ljubljana, Slovenia: Institute for Rehabilitation. Redelmeier, D. A., Tibshirani, R. J., & Evans, L. (2003). Traffic-law enforcement and risk of death from motor-vehicle crashes: Casecrossover study. Lancet, 361(9376), 2177e2182. Retting, R. A., Ferguson, S. A., & Hakkert, S. A. (2003). Effects of red light cameras on violations and crashes: A review of the international literature. Traffic Injury Prevention, 4, 17e23. Roberts, I. G., & Kwan, I. (2001). School-based driver education for the prevention of traffic crashes. Cochrane Database of Systematic Reviews, 2001(4), CD003201, doi: 10.1002/14651858.CD003201. Sagberg, F. (2006). Driver health and crash involvement: A caseecontrol study. Accident Analysis and Prevention, 38(1), 28e34. Sampalis, J. S., Lavoie, A., Williams, J. I., Mulder, D. S., & Kalina, M. (1993). Impact of on-site care, prehospital time, and level of inhospital care on survival in severely injured patients. Journal of Trauma, 34, 252e261.
PART | I
Theories, Concepts, and Methods
Sa´nchez-Mangas, R., Garcı´a-Ferrrer, A., De Juan, A., & Arroyo, A. M. (2010). The probability of death in road traffic accidents. How important is a quick medical response? Accident Analysis and Prevention, 42(4), 1048e1056. Schultheis, M. T., DeLuca, J., & Chute, D. L. (Eds.). (2009). Handbook for the assessment of driving capacity. San Diego: Academic Press. Shankar, V., Jovanis, P., Aguerde, J., & Gross, F. (2008). Analysis of naturalistic driving data: Prospective view on methodological paradigms. Transportation Research Record, 2061, 1e9. Sihvola, N., Luoma, J., Schirokoff, A., Karkola, K., & Salo, J. (2009). Indepth evaluation of the effects of an automatic emergency call system on road fatalities. European Transport Research Review, 1, 99e105. Simons-Morton, B. G., Ouimet, M. C., & Catalano, R. F. (2008). Parenting and the young driver problem. American Journal of Preventive Medicine, 35(3), 294e303. Smith, J. E., & Hall, M. J. (2005). Injuries caused by seatbelts. Trauma, 7(4), 211e215. Stafford, M. C., & Warr, M. (1993). A reconceptualization of general and specific deterrence. Journal of Research in Crime and Delinquency, 30, 123e135. Sweeney, M., Giesbrecht, L., & Bose, J. (2004). Using data from the 2001 National Household Travel Survey to evaluate accident risk. Paper presented at the 30th annual International Traffic Records Forum, Nashville, TN, July 25e27, 2004. Taxman, F. S., & Piquero, A. R. (1998). On preventing drunk driving recidivism: An examination of rehabilitation and punishment approaches. Journal of Criminal Justice, 26, 129e143. Tay, R. (2005a). General and specific deterrent effects of traffic enforcement. Journal of Transport Economics and Policy, 39(2), 209e223. Tay, R. (2005b). The effectiveness of enforcement and publicity campaigns on serious crashes involving young male drivers: Are drink driving and speeding similar? Accident Analysis and Prevention, 37, 922e929. Thomas, L. J., Srinivasan, R., Decina, L. E., & Staplin, L. (2008). Safety effects of automated speed enforcement programs: Critical review of international literature. Transportation Research Record, 2078, 117e126. Vanlaar, W., Mayhew, D., Marcoux, K., Wets, G., Brijs, T., & Shope, J. (2009). An evaluation of graduated driver licensing programs in North America using a meta-analytic approach. Accident Analysis and Prevention, 41(5), 1104e1111. Wahl, G. M., Islam, T., Gardner, B., Marr, A. B., Hunt, J. P., McSwain, N. E., Baker, C. C., & Duchesne, J. (2010). Red light cameras: Do they change driver behavior and reduce accidents? Journal of Trauma: Injury, Infection and Critical Care, 68(3), 515e518. Watling, C. N., Palk, G. R., Freeman, J. E., & Davey, J. D. (2010). Applying Stafford and Warr’s reconceptualization of deterrence theory to drug driving: Can it predict those likely to offend? Accident Analysis and Prevention, 42, 452e458. Wells, P., Tong, S., Sexton, B., Grayson, G., & Jones, E. (2008). Cohort II: A study of learner and new drivers: Volume 1dMain Report. (Road Safety Research Report No. 81). London: Department for Transport.
Chapter 2
Driver Control Theory From Task Difficulty Homeostasis to Risk Allostasis Ray Fuller Trinity College Dublin, Dublin, Ireland
1. INTRODUCTION Driving may be described as a control task in an unstable environment created by the driver’s motion with respect to a defined track and stationary and moving objects. The task includes requirements for route choice and following, coordination of maneuvers in support of navigational objectives, and ongoing adjustments of steering and speed (Allen, Lumenfeld, & Alexander, 1971). Figure 2.1 shows speed adjustments by a driver on a winding country lane, sampled at 5-s intervals. A fundamental issue in understanding driver behavior is the nature of the control process that produces such variations in speed. Control theory is predicated on the assumption that driver control actions are dependent on perceptual processes that select information that is compared to some standard or standards. Drivers act to keep resulting
40 Speed
Speed in mph
35
discrepancies within acceptable limits in a negative feedback loop as the means of control in their goal-directed behavior (Figure 2.2). Ranney (1994), in his review of the evolution of models of driving behavior, makes a distinction between motivational and cognitive models and by implication includes control theory (e.g., Wilde’s risk homeostasis theory; Wilde,1982) within his motivational rubric. However, as can be inferred from the previous description, control theory encompasses both motivational (setting of standard) and cognitive (perceptual process) dimensions and is a characteristic not just of risk homeostasis theory (RHT) but also of zero-risk theory (Summala, 1986), Vaa’s (2007) “monitor model,” Summala’s (2007) “comfort zone model,” and the taskecapability interface (TCI) model (Fuller, 2000). It is with such models that we have seen the most evolution in recent years. All these models differ, however, in terms of their claims regarding what is the reference standard(s) in the control system. The principal aim of this chapter is to describe developments in how the TCI model conceptualizes these standards. It concludes by exploring whether, in the interests of theoretical parsimony, the different reference standards that have been proposed by Vaa and Summala can be assimilated into the developed TCI model.
30
2. THE TASKeCAPABILITY INTERFACE MODEL
25
20
15 1
3
5
7 9 5-sec sample
11
13
FIGURE 2.1 Variation in speed on a country lane.
Handbook of Traffic Psychology. DOI: 10.1016/B978-0-12-381984-0.10002-5 Copyright Ó 2011 Elsevier Inc. All rights reserved.
15
The TCI model is an attempt to understand what motivates driver decision making, with a particular emphasis on implications for performance safety. It starts from a recognition that driver perceptual processes and control actions both have rate limitations. Thus, the driver needs to continuously create and maintain conditions for driving within these limitations. That is, he or she must ensure that the demands of the driving task are within his or her capability (Figure 2.3). Loss of control occurs when, for 13
14
PART | I
Theories, Concepts, and Methods
FIGURE 2.2 Schematic description of a feedback loop. Source: Adapted from Carver and Scheier (1990).
Reference value Comparator
Input function (perception)
Output function (behavior)
Impact on environment
2.1. Driving Task Difficulty Lucky escape
Compensatory action by others
Capability (C)
Collision !
Loss of control
CD
Task demand (D)
Control
FIGURE 2.3 Starting point for the taskecapability interface model (2000): Limited capacity concept at the interface between driving task demand and driving task capability.
a multitude of possible reasons, drivers allow task demand to exceed their capability. It is the identification of these reasons that promises to shed light on how safety might be more reliably achieved as a concomitant outcome of our seemingly insatiable desire for greater mobility. From the perspective of the driver, the statistical probability of loss of control and collision (or road run-off) is not some potentially variable phenomenon and influence, as implied, for example, in RHT (Wilde, 1982). Once a driver begins to move his or her vehicle, the statistical probability of collision is essentially one. It is a certain outcome unless, of course, the driver continuously makes adjustments to avoid collision (or road run-off). For this reason, my original theoretical explorations of driver decision making focused on the concept of threat avoidance (Fuller, 1984a). However, that concept provides only a partial account, as has been discussed elsewhere (Fuller, 2005a; Michon, 1989).
The difficulty of the driving task is inversely related to the degree of separation between the demands of the task and the driver’s available capability. In principle, the greater that capability is, relative to task demand, the lower the difficulty of the task and vice versa. In general, the separation between demand and capability is equivalent to concepts such as spare capacity and safety margin (Figure 2.4). Where capability is more-or-less stable, changes in task demand will directly influence task difficulty. In this typical situation, task difficulty will be equivalent to workload and may in part be operationalized in terms of time-to-collision and time-to-line crossing, assuming resource demand (in terms of speed of information processing and response) to be inversely related to the time available (Wickens & Hollands, 2000). As task demand or workload increases, the margin of available capability to deal with additional demands decreases, and the driver becomes more vulnerable to the consequences of a performance error and to acute high demands such as in an emergency situation. Young, Higher Objective driver capability Task demand/ driver capability
Actual safety margin
Objective task demand
Lower
FIGURE 2.4 Driving task difficulty is inversely related to the degree of separation between driver capability and task demand.
Chapter | 2
Driver Control Theory
Mahfoud, Walker, Jenkins, and Stanton (2008) demonstrated this phenomenon in a simulator study of the effects of eating and drinking on driving. They found that although these activities increased subjective ratings of physical workload, initially there was no effect on driving performance measures. However, when a pedestrian unexpectedly walked in front of them, there was a reduced ability to avoid collision. The authors concluded that although drivers may be able to cope with eating and drinking during normal driving, it is the response to a sudden peak in demand that is affected by the additional activity.
2.2. Driving Task Demand Driving task demand has both information input and response output characteristics, corresponding to the requirement to determine the situation ahead and the requirement to maneuver the vehicle appropriately. It arises out of a number of factors, including vehicle performance and information display characteristics, route choice, physical characteristics of the environment (e.g., visibility and road surface), and the presence and behavior of other road users. From a safety perspective, one can think of task demand in terms of the difficulty of information acquisition along dimensions of discriminability and flow rate and the number of potential conflicts for space in the driver’s trajectory. One can also think of it in terms of controllability associated with vehicle handling (but see Section 2.3), road surface quality, and the time available for decision making and response (which for any given situation decreases with increases in speed).
2.3. Driver Capability Driver capability arises from the driver’s basic physiological characteristics, education, training, and experience. These provide conditional rules for action as well as a realtime mental representation or simulation of the situation that enables top-down or feed-forward control decisions (see Section 2.4.3). This capability arms the driver with strategies for information acquisition and the capability of preadaptation to anticipated changes in task demand. It is ultimately expressed in speed and directional control of the vehicle. One could also include vehicle control functions that enhance the driver’s capability, such as antilock brake systems, electronic stability control, and global positioning system support for route and lane choice. Nevertheless, it makes little difference to the basic formulation: Such vehicle control supports can also be construed as reducing task demand (as in earlier formulations of the model) rather than as increasing capability. Fastenmeier and Gstalter (2007) have used task analysis to develop a typology of both road/traffic situations and driver behavioral requirements that they call SAFE
15
(Situative Anforderungsanalyse von Fahraufgabendsituational analysis of behavioral requirements of driving tasks). A road/traffic situation is defined as “a bounded section from traffic reality that the driver experiences as a unit in time and space.” Behavioral requirements are a specification of relevant cognitive and psychomotor performances linked to successful negotiation of each situation. Fastenmeier and Gstalter (2007) support a distinction between a conscious information processing system, which is a sequential processor of limited capacity and speed and underpins reasoning and decision making, and a subconscious processor, which operates as a parallel, distributed system to perform a continuous, dynamic simulation of the environment and the individual’s position within it. This simulation provides the basis for a feed-forward control of the driver’s actions as well as a reference for detecting deviations from intended outcomes. The work of Fastenmeier and Gstalter (2007) provides important first steps in identifying at a micro and measurable level both the nature of driving task demands and the capabilities required of the driver to meet those demands. It is important to note, however, that capability is vulnerable in real time to a range of human factor variables, such as emotion and fatigue.
2.4. Task Difficulty Homeostasis The control theory concept at the center of the TCI model is the hypothesis of task difficulty homeostasis (Fuller, 2005a), the proposition that drivers continuously make real-time decisions to maintain the perceived difficulty of the driving task within certain boundaries, mainly (but not necessarily exclusively) by adjusting their speed (Figure 2.5). Thus, for example, increased task difficulty arising from snow and sleet and darkness additively reduces speed (Kilpela¨inen & Summala, 2007). High proportions of drivers state that they drive more slowly than usual when task difficulty increases, such as in fog (98%), heavy rain (96%), and on unfamiliar roads (88%) (Campbell & Stradling, 2003). Drivers also typically reduce speeds while negotiating intersections, but more so while simultaneously completing car phone tasks (Liu & Lee, 2005). They also choose to drive more slowly on a narrower version of the same road (Lewis-Evans & Charlton, 2006; Uzzell & Muckle, 2005). In Lewis-Evans and Charlton’s simulator study, ratings of difficulty and of subjective risk were higher for the narrower road, but drivers were not aware of the road feature that mediated these differences, suggesting that decision making was occurring at a preconscious level. Occasionally, speed is not the only variable that drivers can adjust in order to control the level of perceived task difficulty. For example, in a following situation, time headway may similarly be used. In a study of the effects of
16
PART | I
Theories, Concepts, and Methods
FIGURE 2.5 Task difficulty homeostasis.
Range of acceptable task difficulty
Comparator
Input function (perceived task difficulty)
Output function (speed and trajectory)
Impact on environment (task demand)
prolonged driving on truck drivers, it was found that as drivers’ ratings of drowsiness increased, so did their time headway (Fuller, 1984b). The corollary to this is that when task difficulty decreases, such as when roads are empty at nighttime, speeds increase (Broughton, 2005; Lam, 2003). Compensatory increases in speed have also been found by Larsen (1995), who measured the free speeds of drivers on different road segments in a 50 km/h zone. He observed an 11 km/h range from 49.2 to 60.2 km/h, with the highest mean speeds associated with what Larsen rated as the easiest driving conditions.
2.4.1. Calibration We need to modify the representation of the control process illustrated in Figure 2.5 to show that perceived task difficulty arises out of the interface between perceived task demand and perceived capability (Figure 2.6). Accuracy of driver perceptions is referred to as the driver’s calibration accuracy. Clearly, if drivers either underestimate task
FIGURE 2.6 Perceived task difficulty arises from perceived capability and perceived demand.
Range of acceptable task difficulty
Comparator
Output function (speed and trajectory)
Input function (perceived task difficulty)
Perceived capability
demand or overestimate their capability, the perceived level of task difficulty will be less than is objectively the case. Unfortunately, both of these conditions pertain to novice drivers in general (de Craen, 2010; Fuller, Bates, et al., 2008), and their poor calibration may explain in part the overrepresentation of this group in collision statistics. Harre´ and Sibley (2007) demonstrated that the disposition of young male drivers, in particular, to believe they are more capable than others occurs with both a traditional explicit measure of attitude and with a new implicit measure. In this latter, participants associated words indicating themselves with words indicating driving ability or driving caution more quickly than they associated words referring to other people with these same positive driving characteristics. The advantage of this implicit measure is that it avoids possible social desirability influences. Younger drivers also appear to be less well calibrated with regard to estimating the effects of distracting events on their performance or taking account of behavioral variables that may undermine capability. Horrey, Lesch, and Garabet (2008) asked younger and older drivers to complete
Perceived demand
Impact on environment (task demand)
Chapter | 2
Driver Control Theory
a handheld or hands-free cell phone task while navigating a closed test track in an instrumented vehicle. Although drivers generally (correctly) rated their performance as poorer when distracted, across all driving measures subjective estimates were not related to the magnitude of the distraction effect. Some drivers who estimated the smallest effects actually exhibited the largest, and these were typically younger males. A review of young driver crashes (Organisation for Economic Co-operation and Development (OECD), 2006) concluded that the rate of inattention-related crashes and near crashes is four times higher for 18- to 20-year-old drivers than for those older than 34 years. Poorly calibrated drivers, who overestimate capability or underestimate task demand, will typically operate with less spare capacity and visit the boundary where task demand meets capability more frequently. Evidence for this in less experienced drivers derives from a study by Patten, ¨ stlund, Nilsson, and Svenson (2006) in which Kircher, O a secondary peripheral detection task in real driving was performed. Less experienced drivers had significantly longer reaction times to the peripheral stimuli and higher miss rates (although these results may have been confounded in part by familiarity with the route and sex of participant).
2.4.2. Evidence for Task Difficulty Homeostasis The concept of task difficulty homeostasis is not exclusive to the TCI model, and Summala (2007, p. 194) has argued a similar case. With reference to time-to-line crossing, he cites evidence that on a wider road, more time is available and hence drivers allow longer glances and more time for subsidiary tasks (Wikman, Nieminen, & Summala, 2008; Wikman & Summala, 2005). Similarly, on a road with a series of bends, available spare capacity diminishes and subsidiary tasks typically drop out (Summala, 2007). Evidence for the proposition that drivers try to keep task difficulty more or less constant over the short term derives from the work of Godthelp (1988), who instructed drivers in open road conditions to correct their path only at the moment when it could still be corrected comfortably (to prevent lane boundary crossing). Godthelp found that over a wide range of speeds, time-to-line crossing at the point of decision was essentially constant. In a field study, van der Horst (2007) asked drivers to brake hard at the last moment at which they thought they could stop in front of the simulated rear end of a stationary passenger car. He similarly found that time to collision appeared to be independent of approach speed. In addition, in a simulator study of car following, Van der Hulst, Meijman, and Rothengatter (1999) found that when drivers expected decelerations in the lead vehicle, they maintained the same minimum headway irrespective of whether the lead vehicle decelerated or not,
17
implying that they were maintaining a consistent safety margin (time to collision).
2.4.3. Hysteresis and Top-Down and FeedForward Control There is evidence that under certain conditions there can be a hysteresis effect (i.e., response delay) in this homeostatic process, in which drivers’ adjustments lag behind changes in task demand. Thus, for example, Andrey, Mills, Leahy, and Suggett (2003) found that the first snowfall days of the year were especially prone to increased accidents. On the other hand, a key component of driver capability is a valid mental representation of what may happen next. It is this that enables top-down or feed-forward control decisions, where drivers’ adjustments to changes in task demand can anticipate those changes. We can represent this process in the model by including not only actual perceived demand but also perceived demand as immediately anticipated (Figure 2.7). Evidence that less experienced drivers have poorer anticipatory adjustment to changes in task demand derives from the work of Saad, Delhomme, and Van Eslande (1990), who found that younger drivers adjusted their speeds less (than older drivers) when approaching an intersection, and in that of de Craen (2010), who demonstrated that recently qualified drivers performed worse than experienced drivers in a test measuring adaptation of speed to increases in task demand. Real challenges for research and development are how to accelerate novice driver progression to this kind of anticipatory task difficulty management and how to sustain it in the face of frequent feedback that it was not actually necessary. In addition, the driver’s action creates, by and large, the future with which he or she has to deal. This is why the driver needs to know the effects of those actions in the context of the unfolding road and traffic situation ahead. Evidence of poorer knowledge of this type in less experienced drivers is clearly exemplified in their higher involvement in single-vehicle crashes.
2.4.4. Boundaries of Preferred Task Demand The lower boundary of a driver’s preferred task demand will be determined by a minimum consistent with making satisfactory progress and providing sufficient stimulus to avoid boredom and perhaps prevent a progressive decline into drowsiness and sleep. The upper level will be determined by such variables as the driver’s perceived capability, motivation to put effort into the task, and goals of the journey in question. Journey goals may, of course, have a direct influence on choice of speed; however, if the choice of speed is higher than would normally be preferred, perhaps because the driver is running late and needs to make up lost time, this will raise the level of task demand
18
PART | I
FIGURE 2.7 Representation of actual perceived demand and anticipated demand.
Range of acceptable task difficulty
Comparator
Output function (speed and trajectory)
Input function (perceived task difficulty)
Perceived capability
Theories, Concepts, and Methods
Perceived demand (actual) Perceived demand (anticipated)
Impact on environment (task demand)
and require the driver to operate with a higher level of task difficulty.
2.4.5. Task Difficulty as Risk Feeling One further point in the elaboration of the TCI model is that drivers appear to experience task difficulty in the same way as they experience feelings of risk (Fuller, McHugh, & Pender, 2008). In the study by Fuller et al., participants were asked to rate video sequences of the same segments of a roadway traveled at a wide range of speeds that were systematically varied. Ratings were recorded of task difficulty but also of feelings of risk. Estimates of the statistical risk of loss of control and collision were also obtained. In several replications, it was found that ratings of task difficulty and feelings of risk covaried very closely: The typical correlation between the two variables was on the order of r ¼ 0.97. However, such ratings were independent of estimates of statistical risk at lower levels of rated difficulty and risk feeling. Thus, risk feeling and statistical risk estimates are not the same thing, but risk feeling can behave as a surrogate for task difficulty. This finding has since been replicated by Kinnear, Stradling, and McVey (2008) and Lewis-Evans and Rothengatter (2009). The pivotal role of risk feeling in driver decision making is discussed further in Section 5. That feelings of risk should be so closely associated with perceived task difficulty should come as no surprise given that the outcome of loss of control of the task may be potentially so punishing. If we consider how task difficulty may be represented in the “comparator” element of the task difficulty homeostasis model, one possibility is that it involves a meta-cognitive process that is sensitive to the degree of deviation from subgoals of the driving task. Relevant subgoals that relate to speed choice, because they
are time critical, are the maintenance of directional control (adhesion to road), sampling and processing of required information, and enabling of required response. Thus, deviations from these subgoals, such as loss of directional control, loss of time to sample needed information, and loss of time to enable response execution, may trigger a fear or anxiety response because of the potentially punishing consequences. It is a question for future research to determine whether or not the degree of fear felt is systematically related to such measurable variables as time-to-line crossing, or time to collision, or is triggered in an all-ornothing manner (i.e., driven by possibility rather than probability; see Loewenstein, Weber, Hsee, & Welch, 2001).
2.4.6. Individual Differences in Preferred Task Demand and Difficulty Accumulating evidence reveals that drivers vary in their individual dispositions to adopt a particular level of task demand. In a study with Steve Stradling’s group at Napier University (project HUSSARdhigh unsafe speed accident reduction), we interviewed a national sample of British drivers, and in part of this we presented respondents with a picture of a single carriageway rural road and asked them about two speeds: What speed would they normally drive and what speed would put them right at the edge of their safety margin? There was wide variation in preferred speed: 81% of the sample were distributed over a range of nearly 30 mph (36e64 mph). Furthermore, 7% indicated a speed lower than 36 mph and 11% a speed higher than 64 mph. There was similarly wide variation in what speed they thought would put them right at the edge of their safety margin. A majority (61%) said that a speed less than 65 mph would do so. Twenty-two percent said a speed of
Chapter | 2
Driver Control Theory
65e74 mph, 11% said a speed between 75 and 84 mph, and 6% said a speed of 85 mph or faster would do so (Stradling et al., 2008). Despite this wide individual variation in preferred and in edge-of-safety margin speeds, there was a consistent relationship between the two speeds: Edge-of-safety margin speed represented a 14% increase over preferred speed. Furthermore, feelings of risk and stress did not vary with speed chosen: The feeling of risk was similar, whether one was a slow or a fast driver on the same segment of roadway. This suggests that despite variations in speed choice, perceived task difficulty may have been much more equivalent among drivers. Project HUSSAR also confirmed, on the basis of a 12year literature review, the national survey, and four focus groups, earlier findings by Musselwhite (2006) that there are four distinguishable groups of drivers. We have labeled them low risk threshold, high risk threshold, opportunistic, and reactive (Fuller, Bates, et al., 2008). Risk threshold in this context refers to the upper limit of task difficulty a driver will accept (i.e., the smallest separation between perceived task demand and capability). Low risk threshold drivers comply with speed limits, reduce their speed if they realize they are traveling faster than the speed limit, and are unlikely to change their driving behavior in a 30 mph (50 km/h) zone as a result of momentary influences, including if they are in a hurry. They are typically older, more experienced, and represent approximately 40% of male and female drivers. In marked contrast, high risk threshold drivers have positive attitudes to high-risk behavior and a thrill-seeking and expressive use of their car (Machin & Sankey, 2008), often as part of a youth subculture that exploits driving as a recreational activity that is functionally related to their life situation (Møller & Gregersen, 2008). They drive at higher speeds; commit more, and more extreme, speed limit violations and other forms of dangerous driving behavior; and have more convictions. Not surprisingly, they are more involved in collisions. Members of this group are typically young, inexperienced, and male, and they are poorly calibrated. They represent approximately 14% of drivers. The origins of the driving style of at least some members of this group may date back to early childhood. In a seminal paper by Vassallo et al. (2007), which was concerned in part with identifying longitudinal precursors of high-risk driving behavior, three clusters of drivers were identifiable at ages 19 and 20 years who differed reliably in their engagement with risk-related driving behaviors, such as excessive speeding, drink driving, drug driving, driving when fatigued, and not using seat belts. Members of the high-risk group, which comprised 7% of their sample of 1135 young adults, were mainly male (77%) and were found to have been involved in more speeding offenses and
19
collisions. Compared with others, they were more antisocial in behavior and choice of friends, more aggressive, more irresponsible, showed less empathy, and were more likely to engage in maladaptive coping (e.g., multisubstance use). However, particularly intriguing in their findings was that the characteristics of antisocial behavior and aggressiveness differentiated between the groups as early as ages 5e8 years and persisted throughout later childhood and adolescence. Does this finding imply that we can identify certain types of high-risk driver as soon as they are old enough to go to school? If so, what implications might this have for early intervention? Opportunistic drivers do not pursue high speed for its own sake, unlike the high risk threshold drivers. They tend to adjust their speed to the conditions rather than to the speed limit, and they will exceed the limit if they believe it is safe to do so. They exploit opportunities to get ahead. Approximately 23% of drivers can be labeled as primarily opportunistic, and they are more likely to be male than female. The latter, on the other hand, are more likely to be reactive drivers. This group is not persistently concerned with making good progress and tends to avoid unsafe high speed and dangerous overtaking. However, such drivers can be strongly influenced by their emotional state, driving faster if annoyed or angry or under time pressure. Consistent with this is the finding in a questionnaire study by Bjo¨rklund (2008) that women drivers report more irritation than men when impeded or exposed to reckless driving, and evidence presented by Lustman and Wiesenthal (2008) that female drivers report more aggression than men when feeling low levels of anger in similar scenarios. Dispositional influences on driver risk threshold, and therefore speed choice (potentially), are partly captured by the social and cognitive variables that form the core elements of the theory of planned behavior (TPB), notably intentions, attitudes, and perceived social norms. However, correlations between measures of these variables and measures of actual behavior are not particularly strong, perhaps explaining approximately 25% of the variance in ˚ berg & Walle´n Warner, 2008), the behavioral variable (A and it is perhaps self-evident that such a conceptual approach cannot provide a comprehensive model of dispositional influence and most certainly not an account of real-time speed decisions by drivers. Thus, Paris and Van den Broucke (2008) conclude in an evaluation of TPB that actual speeding behavior can only partially be predicted from TPB concepts and that the cognitive determinants of safe driving as identified by the TPB need to be complemented by other factors, including less “conscious” cognitive factors such as personal identity and habit formation, as well as external factors, such as cues to action, reinforcers, or the design of roads. (p. 179)
20
3. TASK DIFFICULTY ALLOSTASIS: TEMPORARY INFLUENCES ON RISK THRESHOLD In addition to dispositional differences, a wealth of research has demonstrated that several variables may temporarily raise a driver’s risk threshold. Such variables include feelings of anger and aggression, competitiveness, thrill-seeking to get an “adrenalin rush,” feelings of power, social influences, the pressure of being late, and to find out how fast a vehicle can go (Fuller, Bates, et al., 2008). For example, Ellwanger (2007) has shown that young drivers’ “delinquent” driving responses, such as speeding, aggressive driving, and risk taking, are strongly correlated with individuals’ ascribing their frustration to the voluntary and intentional actions of others on the road. Jamson (2008) reported that drivers drive closer to the car in front when their emotions are aroused. Similarly, King and Parker (2008) showed that relatively high levels of anger are associated with increased commission of both aggressive and highway code violations and that accident-involved drivers are more angry and hostile than accident-free drivers. This evidence of factors that may have an immediate influence on the level of task difficulty that drivers are prepared to accept implies that the hypothesis of task difficulty homeostasis is not completely satisfactory and that a more appropriate concept is that of allostasis. Whereas homeostasis is the process by which a target condition is maintained in the face of external variation in a negative feedback loop system, allostasis refers to adaptation to a more dynamic target condition and is defined as maintaining certain levels of biological conditions that vary according to an individual’s needs and circumstances (Kalat, 2008). So what we should really be discussing here is task difficulty allostasis. As an example of this allostatic variation in needs and circumstances, consider results from a study examining the conditions under which drivers of emergency service vehicles are more likely to crash (Gormley, Walsh, & Fuller, 2008; Walsh, Hannigan, & Fuller, 2008). For both ambulances and fire trucks, significantly more collisions are reported under blue light (BL) conditions (responding to an emergency situation with blue lights on and usually with accompanying siren) than under non-blue light (nBL) conditions. For every one nBL collision there were three BL collisions. This contrast is useful in the sense that fire trucks provide their own controls for a comparison of driving under time pressure in one direction and without that pressure in the other (albeit confounded by condition order). Drivers were quite open about their acceptance of an increased task demand level on the way to a serious case: You can justify driving at a certain speed when its three kids in a house, if you’re standing in front of a judge. You can’t justify that kind of driving if it’s a bin on fire. dParticipant 3
PART | I
Theories, Concepts, and Methods
In a questionnaire survey of these ambulance and fire truck drivers, they were significantly more likely to say they would overtake and take risks whenever possible when driving under blue lights. Thus, under certain conditions, the reference criterion for acceptable task difficulty can change: allostasisdnot homeostasis.
4. COMPLIANCE One further component now needs to be added to the model to make it more complete. The decision output from the process of task difficulty allostasis may be an achievable speed but which is in excess of the legal limit for the road segment in question. Hence, we need to include the driver’s disposition to transfer from choosing a speed based solely on task difficulty to a speed consistent with the legal limit (Figure 2.8). Evidence indicates that there is considerable variation in drivers’ dispositions to comply with limitsdvariation that represents both more or less stable individual differences (as discussed in Section 2.4.6; see Fuller, Bates, et al., 2008) and momentary influences on compliance (Stradling et al., 2008). With regard to individual differences, cognitive style may relate to degree of noncompliance with limits (and other forms of deviant behavior), particularly whether or not the individual tends to focus on potentially positive outcomes of choices and discount potentially harmful consequences or vice versa. Lev, Hershkovitz, and Yechiam (2008) showed that in a gambling task, traffic offenders give more weight to gains compared with losses, relative to control drivers, with the implication that when speeding their minds are more focused on the gains involved rather than the possible costs in terms of detection and penalty or loss of control. Interestingly, they were also found to be more extroverted, which would also dispose them to be more sensation-seeking in their profiles. Despite individual differences in disposition to comply, whatever its basis, violation of speed limits is a pervasive phenomenon: The OECD estimates that at any one time, approximately 50% of drivers are exceeding the speed limit (OECD/European Conference of Ministers of Transport, 2006). It is important to note that for a large proportion of drivers, this behavior does not necessarily represent some kind of willful contempt for rules and regulations but is rather an expression of their maintaining a preferred level of task difficultydadjusting task demand to the prevailing conditions as they perceive themdhence their anger at getting caught and fined and the general lack of social censure from others for minor violations. In the HUSSAR study discussed previously, all four of the focus groups supported the view that noncompliance is not necessarily unsafe and that an immediate influence
Chapter | 2
Driver Control Theory
21
Disposition to comply with speed limit Range of acceptable task difficulty Comparator
Input function (perceived task difficulty)
Perceived capability
Immediate influences on compliance
Output function (speed and trajectory) Perceived demand (actual) Perceived demand (anticipated)
Impact on environment (task demand)
FIGURE 2.8 Influences on compliance.
supporting noncompliance is that the speed limit is perceived to be too low.
5. RISK ALLOSTASIS THEORY Prevalent emotions concerning speed choice are likely to be fear and frustration, with fear associated with the upper level of difficulty tolerated (the driver’s risk threshold) and frustration the lower level, arising from deviations from driving goals that would otherwise have been positively or negatively rewarding. In relation to fear, in 1964, Taylor concluded from on-road observations of drivers’ autonomic activity that drivers adopt a level of anxiety that they wish to experience when driving and then drive so as to maintain it. Mesken, Hagenzieker, Rothengatter, and de Waard (2007) studied participants who drove an instrumented car in real road environments and gave self-reports at critical points. They found that anxiety associated with safetyrelated events was the most frequent on-road emotion (from a choice restricted to anger, nervousness, and happiness), and this was in turn associated with increased perceived risk (and heart rate). In the HUSSAR study (Stradling et al., 2008), in response to the open road scenario, feelings of risk were positively correlated with ratings of task difficulty (r ¼ 0.64) and significantly inversely related to perceived safety margin: The larger the margin, the less the feeling of risk. Most respondents (76%) agreed that if they drove any faster than normal, they would feel less in control, the task of driving would be more difficult (67%), and it would feel too risky (75%).
The upshot of this link between perceived task difficulty and risk feeling, discovered in our digital video studies mentioned previously (Fuller, McHugh, et al., 2008), is that we can now refer to the model as risk allostasis theory (RAT) (which is somewhat more pithy than the hypothesis of task difficulty allostasis subsumed within the TCI model). With this change in nomenclature, it is important to stress that we are not simply substituting allostasis for homeostasis in the theory known as risk homeostasis (Wilde, 1982). In Wilde’s model, the risk concept is operationalized in terms of feelings of risk in conjunction with statistical risk estimates (Simonet & Wilde, 1997). Estimates of statistical risk have no part to play in RAT, and this rejection of the role of statistical risk in driver decision making has also been emphasized by Vaa in his critical analysis of RHT (Vaa, 2007, pp. 214 and 266). Furthermore, in RHT, the determination of preferred risk levels (“target risk” in Wilde’s terminology) arises out of an inferred costebenefit analysis of safe and risky behavioral choices rather than the variables of perceived capability, journey goals, effort motivation, and dispositional and immediate factors identified in RAT.
5.1. The Role of Feelings in Decision Making The role of feelings in decision making has a long history, being explored, for example, in the early work on emotion of William James and Carl Lange in the nineteenth century and significantly developed as a concept in the work of Zajonc in the twentieth century (Zajonc, 1980). Nevertheless, one gets a sense that the so-called cognitive revolution has until relatively recently largely neglected this role.
22
However, a new emphasis on the importance of emotion and feeling in decision making has emerged in particular through the work of Damasio (1994, 2003), with his concept of the “somatic marker hypothesis.” Before discussing this hypothesis, it is useful to briefly consider the fundamental role of feelings in motivation. Feelings are the experiences concomitant with reward and punishment, with incentives and deterrents, with things we seek and things we avoid. They are the engines of the values we have and the goals we seek. Thus, although our decisions about how to realize our goals may principally involve cognitive operations, it is feelings that select our goals, enabling us to choose between them, and that energize or motivate our approach to them. Driving goals are no different. They are similarly feelings motivated and must involve at the same time both positive, approach-motivating feelings associated with the achievement of the mobility goal (destination, journey, or both), and negative, avoidance-motivating feelings associated with collision or road run-off. Once goals are identified, to a certain extent we can leave it up to cognitive operations to guide decisions that enable us to attain these goals. From a feeling perspective, however, there is one important difference between goals of approach and goals of avoidance. In the former, feelings (positive) intensify as the goal becomes nearer and are presumably fully experienced when the goal is reached. In the latter, feelings (negative) decrease to the extent that the avoidance goal is achieved. Thus, negative feelings are presumably rarely fully experienced when avoidance is successful (as is the case nearly all of the time when driving), thus giving the impression that feelings of risk, for example, are generally not important in decision making (Lewis-Evans & Rothengatter, 2009). As Carver and Scheier note (1981, p. 199), “self-regulation is relatively affect free as long as normal discrepancy reduction processes are uninterrupted and are proceeding without difficulty [italics added].” However, “when discrepancies cannot be easily reduced, then affective processes become important” (pp. 360e361). Despite this assertion, it is most important to stress that the stimuli that trigger avoidance responses must retain their emotive characteristic; otherwise, they will become neutral stimulidmere shadows that are powerless to elicit an avoidance response. The driver still has risk feelings associated with objects to be avoided: These are what determine avoidance goals. However, when operating with a safe margin from objects to be avoided, those feelings are not intensified and may not enter conscious awareness. It is this condition that is captured so well by Summala’s zero-risk hypothesis, with risk feelings only kicking in when the safety margin has shrunk to some critical level. Summala’s (2007) model argues that action is continuously monitored by a subjective risk/fear monitor, but this only
PART | I
Theories, Concepts, and Methods
plays a role in decisions when some threshold has been reached. However, Summala’s view is surely untenable: Risk feelings must continuously play a part to enable the driver to maintain safety margins, even if they have the characteristic of “whispers of affect,” as Slovic, Finucane, Peters, and MacGregor (2002) described such responses. Without continuously taking account of the emotions triggered by elements in the road and traffic environment and discrepancies between current and goal states, the driver would have no basis for decision making for his or her choices. (To test this phenomenon directly when driving, keep your eyes closed and note the rapid onset of risk feeling as you proceed. Note that this test is not recommended from a safety perspective.) Interestingly, in a study for the Irish Road Safety Authority, we obtained evidence that suggests that younger drivers are less disposed to think (and presumably therefore feel) immediately of the severest consequences of extremely dangerous behavior (Gormley & Fuller, 2008). In an interview survey of 1039 male drivers attending the World Rally Championships in Ireland in 2007, we presented the following crash scenario and asked respondents to list as many consequences as occurred to them. Participants were distributed approximately equally across the four age groups of 17e19, 20e22, 23e25, and 26e28 years: I am now going to describe to you a crash and when I finish I would like you to tell me what you think the consequences might be: “John, a young man of 20, loved driving fast and showing his mates how he could push his car to the limit. One rainy day, with two of his mates with him in the car, he took a corner too fast, lost control, and slammed into a tree at 120 km/h (approximately 75 mph).” What do you think might be the consequences of this crash?
In the subsequent analysis, attention was paid to the order in which particular responses were given. The three main categories of consequence identified in order of frequency were death, serious injury, and damage to car/ property. No differences between age groups were found in the frequency of reporting any particular category. However, death was significantly less likely to be mentioned early as a consequence by the youngest group of drivers. Consistent with this, in a survey of young drivers in compulsory service in the Israeli Defense Forces, TaubmaneBen-Ari (2008) found that the cost of risk to life was not a predictor of any reckless driving measure. In a comprehensive review of brain imaging studies and decision making, Glendon (2008) noted that less welldeveloped executive functions of the brain in late adolescence may mean that implications of hazards are not so readily accessed. Linked to this is the observation that the
Chapter | 2
Driver Control Theory
integration of emotion with cognition, which appears to be mediated by the amygdala and hippocampus, is still maturing during this period.
5.2. The Somatic Marker Hypothesis In the somatic marker hypothesis, Damasio argues that elements of experience, such as objects, persons, and scenarios, automatically trigger an emotional response, albeit often only a weak one, whenever their representation is activated in the brain by either external or internal stimuli. Damasio proposes that in any situation requiring a decision, emotional signals “mark options and outcomes with a positive or negative signal that narrows the decisionspace and increases the probability that the action will conform to past experience” (Damasio, 2003, p. 148). This emotional signal has an auxiliary role that increases the efficiency of the reasoning process and is not usually a substitute for it. However, when we immediately reject an option that would lead to certain disaster, reasoning may be “almost superfluous”: The action may be taken without some intervening conscious cognitive processing. Because emotional signals are body related, Damasio labeled this set of ideas the somatic marker hypothesis. Through learning, somatic markers can become linked to stimuli and patterns of stimuli. When a negative somatic marker is linked to an image of a future outcome, it sounds an alarm. Slovic et al. (2002) refer to a similar concept as the “affect heuristic.” The key conclusions, however, are that not only is affect essential to rational action but also affective responses have a direct effect on cognitive operations (see Fuller (2007) for a complete exposition of Damasio’s conceptualization). Note that, as discussed previously, emotional responses in the form of somatic markers arise not only from stimuli external to the driver but also from perceived discrepancies between goal states and current states. Included in these discrepancies are where task demand exceeds the upper limit preferred by the driver (yielding a conscious feeling of anxiety, risk, or fear) and where progress goals are thwarted (yielding a conscious feeling of frustration, anger, or rage). The relevance of the somatic marker hypothesis for driver decision making has been discussed by Summala (2007), who suggests that “in dynamic time-limited situations like driving, fast affective heuristics must have a big role” (p. 198), and its potential implications for driver safety have been discussed by Fuller (2005b, 2007). Increases in risk may not be felt because of suppressed emotional reactivity (e.g., through alcohol, depression, denial, desensitization, and perhaps in conditions in which the outcome of the decision is uncertain; van Dijk & Zeelenberg, 2006) or because of the swamping effect of other emotions (e.g., anger and exhilaration). If felt, risk feelings may be misattributed to events other than those related to accomplishing the driving task (e.g., anxiety
23
from interaction with a passenger). Furthermore, experience may not have been sufficient to provide learning opportunities to link particular scenarios to feelings of risk (as with a novice driver) (Kinnear et al., 2008; Wickens, Toplak, & Wiesenthal, 2008). These and related issues for further research are discussed by Fuller (2005c), and a preliminary study of the contrasting roles of emotion and cognitive decision making as dispositional characteristics of drivers has been reported by Wickens et al. (2008).
6. ALTERNATIVE CONCEPTUALIZATIONS OF DRIVER GOALS Turning to recent proposals for what constitutes the driver’s control goals when driving, Vaa (2007) develops the implications of the somatic marker hypothesis for driver behavior in his “monitor model.” This argues that drivers may make adjustments to the prevailing road and traffic conditions with varying degrees of conscious awareness, on a continuum from unconscious adjustment to fully conscious decision making. He proposes that although risk feeling may describe one homeostatic target for drivers (referred to as tension/anxiety), other feelings may also be targeted. Candidates he suggests as “other feelings” are avoidance of threat or difficulty, compliance and noncompliance, arousal, sensation, joy, and relaxation. Feelings of avoidance of threat or difficulty are clearly related to task difficulty and feelings of risk, as discussed as targets in RAT, which also now incorporates dispositional and immediate influences on compliance. However, the wider range of target states motivating driver decision making proposed by Vaa’s monitor model describe rather the dispositional motives and immediate influences on risk threshold as described in RAT. Rather than being target conditions in themselves, I argue that dispositional motives and immediate influences operate to “set” the target level of risk feeling in the negative feedback control loop. Thus, the driver looking for more arousal raises his or her risk threshold to achieve that state, and the driver wanting to relax does the opposite. Summala’s theoretical development also appears to be moving in a more inclusive direction. Whereas the 1976 conceptualization developed with Risto Na¨a¨ta¨nen (Na¨a¨ta¨nen & Summala, 1976) postulated a subjective risk monitor that kicked in when risk experience exceeded a risk threshold, to both alert the driver and influence decision making, Summala now suggests that drivers operate not with just one target variable but with a whole range of them (Summala, 2007). He invokes the umbrella concept of a “comfort zone” to represent the range of values relating to each variable that drivers are assumed to be motivated to target: “It is hypothesized that drivers normally keep each of them within a certain range (or above a certain threshold) in a comfort
24
zone” (p. 201). Comfort is defined as a general mood or emotion that is “pleasant but not especially aroused, tense, or activated” (p. 201). Included in Summala’s (2007) target variables are space and time margins and mental load specifically relating to control. He also includes motivation for compliance. These variables may be translated in terms of the concept of a target range of task difficulty (operationalized in terms of time to collision and time to line crossing) and influences on compliance as represented in RAT. However, Summala adds various other target variables, including comfort in relation to thermal state, seating, vibration, glare, and rate of speed change and progress. Clearly, Summala’s model is shifting from one specifically concerned with control and collisions to one concerned with more general motives that inform driver decision making. From the perspective of RAT, “glare” and “rate of speed change and progress” may be subsumed under task demand (and therefore task difficulty) elements. However, Summala’s other comfort motives must be secondary to those relating to safety motivation. A driver will hardly survive for very long without crashing if he or she prioritizes temperature, seating comfort, or vibration as the target states that direct decision making. As pointed out by Carver (1994), “certain kinds of discrepancies are more demandingdmore importantdthan others.. For example, the experience of threat to one’s physical safety can override an attempt to engage in activities that are otherwise quite important” (p. 389). Nevertheless, these suggestions by Summala have enriched our conceptualization of potential aspects of driver motivation (we await empirical validation), even though their relevance to our understanding of why collisions occur is unclear. RAT is concerned with representing the process of driver decision making and in particular how motivations influence the outcome for system safety. However, in principle, it can be expanded to include the kinds of motives proposed by Summala. Their influence may be included in RAT as a top-down controlled hierarchy of secondary reference targets in decision making. Because a safe outcome normally has to be prioritized, they must enter the decision-making process after risk allostasis decisions have been made, perhaps at the point in the process where influences on compliance also have their effect. Summala’s extended target variables nevertheless raise a further question: When a control system has multiple reference standards, as he suggests, how do they operate in relation to each other? For example, are they implemented in serial order, as suggested in RAT, where compliance standards emerge as secondary to task difficulty targets, or can they operate in parallel? If the latter, the further question remains as to how their separate outputs are eventually integrated into the behavioral decision. Thus, if
PART | I
Theories, Concepts, and Methods
task difficulty is calling for an increase in speed and simultaneously compliance is calling for a reduction, how is the conflict resolved by his system? Perhaps the main conclusion to be drawn here is that despite the discrepancies that have emerged in conceptualizations of what drivers are aiming for in their decision making, these apparent tensions may in fact reflect a hidden consensus. At least, from the perspective of RAT, that is what I have tried to demonstrate. RAT proposes that driver control decisions are motivated by a desire to maintain feelings of risk (and its corollary task difficulty) within an acceptable range, even though for much of the time these feelings may be below the level of conscious awareness. The acceptable level of risk feeling and task difficulty may vary as a function of factors such as journey goals and emotional state, and there appear to be individual differences in preferred levels related to age, experience, gender, and personality. Constraints on the driver’s freedom to manage this process unavoidably arise from performance limitations of the vehicle as well as through obstructions caused by congested traffic flow that force driving at a lower level of task demand than that preferred. Freedom may also be restricted by compliance with regulated speed limits. With the development of RAT and related concepts advanced by Vaa and Summala, there is a current convergence in recognizing the primacy of the role of feeling in driver decision making. As Laertes says in Shakespeare’s Hamlet, “best safety lies in fear,” and this recognition opens up a whole new set of exciting and promising research questions.
REFERENCES ˚ berg, L., & Walle´n Warner, H. (2008). SpeedingdDeliberate violation A or involuntary mistake? European Review of Applied Psychology, 58, 23e30. Allen, T. M., Lumenfeld, H., & Alexander, G. J. (1971). Driver information needs. Highway Research Record, 366, 102e115. Andrey, J., Mills, B., Leahy, M., & Suggett, J. (2003). Weather as a chronic hazard for road transportation in Canadian cities. Natural Hazards, 28(2e3), 319e343. Bjo¨rklund, G. M. (2008). Driver irritation and aggressive behaviour. Accident Analysis and Prevention, 40, 1069e1077. Broughton, J. (2005). Les vitesses en Grande-Bretagne. In R. Delorme (Ed.), Les regimes franc¸ais et brittanique du re´gulation du risque routie`re: La vitesse d’abord (pp. 29e38). Paris: Centre Pour la Recherche E´conomique et ses Applications. Campbell, M., & Stradling, S. G. (2003). Factors influencing driver speed choices. In Behavioural research in road safety: Thirteenth seminar (pp. 233e244). London: Department for Transport. Carver, C. S. (1994). Cognitive processes and self-regulation: Determinants of concentration and distraction. Applied Psychology: An International Review, 43(3), 387e391.
Chapter | 2
Driver Control Theory
Carver, C. S., & Scheier, M. F. (1981). Attention and self-regulation: A control-theory approach to human behavior. New York: SpringerVerlag. Damasio, A. R. (1994). Descartes’ error: Emotion, reason and the human brain. New York: Putnam. Damasio, A. R. (2003). Looking for Spinoza: Joy, sorrow and the feeling brain. London: Heinemann. de Craen, S. (2010). The X-factor. A longitudinal study of calibration in young novice drivers (TRAIL thesis series T2010/2). Delft, The Netherlands: TRAIL Research School. Ellwanger, S. J. (2007). Strain, attribution, and traffic delinquency among young drivers: Measuring and testing general strain theory in the context of driving. Crime and Delinquency, 53(4), 523e551. Fastenmeier, W., & Gstalter, H. (2007). Driving task analysis as a tool in traffic safety research and practice. Safety Science, 45(9), 952e979. Fuller, R. (1984a). A conceptualisation of driving behaviour as threat avoidance. Ergonomics, 27, 1139e1155. Fuller, R. (1984b). Prolonged driving in convoy: The truck driver’s experience. Accident Analysis and Prevention, 16, 371e382. Fuller, R. (2000). The taskecapability interface model of the driving process. Recherche Transports Se´curite´, 66, 47e59. Fuller, R. (2005a). Towards a general theory of driver behaviour. Accident Analysis and Prevention, 37, 461e472. Fuller, R. (2005b). Control and affect: Motivational aspects of driver decision-making. In L. Macchi, C. Re, & P. C. Cacciabue (Eds.), Proceedings of the international workshop on modelling driver behaviour in automotive environments (pp. 45e52). Luxembourg: Office for Official Publication of the European Communities. Fuller, R. (2005c). Driving by the seat of your pants: A new agenda for research. In Behavioural research in road safety 2005, fifteenth seminar (pp. 85e93). London: Department for Transport. Fuller, R. (2007). Motivational determinants of control in the driving task. In P. Cacciabue (Ed.), Modelling driver behaviour in automotive environments: Critical issues in driver interactions with intelligent transport systems (pp. 165e188). London: Springer-Verlag. Fuller, R., Bates, H., Gormley, M., Hannigan, B., Stradling, S., Broughton, P., Kinnear, N., & O’Dolan, C. (2008). The conditions for inappropriate high speed: A review of the research literature from 1995 to 2006 (Road Safety Research Report No. 92). London: Department for Transport. Fuller, R., McHugh, C., & Pender, S. (2008). Task difficulty and risk in the determination of driver behaviour. European Review of Applied Psychology, 58, 13e21. Glendon, I. (2008). Changes in brain structures of younger and older drivers: Meeting the challenge. Paper presented at the 4th International Conference on Traffic and Transport Psychology, Washington DC, August 31eSeptember 4, 2008. Godthelp, J. (1988). The limits of path error neglecting in straight lane driving. Ergonomics, 31, 609e619. Gormley, M., & Fuller, R. (2008). Investigation of high risk behaviour in Irish young male drivers. Ballina, Ireland: Report to the Road Safety Authority, Government Offices. Gormley, M., Walsh, T., & Fuller, R. (2008). Risks in the driving of emergency service vehicles. Irish Journal of Psychology, 29(1e2), 7e18. Harre´, N., & Sibley, C. G. (2007). Explicit and implicit self-enhancement biases in drivers and their relationship to driving violations and crashrisk optimism. Accident Analysis and Prevention, 39(6), 1155e1161.
25
Horrey, W. J., Lesch, M. F., & Garabet, A. (2008). Assessing the awareness of performance decrements in distracted drivers. Accident Analysis and Prevention, 40(2), 675e682. Jamson, S. L. (2008). Reducing the anger and anxiety experienced at motorway roadworks. Paper presented at the 4th International Conference on Traffic and Transport Psychology, Washington DC, August 31eSeptember 4, 2008. Kalat, J. W. (2008). Introduction to psychology (8th ed.). New York: Thomson. Kilpela¨inen, M., & Summala, H. (2007). Effects of weather and weather forecasts on driver behaviour. Transportation Research Part F, 10, 288e299. King, Y., & Parker, D. (2008). Driving violations, aggression and perceived consensus. European Review of Applied Psychology, 58, 43e49. Kinnear, N., Stradling, S., & McVey, C. (2008). Do we really drive by the seat of our pants? In L. Dorn (Ed.), Driver behaviour and training, Vol. 3 (pp. 349e365) Aldershot, UK: Ashgate. Lam, L. T. (2003). Factors associated with fatal and injurious car crash among learner drivers in New South Wales, Australia. Accident Analysis and Prevention, 35, 333e340. Larsen, L. (1995). Factors affecting drivers’ choice of speed. Madrid: IAAP Symposium. Lev, D., Hershkovitz, E., & Yechiam, E. (2008). Decision making and personality in traffic offenders: A study of Israeli drivers. Accident Analysis and Prevention, 40(1), 223e230. Lewis-Evans, B., & Charlton, S. G. (2006). Explicit and implicit processes in behavioural adaptation to road width. Accident Analysis and Prevention, 38, 610e617. Lewis-Evans, B., & Rothengatter, T. (2009). Task difficulty, risk, effort and comfort in a simulated driving task: Implications for risk allostasis theory. Accident Analysis and Prevention, 41, 1053e1063. Liu, B.-S., & Lee, Y.-H. (2005). Effects of car-phone use and aggressive disposition during critical driving maneuvers. Transportation Research Part FeTraffic Psychology and Behaviour, 8, 369e382. Loewenstein, G. F., Weber, E. U., Hsee, C. K., & Welch, N. (2001). Risk as feelings. Psychological Bulletin, 127, 267e286. Lustman, M., & Wiesenthal, D. (2008). Narcissism and aggressive driving: Is inflated self-esteem a road hazard? Paper presented at the 4th International Conference on Traffic and Transport Psychology, Washington DC, August 31eSeptember 4, 2008. Machin, M. A., & Sankey, K. S. (2008). Relationships between young drivers’ personality characteristics, risk perceptions, and driving behaviour. Accident Analysis and Prevention, 40(2), 541e547. Mesken, J., Hagenzieker, M. P., Rothengatter, T., & de Waard, D. (2007). Frequency, determinants, and consequences of different drivers’ emotions: An on-the-road study using self-reports, (observed) behaviour, and physiology. Transportation Research Part F, 10, 458e475. Michon, J. A. (1989). Explanatory pitfalls and rule-based driver models. Accident Analysis and Prevention, 21, 341e353. Møller, M., & Gregersen, N. P. (2008). Psychosocial function of driving as predictor of risk-taking behaviour. Accident Analysis and Prevention, 40(1), 209e215. Musselwhite, E. C. (2006). Attitudes towards vehicle driving behaviour: Categorising and contextualising risk. Accident Analysis and Prevention, 38, 324e334.
26
Na¨a¨ta¨nen, R., & Summala, H. (1976). Road user behaviour and traffic accidents. Amsterdam/New York: North Holland/Elsevier. Organisation for Economic Co-operation and Development. (2006). Young drivers: The road to safety. Paris. Organisation for Economic Co-operation and Development/European Conference of Ministers of Transport. (2006). Speed management. Paris. Paris, H., & Van den Broucke, S. (2008). Measuring cognitive determinants of speeding: An application of the theory of planned behaviour. Transportation Research Part F, 11, 168e180. ¨ stlund, J., Nilsson, L., & Svenson, O. Patten, C. J. D., Kircher, A., O (2006). Driver experience and cognitive workload in different traffic environments. Accident Analysis and Prevention, 38, 887e894. Ranney, T. A. (1994). Models of driving behavior: A review of their evolution. Accident Analysis and Prevention, 26, 733e750. Saad, F., Delhomme, P., & Van Eslande, P. (1990). Drivers’ speed regulation when negotiating intersections. In M. Koshi (Ed.), Transportation & traffic theory (pp. 193e212). Amsterdam: Elsevier. Simonet, S., & Wilde, G. J. S. (1997). Risk: Perception, acceptance and homeostasis. Applied Psychology: An International Review, 46, 235e252. Slovic, P., Finucane, M. L., Peters, E., & MacGregor, D. G. (2002). Risk as analysis and risk as feelings. Some thoughts about affect, reason, risk and rationality. Paper presented at the annual meeting of the Society for Risk Analysis. New Orleans: Louisiana. December 10, 2002. Stradling, S., Broughton, P., Kinnear, N., O’Dolan, C., Fuller, R., Bates, H., Gormley, M., & Hannigan, B. (2008). Understanding inappropriate high speed: A quantitative analysis (Road Safety Research Report No. 93). London: Department for Transport. Summala, H. (1986). Risk control is not risk adjustment: The zero-risk theory of driver behavior and its implications. In Reports (No. 11). Helsinki: University of Helsinki Traffic Research Unit. Summala, H. (2007). Towards understanding motivational and emotional factors in driver behaviour: Comfort through satisficing. In P. Cacciabue (Ed.), Modelling driver behaviour in automotive environments: Critical issues in driver interactions with intelligent transport systems (pp. 189e207). London: Springer-Verlag. TaubmaneBen-Ari, O. (2008). Motivational sources of driving and their associations with reckless driving cognitions and behavior. European Review of Applied Psychology, 58, 51e64. Taylor, D. H. (1964). Drivers’ galvanic skin response and the risk of accident. Ergonomics, 7, 439e451.
PART | I
Theories, Concepts, and Methods
Uzzell, D., & Muckle, R. (2005). Simulating traffic engineering solutions to predict changes in driving behaviour. Transportation Research Part FdTraffic Psychology and Behaviour, 8, 311e329. Vaa, T. (2007). Modelling driver behaviour on basis of emotions and feelings: Intelligent transport systems and behavioural adaptations. In P. Cacciabue (Ed.), Modelling driver behaviour in automotive environments: Critical issues in driver interactions with intelligent transport systems (pp. 208e232). London: Springer-Verlag. Van der Horst, R. (2007). Time-related measures for modeling risk in driver behaviour. In P. Cacciabue (Ed.), Modelling driver behaviour in automotive environments: Critical issues in driver interactions with intelligent transport systems (pp. 235e252). London: Springer-Verlag. Van der Hulst, M., Meijman, T., & Rothengatter, T. (1999). Anticipation and the adaptive control of safety margins in driving. Ergonomics, 42(2), 336e345. van Dijk, E., & Zeelenberg, M. (2006). The dampening effect of uncertainty on positive and negative emotions. Journal of Behavioral Decision Making, 19, 171e176. Vassallo, S., Smart, D., Sanson, A., Harrison, W., Harris, A., Cockfield, S., & McIntyre, A. (2007). Risky driving among young Australian drivers: Trends, precursors and correlates. Accident Analysis and Prevention, 39, 444e458. Walsh, T., Hannigan, B., & Fuller, R. (2008). Risk and the driving of emergency service vehicles. Behavioural research in road safety 2008. London: Department for Transport. Wickens, C. D., & Hollands, J. G. (2000). Engineering psychology and human performance (3rd ed.). Upper Saddle River, NJ: Prentice Hall. Wickens, C. M., Toplak, M. E., & Wiesenthal, D. L. (2008). Cognitive failures as predictors of driving errors, lapses, and violations. Accident Analysis and Prevention, 40, 1223e1233. Wikman, A. S., Nieminen, T., & Summala, H. (2008). Driving experience and time-sharing during in-car tasks on roads of different width. Ergonomics, 41(3), 358e372. Wikman, A. S., & Summala, H. (2005). Aging and time-sharing in highway driving. Optometry and Vision Science, 82(8), 716e723. Wilde, G. J. S. (1982). The theory of risk homeostasis: Implications for safety and health. Risk Analysis, 2, 209e225. Young, M. S., Mahfoud, J. M., Walker, G. H., Jenkins, D. P., & Stanton, N. A. (2008). Crash dieting: The effects of eating and drinking on driving performance. Accident Analysis and Prevention, 40(1), 142e148. Zajonc, R. B. (1980). Feeling and thinking: Preferences need no inferences. American Psychologist, 35, 151e175.
Chapter 3
CaseeControl Studies in Traffic Psychology Martha Hı´jar,* Ricardo Pe´rez-Nu´n˜ez* and Cristina Incla´n-Valadezy *
National Institute of Public Health, Cuernavaca, Morelos, Mexico, y London School of Economics and Political Science, London, UK
1. INTRODUCTION Although the effects of urbanization and industrialization in most countries suggest a degree of inevitability, substantial reductions in rates of road crash fatalities have been achieved in high-income countries despite increasing motorization. Evidence suggests that for the majority of the world’s population, the burden of road traffic injuries is increasing dramatically (Ameratunga, Hijar, & Norton, 2006). Risk in road traffic derives from a need to travel for different reasons and from a range of factors that determine who uses different parts of the transport system, how it is used and why, and at what times (Tingvall, 1997). The concept of risk in road safety includes factors related to exposure, considered as the amount of movement or travel within the transport system by different users or a given population density; the crash probability given a particular exposure; the probability of being injured after a crash; and the outcome of injury. It has been documented that, although it might not be possible to eliminate all risks, it is possible to reduce the exposure to risk of severe injury or minimize its intensity and fatal consequences. The global and concerted approach required must consider the wider societal burden of road traffic crashes (both fatal and nonfatal outcomes) and particularly focus on efforts to protect vulnerable road users (e.g., motorcyclists, human-powered vehicles, and pedestrians). Addressing the disparities in both the impact of and response to this problem must be high on the global public health policy and research agenda (Ameratunga et al., 2006). A substantial body of literature points to the propensity of some road user groups, particularly pedestrians and those using motorized and nonmotorized two-wheelers, to be vastly overrepresented among crash victims at the global level (Peden et al., 2004; Razzak & Luby, 1998) and be at higher risk of crash-related disability (Mayou & Bryant, 2003). Passengers in formal and informal modes of public
Handbook of Traffic Psychology. DOI: 10.1016/B978-0-12-381984-0.10003-7 Copyright Ó 2011 Elsevier Inc. All rights reserved.
and mass transport constitute another important road user group that is a common feature among road crash data, especially from less resourced environments. Error by a road user may indeed trigger a crash but may not necessarily be its underlying cause. In addition, human behavior is governed not only by individual knowledge and skills but also by the environment in which the behavior takes place (Khayesi, 2003). Indirect influences, such as the design and layout of the road, the nature of the vehicle, and traffic laws and their enforcement or lack of enforcement, affect behavior in important ways. For this reason, the use of information and publicity on their own is generally unsuccessful in reducing road traffic collisions (Allsop, 2002; European Road Safety Action Programme, 2003; Impacts Monitoring Group in the Congestion Charging Division of Transport for London, 2003). Error is part of the human condition. Aspects of human behavior in the context of road traffic safety can certainly be altered. Nonetheless, errors can also be effectively reduced by changing the immediate environment rather than focusing solely on changing the human condition (Wang et al., 2003). This chapter provides general guidelines on how to use the epidemiological approach to study the problem of road traffic injuries using the caseecontrol study design. The chapter begins with a general description of epidemiological study designs. It proceeds to give a more comprehensive description of caseecontrol studies, specifically how these studies are defined and when they are most commonly used, followed by “case” definition and alternatives to select cases. Subsequently, it presents the definition of a “control” and explains how controls can be identified and selected. Reasons to match are then presented, followed by a discussion of the forms of matching and stratification and of the disadvantages of matching strategies. An argumentation on how many controls should be included in caseecontrol studies is 27
28
PART | I
then presented, followed by a discussion on how to analyze caseecontrol studies and what measures of associationecausality are used in this type of study. Several variants of caseecontrol studies are presented. Next, the chapter addresses the problem of representativeness and discusses some of the principal biases that are relevant and characteristic of caseecontrol studies, especially in road safety research and analysis. Finally, the advantages and disadvantages of caseecontrol studies are discussed.
2. EPIDEMIOLOGICAL STUDY DESIGNS The purpose of epidemiology is to describe and explain the population health dynamicsdto identify the elements that compose it and to understand the forces that governs it in order to develop actions aimed to preserve and promote ´ vila & Lo´pezhealth between populations (Herna´ndez-A Moreno, 2007). It is thus not only concerned with the occurrence of disease or other health-related events but also with the identification of factors that cause those conditions, which has become the main focus of modern epidemiology (dos Santos-Silva, 1999b). In general, researchers studying road safety attempt to answer the following questions: 1. Does alcohol consumption increase the risk of pedestrian injuries (Haddon, Valien, McCarroll, & Umberger, 1961)? Alcohol consumption (exposure)
Pedestrian injuries (outcome)
Theories, Concepts, and Methods
2. Does seat belt use decrease the risk of severe road traffic injuries in a car collision (Hı´jar-Medina, Flores-Aldana, & Lopez-Lopez, 1996)? Seat belt use (exposure)
Severe road traffic injuries (outcome)
3. What environmental factors could increase the risk of pedestrian and cyclist road traffic injuries (Kraus et al., 1996)? Environmental factors (exposure)
Pedestrian and cyclist road traffic injuries (outcome)
´ vila and Lo´pez-Moreno (2007) classified Herna´ndez-A epidemiological studies using a multidimensional approach (Table 3.1), including the following: 1. Assignation of exposure: observational, experimental, and Quasi-experimental. 2. Number of measurements performed in each study subject to verify changes in the occurrence of both exposition and its effect (longitudinal versus crossover study design). 3. Criteria employed to select population under study (none, exposition, and effect). 4. Temporal relationship between the start of the study and the measurement of the occurrence of the effect (retrospective, prospective, mixed, or ambispective). 5. Unit of analysis for which all variables of interest are measured (individual, group, and population). It is important to note that the concept of an individual in
TABLE 3.1 Multidimensional Classification of Epidemiological Studies
Type of study
Assignation of Exposition
No. of Observations (Measurements) by Individual
Selection Criteria of Population Temporality of under Study Analysis
Unit of Analysis
Experimental
Controlled (random)
Two or more
None
Prospective
Individual or group
Pseudo-experimental
For/by convenience
Two or more
None
Prospective
Individual or group
Cohort
Out of the control of researcher
Two or more
Exposition
Prospective or retrospective
Individual
Cases and controls
Out of the control of researcher
One or more
Effect
Prospective or retrospective
Individual
Crossover
Out of the control of researcher
One
None
Retrospective
Individual
Ecological
Out of the control of researcher
Two or more
None
Retrospective
Group or population
Source: Herna´ndez-A´vila and Lo´pez-Moreno (2007).
Chapter | 3
CaseeControl Studies in Traffic Psychology
road safety research has not been reduced to a single person. Some studies have also had as a unit of analysis streets (Kraus et al., 1996), street crossing locations (Koepsell et al., 2002), crash sites (Wintemute, Kraus, Teret, & Wright, 1990), etc. The previous classification categorizes epidemiological studies according to the strength of evidence that each study design provides to the causal relationship between exposure variables and a health outcome of interest ´ vila & Lo´pez-Moreno, 2007). In this sense, (Herna´ndez-A the best study design to establish cause-and-effect relationships is the experimental randomized design. For strictly medical interventions, the “gold standard” is the double-blind, randomized controlled trial. This study design involves the random allocation of different interventions (treatments or conditions) of studied subjects to compare treatment groups with control groups not receiving the treatment. Participants, caregivers, or outcome assessors are not allowed to know which intervention are they receiving. Although these studies may be ideal for testing the efficacy of interventions, there are many instances in which trials would be impossible, impractical, and/or unethical. For example, it would generally be considered unethical to randomly assign research subjects to be exposed to alcohol in order to evaluate the substance’s effects on their driving ability. In this sense, the choice of study design is affected by numerous factors and considerations. According to Robertson (1992), this decision depends on
29
what is the unit of analysis (people, vehicles, environment)? In what population should the study be conducted? To what population of people, vehicles, or environments will the results be generalized? What kind of measurements of the factors are available or could be obtained? How reliable and valid are the measurements? Can the data be collected without violating ethical guidelines? How can the study isolate the effects of given factors independent of, or in combination with, other relevant factors? How much time will be needed to complete the study? How much will the study cost? (pp. 84e85)
The next section provides a general definition of casee control studies and discusses examples of caseecontrol studies in the road safety field.
3. CASEeCONTROL STUDIES 3.1. Definition and Characteristics Caseecontrol studies represent a sampling strategy in which the population under study is selected based on the presence (case) or absence (control) of an event of interest (i.e., health condition, disease, and death) (Lazcano-Ponce, Salazar-Martinez, & Hernandez-Avila, 2001). The underlying purpose of these studies is to identify causal factors of the events of interest by comparing characteristics of both groups (cases and controls). As shown in Figure 3.1, caseecontrol studies start by identifying the study population, from which cases are identified and their exposure status is determined retrospectively. Then, a control group FIGURE 3.1 Design of caseecontrol studies. Source: Herna´ndez-A´vila and Lo´pez-Moreno (2007).
Non-eligible Population
Nonparticipants Eligible Participants
Identification of cases
Selection of controls
Cases
Controls
Study population
Reconstruction of exposition
Beginning of study
30
of study subjects is sampled from the entire source population that gives rise to the cases (Rothman & Greenland, 1998). Once the exposure status is also identified in the control group, comparison between both groups may evidence whether exposure to a specific factor is higher in the case group (risk factor) or lower (protective factor) or the same as in the control group (no evidence of associa´ vila & Lo´peztion) (dos Santos-Silva, 1999a; Herna´ndez-A Moreno, 2007). It is important to remember that assignation of exposure in caseecontrol studies is out of the control of researchers. A good example of a caseecontrol study in road safety is the study of Jones, Harvey, and Brewin (2005). This study explored the symptom profiles of acute stress disorder and post-traumatic stress disorder (PTSD) in participants who did or did not sustain traumatic brain injury following a road traffic accident. This study selected as “case” all survivors of a road traffic collision during a period of time who had been diagnosed with traumatic brain injury (TBI). The control group was composed of survivors of a road traffic accident during a period of time with no TBI (Jones et al., 2005). Once the exposure status was identified, comparison between both groups could evidence whether exposition to a specific factor is higher in the case group (risk factor) or lower (protective factor) or the same as in the control group (no evidence of associa´ vila & Lo´peztion) (dos Santos-Silva, 1999a; Herna´ndez-A Moreno, 2007). This study found that at 3 months posttrauma, there was no difference in PTSD symptom profile between non-TBI (controls) and TBI groups (cases). Caseecontrol studies can be conceptualized within the framework of a hypothetical cohort study (Rothman & Greenland, 1998). Although in practice it can be difficult to characterize the cohort or study base (Wacholder, McLaughlin, Silverman, & Mandel, 1992), caseecontrol studies can be based on special cohorts of interest rather than on the general population (Rothman & Greenland, 1998).
3.1.1. When Are CaseeControl Studies Used? Caseecontrol studies are frequently one of the first approaches used in the etiological study of a disease or health condition. This is in part due to the possibility of incorporating in the analysis many exposition factors simultaneously and relatively quickly and inexpensively (dos Santos-Silva, 1999a). Therefore, caseecontrol studies represent a cost-effective way of identifying risk and protective factors and generating hypotheses for subsequent, methodologically stronger studies (Lazcano-Ponce et al., 2001). Caseecontrol design is simply an efficient sampling technique for measuring exposureedisease associations in a cohort or study base (Wacholder, et al., 1992). In addition, caseecontrol studies are commonly used to study conditions
PART | I
Theories, Concepts, and Methods
that are relatively rare or that have a prolonged induction period (dos Santos-Silva, 1999a). A study carried out in Shanghai, China, is an example of a caseecontrol study measuring the effect of exposition to different risk factors (Yu, Wang, & Chen, 2005). This study explored the risk factors influencing the occurrence of road traffic injuries on drivers with a history of accidents and on controls. The study included physiological, psychological, and behavioral risk factors and found that factors such as tiredness and waking up early were related to the occurrence of road traffic injuries. It is important to highlight that the human host or vector (pedestrian and driver) in road traffic injuries has been the unit of analysis in most caseecontrol studies, to the neglect of factors that may be more subject to change for injury control. However, caseecontrol designs can also provide strong evidence regarding environmental factors (Robertson, 1992).
3.2. Case Definition As can be seen in the previous examples, the definition of a case can be virtually anything that the investigator wishes: an injured person from a specific gender or age group or a specific road userdpedestrian, cyclist, motorcyclist, or car occupant. Whoever the case is, the case definition will implicitly define the source population for cases, from which also the controls should be drawn (Rothman & Greenland, 1998). In this sense, having precise criteria to define a case is highly relevant. Objective documentation that cases actually have the disease or health condition under study is highly recommended. When this is not possible, an alternative is to classify cases as “confirmed,” “probable,” or “likely.” If analysis shows a gradual decrease in relative risk from the confirmed category to the likely category, problems of erroneous classification are suspected. For this reason, this classification gives researchers the opportunity to evaluate the probability that results are affected by an incorrect classification of the disease analyzed (dos Santos-Silva, 1999a). In general, there are two types of cases: incident and prevalent. Incident cases are those new cases that appear in the population under study in a specific period of time (or during a pre-established period of time). Memory of past events and exposures tends to be more accurate in cases recently diagnosed. For this reason, incident cases are preferred over prevalent cases. In addition, it is less probable that incident cases changed their habits (exposures) as a result of disease. Prevalent cases are all cases existent (new and previous) in a population in a specific time (or a short period of time) (dos Santos-Silva, 1999a). Prevalent cases are especially useful when it is not possible to establish a specific date for disease onset.
Chapter | 3
CaseeControl Studies in Traffic Psychology
However, patients with a condition of prolonged duration tend to be overrepresented because those with a condition of short duration drop out of the study due to recovery or death. Unless exposure under study is not related to recovery or survival, incident cases should be privileged when designing a caseecontrol study (dos Santos-Silva, 1999a).
3.2.1. Identification and Selection of Cases Selection of cases should privilege internal validity rather than external validity (dos Santos-Silva, 1999a). Internal validity refers to the absence of errors made during the selection process of the population under study or during the measurement of individual variables of interest. To achieve internal validity, comparability of groups under study should be met (Hernandez-Avila, Garrido, & SalazarMartinez, 2000). On the other hand, external validity refers to the capacity of a study to generalize observed results to the base population. A prerequisite of external validity is the achievement of internal validity. This is the reason why internal validity should be privileged over external validity (Hernandez-Avila et al., 2000). In addition, selection of cases should include only those for whom the reasonable possibility exists that the disease or health condition existed prior to the study (dos SantosSilva, 1999a). Ensuring that cases comprise a relatively homogeneous group will increase the possibility of detecting important etiological relations. It is less important to be able to generalize results to the entire population than to establish an etiologicalecausal relation even when this
31
relation only applies for a small group of the population (dos Santos-Silva, 1999a). Ideally, selection of cases should follow the paradigm of longitudinal studies. It is recommended to select recently diagnosed cases (incident cases). Less recommended is the use of prevalent cases unless other requisites are met or when justified (Table 3.2;Lazcano-Ponce et al., 2001). Commonly used sources of cases in the published literature on road traffic injuries are hospital cases (Celis, Gomez, Martinez-Sotomayor, Arcila, & Villasenor, 2003; Tester, Rutherford, Wald, & Rutherford, 2004), administrative registries such as city police reports (Lightstone, Peek-Asa, & Kraus, 1997; von Kries, Kohne, Bohm, & von Voss, 1998), and coroners’ registries (Wintemute et al., 1990). Also common is a combination of sources (Stevenson, Jamrozik, & Burton, 1996). Even when cases are identified exclusively at hospital points, they can be reasonably assumed to represent all cases in a region or determined study population when, for example, the severity of the condition or disease requires hospitalization. This is the case for severe road traffic injuries, for which health care utilization patterns are different from those of most other diseases due to the severity of injuries and because the urgent demand for treatment eliminates some of the most common access barriers (Hı´jar-Medina & Va´zquez-Vela, 2003). However, slight injuries that are treated in a hospital or emergency room cannot be considered as representative of the total study base. When reporting results of a caseecontrol study, it is important to specify which cases were not included in the
TABLE 3.2 Options for Selection of Cases Option
Characteristics
Utilization of incident cases with long exposure periods or prolonged latency periods
OR tends to be similar to RR when cases under study are incident and preceded by a long-term exposure.
Use of prevalent cases with prolonged exposure period
OR is similar to RR if disease does not affect the status of exposure and there is a long-standing exposure period. Prevalent cases could be included, especially when new cases are not available (low prevalent conditions), lethality of disease is low, and exposure does not modify the clinical outcome of the disease (survival).
Utilization of incident cases and very short exposure periods
OR is similar to RR when the risk period is short and incident cases are used.
Utilization of prevalent cases
OR comes closer to RR when prevalence of cases is low only if outcome is not related with survival before selection, condition, or exposure and if disease does not affect the exposure status.
Utilization of death cases
Inclusion of death cases is only justified in exposures that could be quantified through the use of high-quality secondary sources of data, such as medical records and occupational information sources.
OR, odds ratio; RR, relative risk. Source: Lazcano-Ponce et al. (2001).
32
PART | I
study even though they satisfied all inclusion criteria. Reasons for exclusion and the number of cases by reason should be specified. This information allows the assessment of the level to which study results may be affected by a selection bias (dos Santos-Silva, 1999a). l
3.3. Definition of a Control A control is an individual without the condition of interest who serves as reference for a case. The purpose of the control group is to determine the relative (as opposed to absolute) size of the exposed and unexposed denominators within the source population. From the relative size of the denominators, the relative size of the incidence rates (or incidence proportions, depending on the nature of the data) can be estimated (Rothman & Greenland, 1998). Thus, caseecontrol studies yield estimates of relative effect measures (Rothman & Greenland, 1998). It is important to note that controls should meet all eligibility criteria defined for cases other than those related to the diagnostic of disease, outcome, or health condition under analysis (dos Santos-Silva, 1999a).
l
l
l
3.3.1. Identification and Selection of Controls Conceptually, it can be assumed that all caseecontrol studies are nested inside a particular population (dos Santos-Silva, 1999a). This is called the “study base,” which can also be thought of as the members of the underlying cohort or source population for the cases during the time periods when they are eligible to become cases (Wacholder, et al., 1992). However, the identification of the appropriate study base from which to select controls is the primary challenge in the design of caseecontrol studies (Wacholder, et al., 1992). For this reason, selecting a control group could be the most difficult part of a caseecontrol study (dos Santos-Silva, 1999a). Controls are selected from the same population base as cases but through a mechanism independent from that used for case ´ vila & Lo´pez-Moreno, 2007). selection (Herna´ndez-A Some authors have set basic principles for control selection that are required to minimize bias, including the following: l
l
Cases and controls should be “representative of the same population base experience” (Wacholder, et al., 1992, p. 1020). Operationally, this implies that if a control develops the condition or disease (the event under study), he or she must be included as a “case” in the study (Lazcano-Ponce et al., 2001). Confounding should not be allowed to distort the estimation of effect. This is referred to as the deconfounding principle. Confounders that are measured can be controlled in the analysis. Unknown or unmeasured
l
Theories, Concepts, and Methods
confounders should have as little variability as possible. Because this variability is measured conditionally on the levels of other variables being studied, the use of stratification or matching can, in effect, reduce or eliminate the variability of the confounder (Wacholder, et al., 1992). Controls must be sampled independently of their exposure status to ensure that they represent the population base (Rothman & Greenland, 1998). The degree of accuracy in measuring the exposure of interest for the cases should be equivalent to the degree of accuracy for the controls, unless the effect of the inaccuracy can be controlled in the analysis (Wacholder, et al., 1992). The probability of selection of a control should be proportional to the time a subject has remained eligible to develop the event or condition under study. This implies that not only are all controls at risk of developing the condition but also subjects selected as controls in an early stage could become cases in latter stages (Lazcano-Ponce et al., 2001). The study should be implemented so as to learn as much as possible about the questions being investigated for a fixed expenditure of time and resources (Wacholder, et al., 1992). This has been called the efficiency principle, and it calls for consideration of costs as well as validity in selection of controls. Statistical efficiency refers to the amount of information obtained per subject; more broadly, efficiency encompasses the time and energy needed to complete the study (LazcanoPonce et al., 2001; Wacholder, et al., 1992). An exclusion rule that applies equally to cases and controls is valid because it simply refines the scope of the study base (Wacholder, et al., 1992).
Results of a caseecontrol study become more credible to the extent that these principles are met. The objective of the principles is to reduce or eliminate selection bias, confounding bias, and information bias (Wacholder, et al., 1992). Perfect adherence to a principle can be as difficult to achieve as perfect experimental conditions in a laboratory. Sometimes, one principle can conflict with another. Indeed, tolerating a minor violation of a principle is often the only way to study a particular exposureedisease association. Such a study can still provide valuable information, particularly when the impact of the violation can be measured (Wacholder, et al., 1992). 3.3.1.1. Source of Controls Most of the sources of controls for epidemiological research are presented in Table 3.3 along with their advantages and disadvantages. However, studies of road traffic injuries commonly use controls obtained from hospitals and neighborhoods. For example, as discussed previously, one of the most popular and practical strategies
Chapter | 3
CaseeControl Studies in Traffic Psychology
to identify road traffic injury cases is through a hospital or emergency room. It could be difficult to consider that they represent all persons injured as a result of road traffic collisions, although this may be true when a hospital covers
33
the entire population under study and there are no access barriers for individuals. If not, they might be claimed to be representative only of road traffic injury users of a specific hospital or emergency room. Controls, however, may
TABLE 3.3 Advantages and Disadvantages of Different Types of Controls Type of Controls
Advantages
Disadvantages
Population controls
Same study base: Ensures that the controls are drawn from the same source population as the case series.
Inappropriate when there is incomplete case ascertainment or when even approximate random sampling of the study base is impossible because of nonresponse or inadequacies of the sampling frame.
Exclusions. Definition of the base can encompass the exclusions. Extrapolation to base population: Distribution of exposures in the controls can be readily extrapolated to the base for purposes such as calculations of absolute or attributable risk.
Inconvenience: Definition of the base can encompass the exclusions. Recall bias: Responses by a previously hospitalized case may reflect modifications in exposure due to the disease, such as drinking less coffee or alcohol after an ulcer, or due to changes in perception of past habits after becoming ill. Less motivation to cooperate.
Random digit dialing
In some circumstances, could come close to sampling randomly from the source population.
Probability of contacting each eligible control will not necessarily be the same because households vary in the number of people who reside in them and the amount of time someone is at home. Contact with a household may require many calls at various times of day and various days of the week. Challenging to distinguish business from residential telephone numbers.
Neighborhood controls
Convenient substitute for population-based sampling of controls. Control of environmental or socioeconomic confounding factors.
If a person is injured in a neighborhood, controls who have knowledge of the injury may give misleading information because of denial of personal vulnerability or other psychological factors. Overmatching. Could introduce selection bias because it cannot be assumed that controls represent the base population from which cases were extracted.
School rosters
Especially useful when population under study is of school age.
Selection bias in contexts of high rates of school desertion.
Hospital or disease registry controls
Comparable quality of information.
Different catchments: Catchments for different diseases within the same hospital may be different.
Convenience. Factors such as socioeconomic characteristics, race, and religion can be controlled. Normally, they tend to be willing to participate and to provide complete and exact information.
Other diseases obtained from a population registry
Comparable quality of information.
Controls from a medical practice
Useful strategy when it is otherwise difficult to find controls who are comparable to cases on access to medical care or referral to specialized clinics.
Willing to participate and to provide complete and exact information.
Berkson’s bias: Caused by selection of subjects into a study differentially on factors related to exposure. Disease of controls could be related to exposure (risk factors).
Berkson’s bias: Caused by selection of subjects into a study differentially on factors related to exposure. Disease of controls could be related to exposure (risk factors). The study base principle can be jeopardized with medical practice controls because the exposure distribution for controls may not be the same as that in the study base. (Continued)
34
PART | I
Theories, Concepts, and Methods
TABLE 3.3 Advantages and Disadvantages of Different Types of Controlsdcont’d Type of Controls
Advantages
Disadvantages
Friend controls
More convenient and inexpensive source of controls.
The credibility of representativeness of exposure is low for factors related to sociability, such as gregariousness or, possibly, smoking, diet, or alcohol consumption, because sociable people are more likely to be selected as controls than are loners.
Controls can be selected from a list of friends or associates obtained from the case at little extra effort while the case is being interviewed. Friends may be likely to use the medical system in similar ways. Moreover, biases due to social class are reduced because usually the case and friend control will be of a similar socioeconomic background. Despite serious shortcomings, friend controls may be useful in some exceptional circumstances, such as in a study of exposures unrelated to friendship characteristics, as is likely in a study of a genetically determined metabolic disorder. Relative controls Useful when genetic factors confound the effect of exposure, blood relatives of the case have been used as a source of controls in an attempt to match on genetic background.
“Friendly control” bias: Sociable people are more likely to be selected as controls than are loners. Loners, although not on anyone’s list, can become a case. A less serious problem is that the use of friend controls can lead to overmatching because friends tend to be similar with regard to lifestyle and occupational exposures of interest. Some cases may not be willing to provide names of friends, increasing nonresponse. Cases and controls may be overmatched on a variety of genetic and environmental factors that are not risk factors but are related to the exposure under study.
Spouses might be a suitable control group if matching on adult environmental risk factors is sought. The case series Only patients need to be studied, and recurrences can as the source of be handled easily. controls
For studies of chronic diseases in which the main focus is on more stable time-dependent covariates, the use of a study series of cases only, as might be found in a disease registry, requires a complete and accurate exposure history and the strong assumption that the exposure of interest is unrelated to overall mortality. This study design may also have lower power than more conventional studies.
Proxy Useful when subjects are deceased or too sick to respondents and answer questions or for persons with perceptual or deceased cognitive disorders. controls Provide accurate responses for broad categories of exposure information, and sometimes even better information than the index subjects.
Because proxy respondents will tend to be used more often for cases than for healthy controls, violation of the comparable accuracy principle is likely. More detailed information is usually less reliable. Could violate the comparable accuracy principle.
Source: dos Santos-Silva (1999a), Lazcano-Ponce et al. (2001), Roberts and Norton (1995), Rothman and Greenland (1998), Stevenson et al. (1996), Wacholder, McLaughlin, et al. (1992), and Wacholder, Silverman, et al. (1992a).
sometimes be difficult to identify in the context of this design. For example, what would be a good control for an injured motorcyclist or car occupant? One solution is to select users of the same medical unit who are more comparable to cases with respect to quality of information because they also have been ill and hospitalized (Wacholder, Silverman, McLaughlin, & Mandel, 1992a). They are also the most convenient choice when controls will be asked to provide bodily fluids or to undergo a physical examination (Wacholder, et al., 1992a). Another strategy to identify suitable controls is that used by Wells et al. (2004). They obtained a random sample of motorcycle riding by identifying motorcyclists from 150 roadside survey sites (also randomly selected from a list of all nonresidential roads in the region under study).
Motorcyclists were photographed as they approached the survey site, stopped, and invited to participate in the study. Where survey sites or conditions were too dangerous for motorcyclists to be stopped, vehicles were photographed and followed up through their registration plate details. Although the authors reported that only 42 (3%) drivers refused to participate, this participation rate could be much smaller in other contexts (Wells et al., 2004). However, would this be the best solution to identify a control for an injured pedestrian? Here, neighborhood controls are even more recommended, especially if we consider that pedestrian injuries in some contexts occur near the place of residence 70% of the time (Fontaine & Gourlet, 1997; Muhlrad, 1998). In these cases, neighborhood controls represents an excellent option to sample
Chapter | 3
CaseeControl Studies in Traffic Psychology
controls. In this sense, after a case is identified, one or more controls who reside in the same neighborhood as that case are identified and recruited into the study. Such controls are matched to the cases on neighborhood. This constitutes a convenient substitute for population-based sampling of controls (Rothman & Greenland, 1998). This is relevant if we consider that population-based control studies tend to be more expensive, require more time, and also require a roster of all eligible subjects and families. In both cases, it is possible and, in fact, common that healthy people do not participate, which could introduce selection bias due to nonparticipation (dos Santos-Silva, 1999a). A neighborhood approach was used by Celis et al. (2003) to identify suitable controls to study pedestrian injuries in a sample of children 1e14 years old, although cases were identified through the attorney general’s office and emergency room registries. Upon leaving the house of each case, the interviewer knocked on the door of the house located immediately to the left and asked whether a child 1e14 years old lived there; if the answer was positive, authorization was requested to conduct the interview. If more than one child lived in the house, one of them was chosen randomly as the control. If there were no children living in the house, or permission was denied to conduct the interview, the next house to the left was approached in the same manner. If the cases are a representative sample of all cases in a precisely defined and identified population and the controls are sampled directly from this population, the study is said to be population based. If possible, this is the most desirable option (Rothman & Greenland, 1998). As previously noted for cases, it is important to give reasons why controls do not participate and, when possible, provide additional information about their sociodemographic characteristics (age, sex, etc.) (dos Santos-Silva, 1999a). For example, we consider a population-based caseecrossover and caseecontrol study of alcohol and the risk of injury (Vinson, Maclure, Reidinger, & Smith, 2003). Cases were injured patients recruited from emergency departments. Each case’s alcohol consumption in the 6 h prior to injury was compared to his or her consumption the day before in a caseecrossover analysis. Cases were recruited by telephone and matched to other cases by age, gender, day of week, and hour. Caseecontrol analyses examined recent alcohol consumption (past 6 h), hazardous drinking in the past month, and alcohol use disorders in the past year. Alcohol’s effect on injury risk was related more strongly to acute exposure than to measures of long-term exposure. The risk was significant even at low levels of consumption.
3.3.2. Matching The fundamental question concerning the selection of cases and controls is the following: What should be allowed to
35
vary as the hypothesized cause, or causes, in stratified samples, and what should be held constant? If the variables to be held constant can be other than randomly distributed between case and controls, the purpose of the design is defeated (Robertson, 1992). Matching is a control selection method that can sometimes improve efficiency in the estimation of the effect of exposure by protecting against the situation in which the distributions of a confounder are substantially different in cases and controls (Table 3.4; Wacholder, Silverman, McLaughlin, & Mandel, 1992b). Matching consists of selecting controls based on one or more characteristics of cases, such as sex, age, and socioeconomic status. This strategy increases statistical efficiency and tends to decrease bias associated with well-known confusion factors (Lazcano-Ponce et al., 2001). However, some authors state that the improvement is typically small, except for strong confounders (Wacholder, et al., 1992b). One example from the road traffic injury literature is the study published by Haddon et al. in 1961. They demonstrated the important role that alcohol plays in pedestrian injuries. They measured alcohol levels in fatally injured pedestrians and in randomly selected persons at the same places, walking at the same time of day, on same day of the week, and moving in the same direction as the fatally injured. Consequently, environmental factors were the same for the cases and the controls and did not account for differences in alcohol levels found in the cases and controls (Robertson, 1992). 3.3.2.1. Forms of Matching and Stratification There are two forms of matching: individual matching and frequency matching (Lazcano-Ponce et al., 2001). Individual matching refers to the selection of one or more controls who have exactly or approximately the same value of the matching factor as the corresponding case. The matching factor should not be the exposure under study. Frequency matching or quota matching results in equal distributions of the matching factors in the cases and the selected controls (Wacholder, et al., 1992b). Because cases and controls have similar matching factors, differences in health outcomes may be attributed to other factors (dos Santos-Silva, 1999a). 3.3.2.2. Disadvantages of Matching Matching has some disadvantages as well. In some cases, matching adds more costs and complexity to a sampling scheme by requiring extra effort to recruit controls. In addition, this strategy may result in the exclusion of cases when no matched control can be found, particularly when matching on several variables (Wacholder, et al., 1992b). Matching may also delay a study when cases have to be identified and complex matching variables have be
36
PART | I
Theories, Concepts, and Methods
TABLE 3.4 Reasons to Match Reason
Justification
Control of unmeasured confounders
Identifiable but not quantifiable variables with many categories, such as neighborhood or telephone exchange, can serve as proxies for environmental or socioeconomic confounding factors that are difficult to measure. Matching on such a variable may balance cases and controls with respect to unknown confounders.
Increase statistical power
Matching can ensure that there are sufficient controls to estimate an effect in a particular subgroup or to identify an interaction.
Time comparability
In unmatched studies, it can be difficult to achieve time comparability between cases and controls for exposures that vary over time.
Feasibility
Matching may be the most feasible method of obtaining controls.
Completeness of control for confounding
Perfect matching, followed by a matched analysis, results in complete control for a continuous confounder under a multiplicative model of the joint effects. Alternative strategies, such as regression adjustment for the confounder, can result in bias if its effect is misspecified (e.g., if linearity is wrongly assumed). Categorization may leave some residual confounding, but this is of little importance unless there is a substantial gradient in risk within strata.
Efficiency
Matching reduces the possibility of severe loss of efficiency due to a major discrepancy in the empiric distributions of a strong risk factor between cases and controls. Matching should be considered only for risk factors whose confounding effects need to be controlled for but that are not of scientific interest as independent risk factors in the study. Matching on variables that are unrelated to risk of disease is pointless; it can only reduce a study’s efficiency. Age, sex, and race are often used as matching variables because they are usually strong confounders and because their effects are usually well-known from descriptive epidemiology.
Source: Wacholder, Silverman, et al. (1992b).
obtained for cases and potential controls before control selection can be performed (Wacholder, et al., 1992b). In addition, matching may also present methodological problems. When controls are selected based on one characteristic that tends to hide the association between disease and the exposure of interest, the overmatching problem arises (dos Santos-Silva, 1999a). Matching on a factor that is a surrogate for or a consequence of disease or matching on a correlate of an imperfectly measured exposure can also lead to overmatching and bias (Wacholder, et al., 1992b). In general terms, “overmatching” refers to a matching that is counterproductive by either causing bias or reducing efficiency. Matching on an intermediate variable in a causal pathway between exposure and disease can bias a point estimate downward because the exposure’s effect on disease, adjusting for (conditional on) the intermediate variable, is less than the unadjusted effect (Wacholder, et al., 1992). Overmatching can occur even when matching per se was not used in the selection of controls, such as when an overly homogeneous population base is used for a specific study (Wacholder, et al., 1992b). Therefore, if the role of a variable is doubtful, the best strategy is not matching but, rather, adjusting its effect in statistical analysis (dos Santos-Silva, 1999a). Stratified or matched analyses can be considered even when there is no matching or stratification in the design. However, matching at the design stage reduces the investigator’s flexibility
during the analysis (Wacholder, et al., 1992b) because the effect of matching factors can no longer be studied (dos Santos-Silva, 1999a). Selection of people engaged in the same activity at the same site, time of day, day of week, etc. may not be possible for activities that occur at the case sites infrequently, such as use of “all-terrain” vehicles or snowmobiles. A child injured in a pedestrian collision may have no siblings close enough in age to serve as controls within the household, although children in reasonable proximity in the same neighborhoods may serve as controls depending on the factors of interest (Robertson, 1992).
3.3.3. Number of Controls 3.3.3.1. Ratio of Controls to Cases Determination of the number of controls is another important decision when designing a caseecontrol study. It is useful to consider the ratio of controls to cases. Wacholder, et al. (1992b) argue that there is usually little marginal increase in precision when the ratio of controls to cases is increased beyond four, except when the effect of exposure is large. In general, the best way to increase precision in a caseecontrol study is to increase the number of cases by widening the base geographically or temporally rather than by increasing the number of controls because the marginal increase in precision from an additional case is
Chapter | 3
CaseeControl Studies in Traffic Psychology
37
greater than that from an additional control (assuming there are already more controls than cases in the study). In matched and stratified studies, the most efficient allocation of a fixed number of controls into strata is usually one that sets the ratio of controls to cases to be approximately equal (Wacholder, et al., 1992b). 3.3.3.2. Number of Control Groups Some researchers have suggested choosing more than one control group when one of them has advantages that are missing from the other and vice versa (Rothman & Greenland, 1998). It certainly is reassuring when the results are concordant across control series. The problem is when results are discordant because investigators must decide which result is “correct” and essentially discard the other (Wacholder, et al., 1992b). Wacholder et al. suggest that usually the best strategy is to decide which control series is preferable at the design stage. However, multiple control groups might be helpful when each serves a different purpose, such as when each control group provides the ability to control for a particular confounder. In this situation, the second control group can act as a form of replication. 3.3.3.3. One Control Group for Several Diseases Use of a single control group for more than one case series can lead to savings of money and effort. Systematic errors in assembling the control series would presumably affect each individual series equally, but the availability of a larger number of controls would increase the precision of point estimates (Wacholder, et al., 1992b).
3.4. Analysis in CaseeControl Studies: Measures of AssociationeCausality Caseecontrol studies use the odds ratio (OR) as a measure to evaluate the strength of association between a factor (exposure) and the event (health condition or disease) under study. This measure indicates the relative frequency of exposure between cases and controls, as shown in Figure 3.2. The quotient of OR of exposure in cases and the OR of exposure in controls corresponds to the OR of exposure. In this type of study, the incidence of disease cannot be estimated both in exposed and in nonexposed individuals because they are selected based on the presence or absence of the condition under study and not by their exposure status (with the exception of some variants of caseecontrol studies, such as the nested caseecontrol, the caseecohort, and the caseecrossover designs). On the other hand, although relative risk is not directly calculated, when frequency of disease is low, OR is a nonbiased estimator of the incidence rate ratio or relative risk (Lazcano-Ponce et al., 2001).
FIGURE 3.2 Analysis of a nonmatched caseecontrol study. Source: Herna´ndez-A´vila and Lo´pez-Moreno (2007).
Odds ratio values oscillate between 0 and infinite. The OR obtained from a caseecontrol study indicates how many times higher (when OR is >1) or lower (when OR is 0.055) in the analysis of variance. In other words, simulator sickness affected performance in the emergency task. The authors concluded, “It is strongly recommended that researchers explore and control the potential confounding effects of simulator sickness to assure meaningful performance assessments” (p. 1092). Thus, discomfort can both prevent studies from being completed and affect the results obtained in driving simulators. Anecdotally, it can be stated that the rate of simulator sickness is reduced with a motion system, and especially with a large-scale motion system, but as far as we are aware, this has not been investigated systematically.
8. EXPERIMENTAL DESIGN No particular experimental design can be considered as standard, although within-subject designs have major advantages in terms of experimental power. However, they can also have disadvantages, both in terms of the time required for participants to experience all the required conditions and because repeated-measures designs tend to induce familiarity with the scenarios included in the experiment and may therefore make surprise events nonviable. Similarly, counterbalancing of conditions can be considered as the norm because of the ability to control for learning effects. But the disadvantage is that counterbalancing makes
95
it very difficult to investigate learning and ordering effects when these might be considered to be important. Thus, in experimental designs, as in other aspects of the setting up of simulator experiments, there is no right way and no wrong way. Experimenters should be guided by the research questions and hypotheses that they wish to address, and they should carefully weigh the advantages and disadvantages of alternative designs. Another experimental design issue relates to the amount of control over scenarios. There is a strong impetus to create scenarios that are equal in severity for all participants. Thus, it may be considered desirable to have a carfollowing scenario in which the lead vehicle is controlled in terms of a give time headway to the driven car. Then an event such as a sudden braking of the lead vehicle can be triggered such that all participants have to respond to an event of equal severity. However, participants cannot be forced to drive at a given speed (unless speed control is automated), and a given participant may find that the chosen time headway is too close for comfort. The participant will react by slowing down, the preceding vehicle will come closer, the participant will slow down more, and so on. This phenomenon of participants trying to “override” the scenario design has been observed in the University of Leeds Driving Simulator. It is also discussed by Donmez, Boyle, and Lee (2008), who carried out an analysis of such a scenario using the inverse of actual headway distance (rather than time headway, which was preset) at the time of accelerator release as a covariate. The finding from this analysis was that the experimental results changed depending on whether the covariate was taken into account: Without consideration of the covariate, distraction of various types appeared to improve reaction time, but once the covariate was considered, it was found that distraction resulted in longer reaction times.
9. CONCLUSIONS Simulators provide the opportunity to investigate driving under controlled conditions in a manner that is unparalleled by the alternatives. Real-world studies lack the equivalent control element, whereas test tracks offer a very depleted and inflexible driving environment. Simulator capability, particularly in terms of the graphics performance of PCbased systems, has grown very fast in recent years, and the advent of small-scale and relatively low-cost motion systems means that it may soon become standard for a midrange simulator to be equipped with six degrees of freedom of motion. The number of research simulators worldwide continues to increase, and simulator studies constitute an increasing proportion of the research literature on driving performance and behavior. Simulators may not be total replicates of the real world, and indeed they cannot be. But they offer the researcher of driver behavior an
96
advantage that real-world studies cannot match: the ability to control experimental conditions and create prescripted scenarios.
REFERENCES Alm, H. (1995). Driving simulators as research tools: A validation study based on the VTI driving simulator. In GEM validation studies: Appendix. DRIVE II Project V2065 GEM. Bittner, A. C., Gore, B. F., & Hooey, B. L. (1997). Meaningful assessments of simulator performance and sickness: Can’t have one without the other. In Proceedings of the Human Factors and Ergonomics Society 41st annual meeting (pp. 1089e1093). Blana, E. (2001). The behavioral validation of driving simulators as research tools: A case study based on the Leeds Driving Simulator. PhD dissertation. Institute for Transport Studies. Leeds, UK: University of Leeds. Blana, E., & Golias, J. (2002). Differences between vehicle lateral displacement on the road and in a fixed-base simulator. Human Factors, 44, 303e313. Blauw, G. J. (1982). Driving experience and task demands in simulator and instrumented car: A validation study. Human Factors, 24, 473e486. Dagdelen, M., Reymond, G., Kemeny, A., Bordier, M., & Maızi, N. (2009). Model-based predictive motion cueing strategy for vehicle driving simulators. Control Engineering Practice, 17(9), 995e1003. De Waard, D., & Brookhuis, K. A. (1997). Behavioral adaptation of drivers to warning and tutoring messages: Results from an on-theroad and simulator test. International Journal of Vehicle Design, 4, 222e235. Donmez, B., Boyle, L. N., & Lee, J. D. (2008). Accounting for timedependent covariates in driving simulator studies. Theoretical Issues in Ergonomics Science, 9(3), 189e199. ¨ stlund, J. (2005). Effects of visual and Engstro¨m, J., Johansson, E., & O cognitive load in real and simulated motorway driving. Transportation Research Part F: Traffic Psychology and Behavior, 8(2), 97e120. Evans, L. (1991). Traffic safety and the driver. New York: Van Nostrand Reinhold. Evans, L. (2004). Traffic safety. Bloomfield Hills, MI: Science Serving Society. Greenberg, J., Artz, B., & Cathey, L. (2003). The effect of lateral motion cues during simulated driving. In Proceedings of the Driving Simulator Conference North America, Dearborn, 8e10 October. Irving, A., & Jones, W. (1992). Methods for testing impairment of driving due to drugs. European Journal of Clinical Psychology, 43, 61e66. Jamson, A. H., Lai, F. C. H., & Carsten, O. M. J. (2008). Potential benefits of an adaptive forward collision warning system. Transportation Research Part C: Emerging Technologies, 16(4), 471e484. Jamson, A. H., Whiffin, P. G., & Burchill, P. M. (2007). Driver response to controllable failures of fixed and variable gain steering. International Journal of Vehicle Design, 45(3), 361e378. Jamson, S., Lai, F., Jamson, H., Horrobin, A., & Carsten, O. (2008). Interaction between speed choice and road environment (Road Safety Research Report No. 100). London: Department for Transport.
PART | I
Theories, Concepts, and Methods
Kaptein, N. A., Theeuwes, J., & van der Horst, R. (1996). Driving simulator validity: Some considerations. Transportation Research Record, 1550, 30e36. McGehee, D. V., & Carsten, O. M. (2010). Perception and biodynamics in unalerted precrash response. Annals of Advances in Automotive Medicine, 54, 315e332. Mortimer, R. G. (1963). Effect of low blood-alcohol concentrations in simulated day and night driving. Perceptual and Motor Skills, 17, 399e408. Nahon, M. A., & Reid, L. D. (1990). Simulator motion-drive algorithms: A designer’s perspective. AIAA Journal of Guidance, Control and Dynamics, 13(2), 356e362. Olson, R. L., Hanowski, R. J., Hickman, J. S., & Bocanegra, J. (2009). Driver distraction in commercial vehicle operations (Report no. FMCSA-RRR-09-042). Washington, DC: Federal Motor Carrier Safety Administration, U.S. Department of Transportation. Rizzo, M., McGehee, D., Dawson, J., & Anderson, S. (2001). Simulated car crashes at intersections in drivers with Alzheimer disease. Alzheimer Disease and Associated Disorders, 15, 10e20. Santos, J., Merat, N., Mouta, S., Brookhuis, K., & de Waard, D. (2005). The interaction between driving and in-vehicle information systems: Comparison of results from laboratory, simulator and real-world studies. Transportation Research Part F: Traffic Psychology and Behavior, 8(2), 135e146. Schnabel, E., Hargutt, V., & Kru¨ger, H.-P. (2010). Meta-analysis of empirical studies concerning the effects of alcohol on safe driving (Deliverable D 1.1.2a of DRUID (Driving under the Influence of Drugs, Alcohol and Medicines)). Germany: University of Wu¨rzburg. Wu¨rzburg. Segel, L. (1956). Theoretical prediction and experimental substantiation of the response of the automobile to steering control. In Proceedings of the Institution of Mechanical Engineers Automobile Division (pp. 310e330). Slob, J. J. (2008). State-of-the-art driving simulators, a literature survey (DCT Report No. 2008.107). Department of Mechanical Engineering, Eindhoven University of Technology. In: Eindhoven. The Netherlands: Control Systems Technology Group. Transportation Research Board. (1995). Estimating demand for the National Advanced Driving Simulator. Washington, DC: Author. Virginia Tech Transportation Institute. (2009, July 27). New data from VTTI provides insight into cell phone use and driving distraction. Blacksburg, VA: Author. [Press release]. Ward, N. J., & Dye, L. (1999). Cannabis and driving: A review of the literature and commentary (Road Safety Research Report No. 12). London: Department of Environment, Transport and the Regions. Wang, Y., Mehler, B., Reimer, B., Lammers, V., D’Ambrosio, L. A., & Coughlin, J. F. (2010). The validity of driving simulation for assessing differences between in-vehicle informational interfaces: A comparison with field testing. Ergonomics, 53(3), 404e420. Weir, D. H., & Clarke, A. J. (1995). A survey of mid-level driving simulators (SAE Technical Paper No. 950172) Society of Automotive Engineers. Warrendale, PA. Wierwille, W. W., & Fung, P. P. (1975). Comparison of computergenerated and simulated motion picture displays in a driving simulation. Human Factors, 17(6), 577e590.
Chapter 8
Crash Data Sets and Analysis Young-Jun Kweon Virginia Department of Transportation, Charlottesville, VA, USA
1. INTRODUCTION The focus of this chapter is traffic safety, not transportation safety, which means crashes involving modes using roads are of interest and not those involving other modes, such as air, rail, and marine. Subjects of interest for analysis vary depending on study purposes and design. For example, our interests might be persons (e.g., drivers), vehicles (e.g., trucks), facilities (e.g., signalized intersections), or geographical areas (e.g., cities). The choice of subject for analysis often dictates the level of data aggregation affecting types and formats of data suitable for analysis, which in turn affects analysis methods. For example, if we want to compare the traffic safety situation among cities, the annual number of fatal crashes for each city could be obtained by counting all fatal crashes that occurred for each city in a certain year. In this chapter, data useful for traffic safety analysis are introduced, and typical methods for analyzing such data are described. However, this chapter is not intended to be exhaustive with regard to traffic safety data and analysis but, rather, introduces the most frequently used data sources and analysis methods for traffic safety studies.
2. DATA Two types of data are often used in traffic safety analysis: traffic safety data and supplement data. Traffic safety data can be classified into three types: police crash data, medical crash data, and safety survey data. Use of traffic safety data alone might mislead data analysts in attempts to understand factors that contribute to occurrences and outcomes of crashes. For example, we cannot make a fair comparison of the traffic safety situation between two cities based only on the total annual number of traffic crashes if the two cities are quite different in size. In such a case, information reflecting the size of the cities, such as population, road mileage, and/or registered vehicles of the cities, should also be incorporated. Such information does not usually exist in traffic safety data but can be obtained from other data sources. Because those data supplement traffic safety data Handbook of Traffic Psychology. DOI: 10.1016/B978-0-12-381984-0.10008-6 Copyright Ó 2011 Elsevier Inc. All rights reserved.
so that we can understand factors that contribute to occurrences and outcomes of crashes in a more complete manner, they are called supplement data here. Various kinds of data can serve as supplement data for traffic safety analysis, including roadway and traffic data, license and registration data, travel survey data, and sociodemographic/ economic data. Use of supplement data for traffic safety analysis helps us avoid biases that might exist in traffic safety data or might be introduced when only traffic safety data are used for analysis, and it also helps us better identify and understand factors contributing to occurrences and outcomes of crashes.
2.1. Traffic Safety Data Three types of traffic safety data are police crash data, medical crash data, and traffic safety survey data. Police crash data are the most frequently used data for traffic safety studies.
2.1.1. Police Crash Data Crash data used for traffic safety analysis are typically obtained from police crash reports; thus, they are called police crash data. These data contain the most important information regarding traffic crashesdinformation on a crash and persons and vehicles involved in the crash. Regulation or statute of each jurisdiction establishes a threshold of property damage or injury severity of a motor vehicle crash to be reported to the police; for instance, many jurisdictions in the United States require a report of any crash on a public road sustaining a minimum of $1,000 to $1,500 in property damage or any level of injury severity including fatality. Although crash reports are completed by police officers in most jurisdictions, a state highway agency or vehicle/driver registration agency, not a police agency, typically has custodial responsibility for statewide crash databases containing crash data provided by local and state police departments in the United States. To perform meaningful analysis, it is necessary to extract a uniform data set from a crash database. However, 97
98
PART | I
Theories, Concepts, and Methods
it is possible that different jurisdictions use different definitions and criteria and record different sets of information on crash reports. Thus, caution should be exercised when police crash data from different jurisdictions are used for analysis. In the United States, there have been various efforts to establish uniformity in the fields and values on crash report forms so that crash data across jurisdictions can be shared and analyzed comparably. The American National Safety Institute’s D-16.1 Manual on Classification of Motor Vehicle Accidents has been a source for consistent data definitions on crash data, and the Model Minimum Uniform Crash Criteria (MMUCC) has been a source for a minimum set of data elements to conduct meaningful analysis. The Fatality Analysis Reporting System (FARS) promotes collecting data on all crashes involving a fatality according to a shared standard so that consistent data analysis on fatal crashes can be performed.
(e.g., sequential unique number within each crash), unit type (e.g., vehicle, pedestrian, or bicyclist), vehicle type (e.g., pickup truck and bus), make/model and year, vehicle movement and maneuver prior to the crash (e.g., going straight), and an estimate of property damage.
2.1.1.1. Data Elements Several basic data elements are commonly found in police crash data, five of which are described here.
2.1.1.1.5. Injury Severity All jurisdictions use some method to quantify the level of injury severity; at a minimum, two levels are useddfatal and nonfatal injury. The five-level KABCO (killed, A, B, C injury, and property damage only) severity scale is frequently used to record the severity of injuries to each person on police crash reports. K and O are for fatal and no injury, respectively, and A, B, and C are three levels of injurydincapacitating, capacitating (or non-incapacitating), and possible (or minor) injury, respectively. An alternative to the KABCO scale is the Abbreviated Injury Scale, which is more detailed and generally matches better with codes in medical data systems than the KABCO scale. The police record the injury severity for each person involved in a crash based on visual inspection. Because most jurisdictions require information to be recorded only for injured persons, information on persons who are involved in a crash but not injured based on visual inspection is not found in police crash reports. In the United States, most jurisdictions adopt a definition of fatality as death from injuries resulting from a traffic crash within 30 days of the crash, which is consistent with national crash data systems such as FARS.
2.1.1.1.1. Location Information The location where a crash occurs is critical information for traffic safety analysis. The police typically record the location information of a crash using street names and an estimated distance from a physical marker, such as an intersection, bridge, or milepost. Great effort and resources are required when using this text-based location record and the estimated distance on crash reports to identify crashes systematically, and even if crash location is identified, the accuracy of the location may be in question. 2.1.1.1.2. Environment Information Environment in crash data refers to the conditions that are the same for all vehicles and persons involved in the crash. Environment information includes crash report number (a unique identifier), location (e.g., city and street name), time and date, weather condition (e.g., rain or snow), lighting condition (e.g., dusk or dark), surface condition (e.g., wet or icy), and crash type (e.g., head-on or rear-end). Other useful environment information includes work zone indicator and first and most harmful events. 2.1.1.1.3. Vehicle Information A single crash can involve more than one vehicle; thus, a record containing environment information can be linked to several vehicle (or unit) records. “Vehicle” is somewhat deceiving in crash records because typically pedestrian, bicyclist, and vehicles are viewed as a “unit” involved in the crash and a separate vehicle record is created for each unit. Typical vehicle information includes crash report number (the same as that for environment information), vehicle/unit number
2.1.1.1.4. Driver Information Driver information describes the person operating the vehicle or possibly a pedestrian or bicyclist involved in the crash. Information that can be found on or inferred from a driver’s license, such as gender and age, is recorded on police crash reports. Alcohol involvement is an important piece of information recorded on crash reports, and in some cases, blood alcohol content (BAC) test results are entered into the crash data system. However, not all drinking drivers are tested, and not all BAC test results are submitted for entry into the crash data system.
2.1.1.2. National Databases Three national databases based on police crash reports in the United States are frequently used for traffic safety studies: FARS, General Estimate System (GES), and Highway Safety Information System (HSIS). These databases have been used for various traffic safety studies, ranging from policy studies to engineering studies. 2.1.1.2.1. FARS FARS is a census system of traffic crashes that have resulted in the death of a person involved in a crash within 30 days of the crash, and it contains all fatal crashes that occurred in the 50 states, the District of
Chapter | 8
Crash Data Sets and Analysis
Columbia, and Puerto Rico. Each crash record has more than 100 coded data elements characterizing a fatal crash and persons and vehicles involved in the crash. However, to protect privacy, personal information such as name, address, or specific crash location is not included. FARS data are available for every year since 1975 and have three principal filesdthe accident, vehicle, and person files. The accident file contains environment information such as the time and location of the crash, collision type, roadway alignment, the number of vehicles and persons involved, the first harmful event, and weather and surface conditions. The vehicle file contains vehicle and driver information, such as vehicle type, crash avoidance maneuver, height and weight of a driver, initial and principal impact points, and the most harmful event. The person file contains data on each person involved in the crash, such as age, gender, person type (e.g., driver, occupant, pedestrian, and bicyclist), seating position, injury severity, and restraint use. Some information is found in more than one file, such as hit-and-run status in the accident and vehicle files and the first harmful event in all three files. 2.1.1.2.2. GES National Automotive Sampling System GES is a system of sample traffic crashes involving all levels of severity ranging from no injury (i.e., property damage only) to fatal injury, and it contains crashes reported by approximately 400 police agencies in 60 geographic sites throughout the United States. Each crash record has more than 100 coded data elements, and injury severity is coded using the KABCO scale. GES data are available for every year since 1988, and GES, like FARS, has three principal filesdthe accident, vehicle, and person files. The three files of GES are similar to those of FARS in terms of included information. Since 2000, the event file has also been available in GES data. This file contains a brief description of each harmful event in a crash. There have been efforts to unify the two national systems, FARS and GES, with regard to definitions, entry, and analysis of data. To bring the two systems in alignment with the MMUCC, the first of a three-phase standardization process was implemented in 2009 by unifying 45 data elements; the second phase in 2010 unified more data elements and produced one coding manual; and the final phase scheduled to be implemented in 2011 will produce one data entry system for both systems. 2.1.1.2.3. HSIS The HSIS is a multistate database merging crash, roadway, and traffic data that are processed from databases of nine states selected based on the data quality and the ability to merge data from various files. Each state provides different sets of data in different levels of details. For example, the Illinois data system includes four basic files (accident, road log, bridge, and railroad
99
grade crossing), whereas the Minnesota data system includes seven basic files (accident, road log, reference post, traffic, intersection, bridge, and railroad grade crossing).
2.1.2. Medical Crash Data Medical data on any person injured for any reason, including persons injured in traffic crashes, are recorded in a system, and the collection, storage, and analysis of medical transportation, treatment, and outcome data of injured persons are referred to as the Injury Surveillance System (ISS). The ISS does not refer to a single database but, rather, a system that is used to track causes, extent, treatment, and recovery from injuries. ISS data are usually available only to health-related agencies, such as state departments of health and hospital associations, but aggregate statistics of such data are often made available for use by outside agencies. Among several databases in the ISS, only a few are useful and available for traffic safety studies. In the United States, the National Highway Traffic Safety Administration (NHTSA) recommends three of the databases for use: (1) emergency medical service (EMS) run report database, (2) trauma registry database, and (3) hospital discharge database. Although ISS data provide detailed information on the extent of injuries and long-term consequences of traffic crashes, issues concerning accuracy and completeness of the data exist. 2.1.2.1. EMS Run Report Database The EMS run report database includes records of injured persons who received immediate care and transportation from EMS providers. Typically, EMS providers are required to submit a report on each run to state agencies such as the department of health. The EMS run reports are not standardized in general, even within a jurisdiction, and the National Emergency Medical Systems Information System (NEMSIS) standard was designed to promote the consistent collection of comprehensive data elements throughout the United States. The NEMSIS standard defines data elements in two sets: (1) service information on the EMS run, treatments, and charges and (2) demographic information on the EMS provider, personnel, and equipment. Most of the states agreed to adopt the standard for their statewide reporting systems. 2.1.2.2. Trauma Registry Database The trauma registry database includes records of persons receiving treatment of trauma cases typically at designated trauma centers. The American College of Surgeons certifies trauma centers and sets guidelines for data elements collected on trauma cases. States certifying trauma centers usually establish a statewide trauma registry database to
100
combine registry data on all trauma cases treated at emergency departments of hospitals and designated trauma centers. 2.1.2.3. Hospital Discharge Database The hospital discharge database includes records of inpatients and outpatients admitted to a hospital and contains information on reasons for hospital visits, treatment codes, charges, and a code indicating involvement in traffic crashes. The Universal Billing Code of 1992 (UB-92) is a standard database for generating itemized medical charges at hospitals, and its data requirements serve as the standard for the hospital discharge database. The state hospital associations and the hospitals treating the patients maintain the database. 2.1.2.4. CODES Database Under privacy laws, much of the medical data are strictly available only to authorized personnel of health-related agencies. Even sanitized dataddata for which any personal identifying information, such as name and address, is removeddare not available for use by outside agencies because there is a possibility that a specific individual could be identified when the medical data are combined with other sources of data, such as police crash reports. This use restriction makes it very difficult to link records in the ISS to the police crash data, significantly lessening the usefulness of the medical crash data for traffic safety studies. Because medical data are highly useful for traffic safety studies, the NHTSA developed the Crash Outcomes Data Evaluation System (CODES) for linking the police crash database and the medical crash database. CODES is a probabilistic method matching records in the two databases without using the personal identifying information. The CODES database can be used to produce aggregate statistics so that a specific individual in the database cannot be traced, and it has been proven useful for traffic safety analysis at state and national levels, especially for analyzing outcomes and costs of traffic crashes.
2.1.3. Traffic Safety Survey Data Examples of national traffic safety surveys in the United States are the Motor Vehicle Occupant Safety Survey (MVOSS), National Survey of Drinking and Driving Attitudes and Behavior (DDAB), and the National Occupant Protection Use Survey (NOPUS). NHTSA has periodically conducted national telephone surveys for MVOSS since 1994 and for DDAB since 1991 to understand the public’s attitudes, knowledge, and self-reported behaviors related to restraint use and drinking and driving. MVOSS data contain survey responses regarding occupant protection issues from approximately 6000 respondents, whereas
PART | I
Theories, Concepts, and Methods
DDAB data contain survey responses regarding drinking and driving issues from approximately 7000 respondents. Whereas MVOSS and DDAB are telephone survey data, NOPUS is probability-based observed data on the use of different types of restraints (e.g., front shoulder belts, child restraints, and motorcycle helmets) and driver electronic device use. NOPUS is the only survey that provides nationwide observed data regarding occupant’s restraint use and driver’s electronic device use, and it has been conducted since 2000. NOPUS is composed of two sets of surveysda moving traffic survey and a controlled intersection surveyd and in 2009 its data were collected from approximately 1800 sites throughout the country. There are also other nationwide surveys, such as the American Automobile Association Foundation for Traffic Safety’s telephone survey (Traffic Safety Culture Index), and many local and state-level surveys.
2.2. Supplement Data Traffic safety data alone may not provide a correct view of the traffic safety situation and sometimes might disguise or even distort it. To understand the traffic safety situation more completely, information other than that found in the traffic safety data may be required. Understanding of how traffic safety is affected by various factors such as geometric features of roads, traffic control types, and traffic volume is critical for making informed decisions regarding safety improvements, and these factors are not typically found in the traffic safety data. For example, when we compare the traffic safety situation of two cities with different sizes, we need to obtain not only their crash data (e.g., number of traffic deaths) but also information reflecting their sizes (e.g., number of registered drivers). Among various data that could supplement traffic safety data, four popular sources are roadway and traffic data, license and registration data, travel survey data, and sociodemographic/economic data.
2.2.1. Roadway and Traffic Data It is crucial to know the location where a crash occurs and any roadway and traffic characteristics that may have contributed to the crash. Roadway data help to identify the physical and use characteristics (e.g., horizontal curve and functional classification) of the location that may have contributed to the occurrence or severity of the crash. Traffic data are used to control for use intensity of the location (e.g., traffic volume) and to calculate crash rates (e.g., number of crashes per 100 million, vehicle miles traveled). Highway agencies at the local (e.g., city or county department of public works) and the state (e.g., state department of transportation or highway administration) levels maintain roadway and traffic data on roadways and intersections under their administration.
Chapter | 8
Crash Data Sets and Analysis
Although some guidance on roadway and traffic data elements exist, such as the Highway Performance Monitoring System (HPMS), they are not designed for traffic safety analysis. Model Inventory of Roadway Elements (MIRE) is a data standard for roadway and operations data elements that are critical for traffic safety analysis, and it is anticipated to serve as a companion to the MMUCC. 2.2.1.1. Roadway Inventory Database The roadway inventory database contains physical and use characteristics of a roadway segment and an intersection. The database includes lanes (e.g., the number of lanes and width), shoulders (e.g., type and width), medians (e.g., type and width), access control (e.g., full or partial control), functional classification (e.g., rural/urban interstate highways), speed limits, junction type (e.g., T intersection and ramp), and surface type (e.g., bituminous or Portland cement concrete). There are large variations with regard to elements and the accuracy of the database across jurisdictions. A roadway segment in the database is defined as a stretch of a roadway whose characteristics (e.g., functional classification, speed limit, number of lanes, and shoulder width) are identical, called a homogeneous segment. 2.2.1.2. Traffic Database The traffic database contains traffic volume and characteristics (e.g., traffic counts by vehicle type) of the roadway and can typically be linked to the roadway inventory database. The database includes annual average daily traffic (AADT), speeds, seasonal and directional adjustment factors, and percentage of trucks and buses. 2.2.1.3. Highway Performance Monitoring System The HPMS is a national highway system database containing data on the extent, condition, performance, use, and operating characteristics of highways in the United States, and it supports a data-driven decision process regarding national highway issues. HPMS data are used for assessing performance and investment needs of highway systems and for apportioning federal highway funds. Although the HPMS is not designed specifically for traffic safety analysis, its data contain information useful for traffic safety analysis, covering various aspects of highway characteristics such as roadway inventory (e.g., facility type, turn lanes, and speed limit), traffic operations and controls (e.g., AADT by vehicle type and signals and stop signs), geometric features (e.g., lane width, median type, and grade), and pavement (e.g., surface type, rutting, and base type).
101
departments of motor vehicles or departments of public health. Although the raw data are not typically accessible by outside agencies, aggregate statistics of the data can be made available for use by outside agencies on a regular basis (e.g., annual). For example, percentages of teen and elderly drivers for each jurisdiction can be released from the registration agency and are considered to be important for explaining changes in the traffic safety situation over years or differences across jurisdictions.
2.2.3. Travel Survey Data Individual or household travel survey data for transportation planning purposes are collected periodically or need-based nationally or locally. In the United States, the Nationwide Household Transportation Survey (NHTS) collects such data every 5e7 years, and this database contains various travel-related information that is helpful in understanding factors contributing to occurrences and outcomes of traffic crashes. For example, trip purpose (e.g., work and shopping), travel mode (e.g., car and bus), travel time, and travel time of day and day of week are recorded in the survey. It is not possible to link records of individuals in the survey to records in the crash database. However, we could use aggregate forms of the survey data such as summary statistics for traffic safety studies. For example, annual crash rates can be calculated by combining GES and NHTS (Kweon & Kockelman, 2003). Specifically, crash counts have been estimated from GES, and vehicle miles traveled (VMT) have been estimated from NHTS for groups of drivers defined by age, gender, and vehicle type. The annual crash rates have been calculated by dividing the crash counts by the VMT for different crash types and injury severities.
2.2.4. Sociodemographic/Economic Data Sociodemographic and economic data are collected and made available periodically in an aggregate form. Examples of such data include population, age distribution, education level distribution, and unemployment rate. Such data are helpful in understanding changes in the traffic safety situation of an area over years and differences across areas. For example, when analyzing state-level data, percentages of rural VMT, poverty, interstate highway lane mile and seat belt use, and consumption of beer and wine were found to be useful in predicting traffic fatality rates (Kweon, 2007).
3. DATA ANALYSIS 2.2.2. License and Registration Data Driver’s license and vehicle registration data are maintained by licensing and registration agencies such as state
Data analysis for traffic safety studies refers to the use of data from one or more data sources to describe a traffic safety situation and to understand factors contributing to
102
occurrences and outcomes of crashes using numbers. Data analysis is neither simply extracting and summarizing numbers from the data nor simply creating tables and figures using the data. Data analysis is telling a story embedded in the data and is interpreting the data reliably, validly, and responsibly. Therefore, as data analysts, we are interpreters of the data and storytellers about the data that we analyze. In this respect, we should be able to relate crashes to factors that might have contributed to occurrences and outcomes of the crashes. The ultimate goal of crash data analysis is to reduce occurrences and improve outcomes of crashes. Data analysis can be classified into two types depending on the level of data aggregation: aggregate and disaggregate analyses. Aggregate data analysis is an analysis using numbers aggregated from information on individual records, whereas disaggregate data analysis is one directly using information on individual records. Aggregation can be done temporally and/or spatially. For example, if we are interested in comparing cities with regard to crash occurrences, we could obtain the total number of crashes that occurred in each city in a certain year by counting individual crash records of each city (i.e., spatial aggregation) in that year (i.e., temporal aggregation). Because this uses aggregated forms of individual data records, it is called aggregate data analysis. If we are interested in identifying personal characteristics likely contributing to outcomes of crashes, we need to use individual records of drivers in police crash data (e.g., age, gender, and violation history) and link them to outcomes of crashes. Because this uses individual data records, it is called disaggregate data analysis. Regardless of analysis type, data analysis is performed primarily to describe the traffic safety situation using numbers and to understand contributing factors. This discussion focuses on aggregate data analysis.
3.1. Aggregate Data Analysis Crash frequencies and crash rates are typical data forms for aggregate data analysis, and crash severity can be incorporated in the process of producing these data. Aggregating individual crash data records suitable for the design and purposes of an analysis is the first step in aggregate data analysis and involves a combination of four basic tasks.
3.1.1. Four Basic Tasks Individual crash data records should be aggregated in a suitable and meaningful form for aggregate data analysis, and data aggregation can be performed through a combination of the four basic tasks: selection, aggregation, integration, and normalization. These tasks are performed not only for aggregate data analysis but also for disaggregate data analysis. The selection task is almost always
PART | I
Theories, Concepts, and Methods
performed for any kind of traffic safety analysis, both aggregate and disaggregate levels, whereas the other three tasks might be performed depending on the design and purpose of an analysis. Thus, understanding these tasks is crucial for preparing appropriate and valid data for analysis. 3.1.1.1. Selection Crashes appropriate for the design and purpose of an analysis should first be selected from the crash database. This process is also called “subsetting” or “filtering.” Selection of crash records can be made based on various factors, such as location/area, type of person/vehicle involved in a crash, weather conditions, or a combination of these factors. For example, if we are interested in assessing the crash risk of a driver in a sport utility vehicle (SUV) in rainy conditions, all crashes involving SUVs that occurred in rain should be extracted from the crash database. If we are interested in comparing injury severity of crashes among different types of roadway (e.g., interstate highway, arterial, and collector), crashes should be extracted separately by roadway type. 3.1.1.2. Aggregation Once crashes of interest are selected, we can aggregate those crashes to generate summary statistics in various ways. Using the example of the crash risk of a driver in an SUV in rainy conditions, if we are interested in comparing the crash risk of such a driver across different ages, the selected crash records should be aggregated by age group (e.g., young, middle-aged, and elderly) to generate the total number of such crashes for each age group. Note that factors used as criteria for aggregation should be categorical in nature. If the values of these factors are continuous (e.g., age of a driver) or categorical but contain many unique values (e.g., speed limits ranging from 15 to 75 mph by 5-mph increments, resulting in 13 unique values), these factors can be recoded so as to have a manageable number of categories (e.g., young, middle-aged, and elderly age groups recoded from continuous age values). 3.1.1.3. Integration To understand factors contributing to crashes in a more complete manner, it is often necessary to incorporate information that is not found in traffic safety data, including roadway characteristics (e.g., functional classification, speed limit, lane width, and junction type) and driver and vehicle registrations (e.g., the number of registered drivers and vehicles in a city). For example, it might be necessary to know the type of roadway on which a driver was traveling (e.g., four-lane rural interstate highway) at the time of a crash to properly understand how personal characteristics (e.g., age and gender) affected the occurrence and outcome
Chapter | 8
Crash Data Sets and Analysis
of the crash. In the example of the crash risk of a driver in an SUV in rainy conditions, the crash risk might differ by roadway characteristics (e.g., speed limit) due to the different performance of the SUV on different types of roads (e.g., high-speed vs. low-speed roads). Thus, the roadway inventory database should be integrated with the crash database. In this example, integration should be completed prior to aggregation because individual crash records should be matched with roadway inventory records before aggregation. 3.1.1.4. Normalization Normalization is typically performed to obtain a crash rate that supports valid comparisons among groups of people (e.g., age groups), vehicle types, roadway types, and so on. In the example comparing two cities in terms of the crash risk of SUVs, the number of crashes involving SUVs and the number of registered SUVs for each city in a certain year are obtained through aggregation. Dividing the number of SUVinvolved crashes by the number of registered SUVs for each city gives a crash rate of SUVs so that the SUV safety situation of the two cities can be evaluated in a fair manner by comparing the SUV crash rates between the two cities.
3.1.2. Example Suppose we would like to compare two cities with regard to the traffic safety situation for senior drivers. One way of doing so is to compare the traffic fatality rate per registered drivers for senior drivers in a specific year. Thus, we need to calculate the traffic fatality rate per drivers for senior drivers. First, the selection task is performed. For the numerator of the rate (the number of fatal senior drivers), records of crashes that occurred in the two cities in the specific year are extracted from the crash database. Among the records, drivers who were older than 64 years (here, a senior driver is defined as one who is older than 64 years) and were recorded “fatal” or “killed” in the database are selected. For the denominator of the rate (the number of registered senior drivers), records of drivers who were registered in the two cities and were older than 64 years during that year are extracted from the driver’s license database. Second, the aggregation task is performed. For the numerator, counting those drivers in the selected crashes for each city produces the annual number of senior fatalities in traffic crashes. For the denominator, counting the selected license records for each city produces the number of registered senior drivers. Third, the normalization task is performed. Simply dividing the number of fatal senior drivers by the number of registered senior drivers for each city produces the traffic fatality rate per registered driver for senior drivers. Senior drivers’ fatality risk in a traffic crash in the two cities can
103
now be assessed by comparing the calculated crash rates between the two cities. A couple of points are noteworthy in this example. First, in the aggregation task, for the numerator, we should count drivers in the selected crash records, not the crashes, because there might be crashes involving multiple vehicles resulting in more than one fatal senior driver. Counting the crashes would probably lead to an underestimation of the rate. Second, we may not be able to directly access the crash and/or driver’s license database. Thus, we cannot perform the selection and aggregation tasks on our own but, rather, must request aggregated statistics from agencies managing those databases.
3.1.3. Basic Aggregate Data Analysis Three basic ways of analyzing aggregate data are frequency, rate, and trend analysis. 3.1.3.1. Frequency Analysis A frequency in traffic safety is a tally of crashes, vehicles, or victims, and the traffic safety situation can be quantified by simply counting them, grouped by factors such as location/area (e.g., city or state), group of people (e.g., males age 35 years or older), vehicle type, time period (e.g., holiday weekend night), or a combination of these. Injury severity can be incorporated in the frequency analysis by counting crashes, vehicles, or victims by injury severity sustained. For example, the annual total number of crashes that occur on rural interstate highways during weekends can be obtained for each of three severity levelsdfatal, injurious, and property damage only. 3.1.3.2. Rate Analysis A rate in traffic safety is obtained by dividing the frequency by a normalizing factor that is typically known as a crash exposure measure. Rate analysis is most popular among practitioners because it is straightforward to calculate and easy to understand. Examples of the rate include the fatal crash rate per million vehicles entering into an intersection and the crash rate per 10000 registered drivers in a city. The normalizing factor is intended to support valid comparisons in the traffic safety situation among different groups formed by various factors, such as age, vehicle type, and geographical area. For example, if two roads have the same crash frequency and similar geometric features, the highway with higher traffic volume is deemed safer. Thus, the crash frequencies of the two roads should be normalized by traffic volume (e.g., AADT) so that the safety comparison between the two roads can be made using the resulting crash rate. Examples of frequently used normalizing factors (i.e., exposure measures) include VMT, population, and the numbers of registered drivers and vehicles. VMT is
104
typically obtained from a state highway agency’s database, population from census data, and the numbers of registered drivers and vehicles from a state driver/vehicle registration agency’s database. The normalizing factor (denominator) can be the number in the same kind (e.g., same severity level or same crash type) as the frequency (numerator), and dividing the frequency by such a normalizing factor produces a ratio or proportion, not a rate. For example, dividing the number of fatalities by the number of fatal crashes results in the fatality ratio, and dividing the number of single-vehicle rollover crashes by the total number of single-vehicle crashes results in the single-vehicle rollover proportion (or percentage when being multiplied by 100). However, the ratio, proportion, and percentage can be viewed as the rate in a broader sense. Thus, rate analysis is used to analyze them. For rate analysis, it is best for the numerator and the denominator to be in the same unit (e.g., person, vehicle, crash, or roadway) because it is easier to explain the resulting rate. For example, the number of fatalities divided by population is easier to understand than the number of fatalities divided by registered vehicles. There are numerous ways to calculate the rate, but not all rates make sense, even if the numerator and denominator are in the same unit. For example, the number of teen fatalities divided by the number of registered drivers does not make sense even though the numerator and denominator are in the same unit (i.e., persons). Analysts should carefully examine the meaning of the resulting rate. Also, analysts should examine if rates are valid comparison measures (Kweon, 2008). 3.1.3.3. Trend Analysis A trend is simply listing frequencies or rates in chronological order, typically in a graph, so that patterns in frequency or rate over time can be identified and simple prediction of frequency or rate for the future can be made. For example, for a certain city, the annual number of fatalities or fatality rate per population during the past decade can be listed in a graph and a decreasing or increasing trend during the decade can be identified. In the case in which a large fluctuation in the frequencies or rates is present, the moving average technique is especially helpful for identifying the overall trend.
3.1.4. Regression-Based Aggregate Data Analysis The three aggregate data analyses described previously can account for various factors but in a limited way. For example, the crash rate per VMT can be calculated for two cities and a fair comparison between the two cities is possible only with regard to VMT. We can further divide the rate, for example, by urban/rural classification so that we obtain the crash rates for rural and urban roads and
PART | I
Theories, Concepts, and Methods
TABLE 8.1 Illustration of Simpson’s Paradox in Traffic Safety Analysis Region
Fatality rate (per 100 million vehicle miles traveled) Total
Rural
Urban
A
1.27
0.92
2.68
B
2.12
0.87
2.49
a fair comparison is possible for urban and rural separately. We could continue to narrow the crash rates by introducing more factors into consideration for calculating the rates. However, it is very difficult to make an overall assessment of traffic safety using such fine-tuned crash rates. We can make a valid safety assessment only under the conditions considered to calculate those rates. To make a valid assessment under general conditions, it is necessary to control for various contributing factors simultaneously, and a regression analysis is the most popular way of doing so. Including many factors contributing to crashes in the regression equation normalizes those factors simultaneously so that the safety situation of the cities can be evaluated considering different conditions in those factors among the cities being compared. Analysts should be aware that the aggregate analysis, especially the rate analysis, is subject to Simpson’s paradox. Table 8.1 illustrates the paradox in traffic safety. Suppose two regions are compared in the traffic safety situation based on the fatality rate per 100 million VMT. According to the total rate, region A appears to be much safer than region B. However, when the rate is broken down by urban/ rural classification of roads, region B is safer than region A based on both rural and urban rates. The two regions are very different in their VMT ratio between urban and rural roads. VMT of region A is 20% rural and 80% urban, whereas that of region B is 77% rural and 23% urban. A failure to account for the urban/rural classification of roads in calculating the fatality rates in this example leads to the erroneous conclusion that region A is safer than region B.
3.2. Disaggregate Data Analysis Disaggregate analysis is analysis directly using information from individual crash records. A case study examining crash records individually can also provide insight into the traffic safety situation and uses data to some extents but is not considered a data-driven approach. Thus, it is not viewed as disaggregate data analysis. Disaggregate data analysis is typically performed using regression analysis. In preparing data for disaggregate analysis, the aggregation task is not
Chapter | 8
Crash Data Sets and Analysis
usually involved but the other three basic tasks, especially the integration task, may be involved. For example, if we are interested in identifying the contributing characteristics of drivers in SUVs to occurrences and outcomes of crashes on interstate highways in rain while controlling for roadway characteristics of crash locations, we first need to integrate roadway inventory data into crash data (integration), and then we need to select crashes that occurred on interstate highways in rainy conditions (selection) to obtain data suitable for the disaggregate analysis. Various types of regression models can then be applied to the formed datad Typical types of models suitable for crash data are discrete response models (e.g., binary logit and ordered probit models) and count data models (e.g., Poisson and negative binomial models).
105
ACKNOWLEDGMENT http://www.trafficrecords101.net is an excellent web-based source to learn about traffic safety data and analysis and served as the main source for this chapter.
REFERENCES Kweon, Y.-J. (2007). Prediction of fatality rates for state comparison. Transportation Research Record, 2019, 127e135. Kweon, Y.-J. (2008). Examination of macro-level annual safety performance measures for Virginia. Transportation Research Record, 2083, 9e15. Kweon, Y.-J., & Kockelman, K. M. (2003). Overall injury risk to different drivers: Combining of exposure, frequency and severity models. Accident Analysis and Prevention, 35(4), 441e450.
This page intentionally left blank
Part II
Key Variables to Understand in Traffic Psychology 9. Neuroscience and Young Drivers 10. Neuroscience and Older Drivers 11. Visual Attention While Driving: Measures of Eye Movements Used in Driving Research
109 127
137
12. Social, Personality, and Affective Constructs in Driving 13. Mental Health and Driving 14. Person and Environment: Traffic Culture 15. Human Factors and Ergonomics
149 165 179 193
This page intentionally left blank
Chapter 9
Neuroscience and Young Drivers A. Ian Glendon Griffith University, Queensland, Australia
1. YOUNGER DRIVERS Although not applicable to all individuals, adolescents and emerging adults tend to exhibit various forms of reckless behavior, characterized by sensation seeking and risk taking (Arnett, 1992; Igra & Irwin, 1996; Moffitt, 1993; Spear, 2000). Engaging in one form of risky or reckless behavior increases the likelihood of engaging in others (Dryfoos, 1991; Duangpatra, Bradley, & Glendon, 2009). Risk perception, risk taking, and crash involvement among young novice car drivers, particularly males, has received considerable research attention (Arnett, 2002; Brown & Groeger, 1988; Deery, 1999; Deery & Fildes, 1999; Harre´, 2000; Jonah, 1997; McKnight & McKnight, 2003). Younger drivers are a heterogeneous population aged between 16 and 24 years. Older (20e24 years) and younger (16e19 years) cohorts have been distinguished (Corley, 1999; Jurkiewicz, 2000; Zemke, Raines, & Filipczak, 2000). Giorgio et al. (2008) identified adolescents as aged 13.5e21 years and young adults 22e42 years, reporting that across studies the age differentiating adolescence from young adulthood varied between 17 and 22 years. Notwithstanding individual differences, adolescence could be considered as the teenage years and up to 21 years, whereas emerging adulthood comprises ages 22e29 years. The age range of interest in this chapter is between 16 and 29 years. Driver crash rates increase up to age 18 or 19 years and decline slowly thereafter (Marin & Brown, 2005; Williams, 2003). Crash risk is greatest during the first 6 months or 1000 km (625 miles) of independent driving (Mayhew, Simpson, & Pak, 2003).
2. EVIDENCE FROM DEVELOPMENTAL NEUROSCIENCE RESEARCH 2.1. Prefrontal Cortex The prefrontal cortex (PFC) is the site of executive functions, including high-level cognitive processes through which we develop and execute detailed plans, make judgments about long-term goals, and weigh risks and Handbook of Traffic Psychology. DOI: 10.1016/B978-0-12-381984-0.10009-8 Copyright Ó 2011 Elsevier Inc. All rights reserved.
consequences of actions. The dorsolateral PFC (DLPFC), responsible for executive decision making and impulse control, is among the last brain regions to develop, becoming fully mature during approximately the mid-20s (Giedd, 2004). Key cognitive functions ascribed to the PFC include judgment, decision making, working memory (dorsal and lateral regions), and response inhibition (dorsolateral and orbitofrontal regions) (Casey et al., 1997). Casey et al. suggested that the greater the activation of the orbitofrontal cortex during a decision-making task, the greater the inhibition. Growth in this area between ages 10 and 12 years is followed by a dramatic decline that continues into the early 20s due to pruning unused neuronal pathways. PFC maturation is assumed to correspond to higher level cognitive development throughout childhood and adolescence (Casey et al., 1997). Overman (2004) found that in a decision-making gambling task involving the PFC, children were better than adolescents at probability matching, and that adolescents tended to become frustrated by the task, making more errors than did adults. Adolescent behaviors that may be labeled as “irrational” and “disorganized,” including those that can increase driving crash risk, may result from the process of dendritic tree maturation as learning affects higher cortical functions (Dicks, 2005). Implications of this maturation process for driving might include attenuated competence in developing plans and goals for the driving task, impaired ability to weigh consequences of risk-taking behaviors, and a lower threshold for impulse control. Further specific neuroscience evidence on the driving task is required to confirm these implications. Implications of late DLPFC maturation range from whether teenagers should be allowed to drive to whether minors are sufficiently cognitively mature to be subject to the death penalty (Lenroot & Giedd, 2006; Steinberg, Cauffman, Woolard, Graham, & Banich, 2009; Steinberg & Scott, 2003). Lenroot and Giedd pointed to a tendency in these debates to overestimate current knowledge of brain biology, cognition, and behavior, particularly ignoring substantial individual differences. They explained that the 109
110
interplay of genetic, epigenetic, and environmental factors means that the relationship between brain development and resultant brain structures is exceedingly complex. As neurons develop, they are encased in a layer of white myelinda lipid and protein sheathdhereafter referred to as “white matter,” which enables a 100-fold increase in transmission speed of electrical impulses between neurons (Blakemore & Choudhury, 2006). Whereas some sensory and motor neurons become fully myelinated in early childhood (e.g., large efferent fibers from the spinal segment), myelination of frontal cortex axons continues well into adolescence. Phylogenetically older brain regions mature earlier than newer regions so that the cortex matures from the back to the front (occipital cortex to frontal cortex) as gray matter (cell bodies and dendrites) volume reduces. Critical changes also occur in the brain’s wiring such that synaptic densitydthe network of connections between neurons or “gray matter”dpeaks during early postnatal development, being higher throughout childhood than in adulthood. Synaptic elimination (pruning), in which frequently used connections (e.g., for the native language) are strengthened and infrequently used connections (e.g., for skills that are no longer practiced) are eliminated, occurs at a faster rate in the first few months of life and again soon after puberty. Neuronal reorganization resulting from synaptic pruning continues throughout adolescence, resulting in a net decrease in synaptic density of the frontal lobes, essential for improving neuronal network efficiency. This dendritic pruning, which results in increased synaptic strength, is particularly important to processing information that is essential to learning as motor information becomes “chunked,” a process that also involves the basal ganglia and cerebellum in feedback that is critical to learning skills such as those involved in driving. Although neuronal death is a natural process, also potentially relevant to driving is the fact that alcohol and other drugs can speed up and distort this process.
2.2. White and Gray Matter 2.2.1. General Changes Increased integration of distributed brain regions, reflected in white matter changes, is associated with greater associative connectivity and more extensive neural networks. Reporting on longitudinal studies of participants aged 3e30 years, Giedd (2008) noted that childhood gray matter peaks were followed by declines during adolescence, with the changing balance between limbic/subcortical and frontal lobe functions extending into young adulthood. Giedd noted the adaptive potential of the neuronal elimination process along with increased connectivity and brain function integration. The changing reward systems and frontal/limbic balance that inter alia serve to increase risk
PART | II
Key Variables to Understand in Traffic Psychology
taking and sensation seeking had been highly adaptive for our ancestors. Although risk taking and sensation seeking might continue to be adaptive in some contexts, such as extending social networks in the search for individual identity, in an unattenuated driving context they would be more likely to be maladaptive. Although brain size does not increase significantly after age 5 years, there is considerable subsequent progressive and regressive growth in different brain regions (Durston et al., 2001; Gogtay et al., 2004; Sowell, Thompson, Holmes, Jernigan, & Toga, 1999; Sowell, Thompson, Tessner, & Toga, 2001). Sowell et al. (2001) reported simultaneous regressive (e.g., synaptic pruning) and progressive (i.e., myelination, synaptic formation, and strengthening of extant synapses) cellular events from childhood to young adulthood. During adolescence, regressive changes dominate progressive changes (Suzuki et al., 2005). Sowell et al. (2001) suggested that improved cognitive task performance through adolescence and into young adulthood could result from regressive changes, such as synaptic pruning, as less efficient or infrequently used connections are discarded. Implications for learning skills associated with complex tasks such as driving lead to the dilemma that whereas this could be an optimum period for learning such skills, it coincides with the time of peak risktaking behavior. Enhanced task performance, such as required in driving, could also result from increased myelination, which improves conduction speed of electrical impulses. Further research is required to identify which cortical regions are associated with particular task types. For example, Casey et al. (1997) found that performance on a decision-making task was only related to activity in the orbitofrontal and anterior cingulate cortices. Baxter, Parker, Linder, Izquierdo, and Murray (2000) found that the ventromedial PFC and amygdala are important in guiding people to use information efficiently concerning positive and negative outcomes required to make good decisions about future behaviors (e.g., driving more safely; Montague & Berns, 2002). A positive relationship has been found between the size of the anterior cingulate gyrus and harm avoidance (Pujol et al., 2002). The cingulate gyrus may be more involved in generating responses than in inhibiting them: The greater the activity in this region, the more likely it is that a motor response will be made (e.g., to avoid harm during driving; Casey et al., 1997). It is also likely that morphological changes could result from practicing certain skills, as in driver training (Draganski et al., 2004).
2.2.2. White Matter Giedd (2008) described myelination as “the wrapping of oligodendrocytes around axons, which acts as an electrical insulator and increases the speed of neuronal
Chapter | 9
Neuroscience and Young Drivers
signal transmission” (p. 336). Myelination also modulates synchronicity of neuronal firing to convey meaning (Fields & Stevens-Graham, 2002), with white matter pathways being particularly important for smooth information flows. Whereas myelination rate varies according to life stage, white matter volumes increase throughout childhood, adolescence, and into the third decade of life (Casey, Galvan, & Hare, 2005; Giedd et al., 1999; Giorgio et al., 2008; Gogtay et al., 2004; Toga, Thompson, & Sowell, 2006), being associated with higher processing speed. White matter volume could continue to increase up to age 60 years (Sowell et al., 2003). Giorgio et al. found that white matter pathways matured at different rates, with the most significant changes in the right body of the corpus callosum, associated descending motor pathways (basal ganglia), and the right superior region of the corona radiata. Although raw processing speed may be important in learning complex tasks such as driving, it is only one component of higher level cognitive functioning. Young drivers’ sometime overreliance on fast reaction time as a buffer against harm is therefore likely to be misplaced.
2.2.3. Corpus Callosum The corpus callosum (CC) is the most prominent white matter structure, comprising approximately 200 million axons connecting equivalent regions of the two cerebral hemispheres. The CC integrates sensory, memory storage and retrieval, attention and arousal, language, and auditory functions (Giedd, 2008; Lenroot & Giedd, 2006). The CC is among the last of the brain’s structures to complete maturation, undergoing rapid growth before and during puberty and lasting through adolescence until the mid-20s (BarneaGoraly et al., 2005; Giedd et al., 1999). CC signal intensity decreases between ages 7 and 32 years, with the most rapid changes during childhood, stabilizing in early adulthood as cerebral functioning becomes more lateralized (Keshavan et al., 2002). The number of connections increases during adolescence, and CC fibers are important in connecting motor and sensory cortices so that increased white matter in this location may be associated with improved motor skills during development (Barnea-Goraly et al., 2005) and in adulthood (Johansen-Berg, Della-Maggiore, Behrens, Smith, & Paus, 2007), such as are required for skilled driving performance. The CC influences handednessdwhether an individual has a strong preference (usually for right-handedness) or is “mixed-handed.” Wolman (2005) reported that rather than left-handers being more prone to have vehicle crashes, as was once thought, it is mixed-handers who are more at risk. Consistent with an interhemispheric model would be the enhanced risk of someone talking on a cell phone (a predominantly left-hemisphere task involving language) while driving with the left hand (a predominantly
111
right-hemisphere task for the motor performance component). Further neuroscience evidence is required on potential compromises to the driving task of “cross-talk” between the brain hemispheres that could result from engaging in various secondary tasks during driving.
2.2.4. Gray Matter The childhood peak in gray matter volume and number of synapses is followed by a decline during adolescence, reflecting continuous brain development throughout this period, and a further post-adolescent decrease (De Bellis et al., 2001; Giedd, 2008; Giedd et al., 1999; Paus, 2005; Suzuki et al., 2005). In the nonlinear pattern of gray matter changes, the decrease is most dramatic during adolescence (Sowell et al., 2003). Sowell et al. (2001) found that frontal lobe gray matter loss in the left hemisphere is much greater between adolescence and adulthood than it is between childhood and adolescence. Cortical gray matter loss occurs earliest in the primary sensorimotor areas and latest in the DLPFC and lateral temporal cortex (Gogtay et al., 2004). Whitford et al. (2007) found that gray matter decreased in participants aged 10e30 years in the frontal and parietal cortices, with the greatest change occurring during adolescence. Gray matter changes appear to be region specific, being sometimes progressive and at other times regressive (Blakemore & Choudhury, 2006). For example, gray matter volume in the temporal lobes was found to peak at approximately age 17 years, whereas in the occipital lobes its development was relatively linear (Giedd et al., 1999). In the frontal lobes, gray matter decline was particularly pronounced between adolescence and adulthood. The decline in PFC gray matter volume accelerates during the 20s, although the brain does not reach full maturity until approximately age 30 years (Sowell et al., 2001) and developmental changes continue throughout the adult life span.
2.3. Other Brain Regions Likely to Be Important in Driving Behavior 2.3.1. Amygdala, Hippocampus, and Associated Structures The amygdala and hippocampus, which are concerned with experiencing and expressing emotions, increase in volume with age (Durston et al., 2001; Suzuki et al., 2005). Along with the temporal lobes, the amygdala and hippocampus are involved in emotion, language, and memory for which human capacity changes radically from ages 4 to18 years (Lenroot & Giedd, 2006). The amygdala is critical in assessing the salience of environmental stimuli to survival. The hippocampus, which is involved in memory storage,
112
consolidation, and retrieval, has connections with other limbic structures and the neocortex, and it has a role in integrating emotion with cognition (Benes, 1994). Connections between the neocortex and limbic systems are important in memory for stimuli with high salience. Episodic memory operates optimally during adolescence and early adulthood, perhaps due to changing levels of hormones and neurotransmitters (Janssen, Murre, & Meeter, 2007). This feature might have survival value from the perspective of enhanced learning during a developmental stage when sensation seeking and risk taking expose individuals to high levels of environmental stimuli (Giedd, 2008), as during driving. This developmental change would also tend to be positive for early introduction to learning complex skills, such as are required for safe driving performance. The pineal gland produces the hormone melatonin, the levels of which rise in the evening to signal to the body that it is time to sleep. During adolescence, melatonin peaks later in the day than it does in children or adults, which helps to explain why teenagers often prefer to both rise and go to bed later than do adultsdthe “delayed phase preference” (Carskadon, Acebo, Richardson, Tate, & Seifer, 1997; Steinberg, 2008a). Combined with other features of adolescent behavior, this could also help to explain why they may be overrepresented in vehicle crashes at night (Maycock, 2002). Because melatonin levels remain high upon waking for school or work, adolescents tend to be least alert between 8:00 and 9:00 a.m. and may lose up to 2 h of sleep at night as a result of the shift in the timing of the melatonin cycle (Steinberg, 2008a). Chronic sleepiness that could result from this sleepewake cycle, interacting with loss of vigilance in the standard circadian cycle, has the potential to degrade safe driving performance in this age group. Smith, Horswill, Chambers, and Wetton (2009) found that younger inexperienced drivers’ hazard perception skills were significantly impaired by a mild increase in sleepiness. They reviewed research indicating that increased sleepiness impairs a range of cognitive processes vital to safe driving and that many crashes reported as due to “inattention” may be related to sleepiness.
2.3.2. Cerebellum The cerebellum, a component of the oldest part of the brain, governs posture and movement, helping to maintain balance and ensuring that movements are smooth and directed. It also influences other areas of the brain responsible for motor activity and continues to grow until late adolescence. Traditionally associated with balance and motor control, having links with the DLPFC, the medial frontal cortex, and the parietal and superior temporal areas, the cerebellum also has a role in higher cognitive functions, motor learning, and adaptation (Giedd, 2008). All these
PART | II
Key Variables to Understand in Traffic Psychology
functions are likely to be critical to driving. Cerebellum volume peaks approximately 2 years later than does cerebral volume, and it is the only brain structure to remain significantly larger in males after co-varying for total cerebral volume (Mackie et al., 2007). Involvement of motor and cerebellar networks in driving was confirmed by Calhoun, Pekar, and Pearlson (2004), who also found that the cerebellar network exhibited a highly significant alcohol dose-related effect for high-speed driving and the number of times the speed limit was exceeded. Compared with drivers aged 30e39 years, male and female drivers aged 17e19 years have approximately twice the proportion of their crashes while negotiating a bend (Maycock, 2002). Maycock also found that 17- to 19-year-old male drivers were more than 50% more likely than were female drivers of the same age group to crash when negotiating a bend.
2.3.3. Temporal Lobes The last parts of the temporal lobes to mature are the superior temporal gyrus and sulcus, which have a role (along with prefrontal and inferior parietal cortices) in integrating memory, audiovisual input, and object recognition. This integration is likely to be critical to the driving task.
2.3.4. Parietal Lobes Brain regions serving primary functions, such as sensory and motor systems, mature earliest; higher level control and integrative functional regions mature later (Casey, Getz, & Galvan, 2008; Gogtay et al., 2004; Sowell, Thompson, & Toga, 2004). Sowell et al. (1999) found that visuospatial functions typically associated with parietal lobes operated at a more mature level earlier than executive functions typically associated with frontal brain regions. Adolescents typically develop their visuospatial abilities prior to the means to fully interpret the meaning of all stimuli. Hence, although young drivers can “see” the same things (including obvious hazards) as adults, they cannot always perceive risks appropriately because they have yet to fully develop higher level cognitive interpretive functions. Critical to risk perception in driving is the ability to appreciate the possibility of hidden hazards and interpret conditional probabilities of the following form: “If that pedestrian has not seen my vehicle and decides to step off the sidewalk . then I had better be prepared to stop quickly.” Differential development of the cerebellum and cortical regions responsible for higher order cognitive processing is consistent with findings from studies revealing younger drivers’ difficulties in perceiving risk accurately. All drivers frequently need to make complex decisions under conditions of uncertainty, which improve with experience. Anticipation and risk avoidance skills develop
Chapter | 9
Neuroscience and Young Drivers
with experience (Simons-Morton, 2002). In gaining this experience, young drivers have the compounding issue of inhibiting greater risk-taking impulses.
2.3.5. Other Subcortical Structures The basal ganglia (caudate nucleus, putamen, globus pallidus, and amygdala) are involved in mediating movement, higher cognitive functions, attention, and emotions (Giedd, 2008; Lenroot & Giedd, 2006), and they have reciprocal connections with the substantia nigra in the midbrain. Large developmental changes occur in the basal ganglia, particularly in males (Casey et al., 2008; Giedd, Snell, et al., 1996). Spear (2000) reported that the balance between dopamine systems begins to shift from subcortical toward cortical levels during adolescenceda change that reflects enhanced acquisition of higher level executive function during this period.
3. CRITICAL ASPECTS OF DRIVING LINKED WITH NEUROLOGICAL DEVELOPMENT 3.1. Response Inhibition Experiments using “go/no-go” tasks require multiple executive functions, including working memory and inhibiting a normal/prepotent response, when participants are shown a certain stimulus (Blakemore & Choudhury, 2006). A driving example would be inhibiting motor responses to start or stop a vehicle at traffic lights when a directional turn (filter) arrow incongruent with the main light is displayed. Studies with children and adults have indicated that inhibiting the normal response (e.g., to proceed or not to proceed when a traffic light shows either green or red, respectively) involves several regions of the frontal cortex, including the anterior cingulate, orbitofrontal cortex, and the inferior and middle frontal gyri (Casey et al., 1997). Individual differences were evident; for example, participants with the lowest error rates showed the greatest orbitofrontal activation and the least DLPFC activation. Casey et al. suggested that during adolescence, the neural network recruited for such tasks matures so that by adulthood a smaller region of the PFC is used to perform this type of task. Continual development of the PFC and parietal cortex during adolescence is reflected in executive functiondcontrolling and coordinating thoughts and behavior. This includes such cognitive features as selective attention, decision making, voluntary response inhibition, and working memory (Blakemore & Choudhury, 2006). During driving, these functions might be involved in filtering out unimportant information (e.g., irrelevant road signage), prospective memory (remembering to carry out an intended action in the future such as a route to
113
a destination), and inhibiting impulses (e.g., expressing anger at other road users). These skills rely heavily on the frontal lobes, and studies have shown that adolescents’ task performance continues to develop, for example, in inhibitory control (Leon-Carrion, Garcia-Orza, & Perez-Santamaria, 2004; Luna, Garver, Urban, Lazar, & Sweeney, 2004), processing speed (Luna et al., 2004), and working memory and decision making (Hooper, Luciana, Conklin, & Yarger, 2004; Luciana, Conklin, Cooper, & Yarger, 2005). Speed of switching between tasks, essential as a driving skill, continues to develop during adolescence, which could be another argument for early acquisition of driving competence.
3.2. Controlling Risky Behavior Steinberg (2008b) attributed the increase in risk taking between childhood and adolescence to changes in the brain’s socioemotional system, which tends to increase reward seeking, particularly in the presence of peers. Steinberg attributed the decline in risk taking between adolescence and adulthood to changes in the brain’s cognitive control system, which gradually improves an individual’s capacity for self-regulation. The right ventral striatum, thought to be involved in motivating rewardseeking behavior, is less active in teenagers than in adults, suggesting that adolescents could be more prone to risky but potentially high-reward behaviors (Dicks, 2005). Seeking to account for adolescents’ greater risk-taking propensity, Bjork et al. (2004) suggested that differences in brain activation in mesolimbic regions during incentivemotivated behaviors might be involved. Compared with adults, Bjork et al. found that when anticipating gains, adolescents showed reduced activity in the right ventral striatum and right amygdala, which could be associated with reduced levels of fear. These authors suggested that adolescents’ risk behaviors compensated for low recruitment of networks associated with these brain regions by seeking more extreme incentives. Driving examples include speeding and tailgating, as well as more deviant forms of driving behavior. Steinberg reviewed evidence indicating that during adolescent brain development, peer acceptance may be processed in similar ways as other types of rewards, which could be critical for risk taking during driving. Baird and Fugelsang (2004) studied that aspect of human reasoning concerned with imagining alternative outcomes and the consequences of different courses of actiondessential aspects of risk perception and risk-taking behaviors. Comparing brain activity in teenage and adult samples faced with dangerous and safe scenarios, adults showed greater activity in parts of the brain creating mental imagery and signaling internal distressdboth associated with a rapid and automatic response to danger. Baird,
114
Fugelsang, and Bennett (2005) found that compared with adolescents, adults showed greater activation in the insula and right fusiform face area when assessing potentially dangerous actions, indicating that these processes had been automated in adults but not yet in adolescents. Teenagers showed increased PFC activity, with reasoning and judgment activity resulting in extended decision making, revealing a less appropriate response to inherently dangerous situations (e.g., driving at high speed), which do not require extended reasoning for most adults. Confronted with a potentially dangerous scenario, adults were more likely to create a mental and visceral image (“gut feeling”) of possible outcomes (e.g., injuries that could be sustained) and to have a readily available appropriate aversive response based on affect (e.g., “this feels like a very bad idea”). Teenagers, who took significantly longer to decide upon the dangerousness of the scenarios and to produce a correct response, appeared to use reason, involving greater DLPFC activation, to determine whether a scenario was dangerous because they did not have a general mental image and associated visceral response available to guide their decision making. Reyna and Farley (2006) argued that adolescents are able to reason and understand the risks of the behaviors in which they engage, using an intuitive form of decision making. However, actions taken in rewarding and emotional contexts explain why some teenagers are at greater risk for poor decision making and bad outcomes (Galvan et al., 2006; Galvan, Hare, Voss, Glover, & Casey, 2007). Important to successful maturation is developing an ability to suppress inappropriate thoughts and actions in favor of goal-directed ones, particularly within a context of competing and more immediate rewards (Casey, Tottenham, Liston, & Durston, 2005; Casey et al., 2008). All these features may be represented in driving. Risk taking, which is distinct from impulsivity, tends to be higher in adolescence than in either adulthood or childhood. It is associated with subcortical regions, specifically the nucleus accumbens, a region of the basal ganglia involved in making risky choices when evaluating potential rewards (Bjork et al., 2004; Casey et al., 2008; Ernst et al., 2005; Ernst, Pine, & Hardin, 2006; Galvan et al., 2006, 2007; Kuhnen & Knutson, 2005; Matthews, Simmons, Lane, & Paulus, 2004; May et al., 2004; Montague & Berns, 2002). Activity in this region immediately prior to making risky choices is higher in adolescents than in either children or adults (Ernst et al., 2005; Galvan et al., 2006). Casey et al. postulated that adolescents’ increased tendency to higher levels of both impulsivity and risk taking, compared with other life cycle stages, could be explained by their bias to immediate rewards over achieving longer term goals imposed by the combination of changes in the relative rates of development of subcortical regions (e.g., accumbens) to control regions
PART | II
Key Variables to Understand in Traffic Psychology
(e.g., ventral PFC) and white matter tracts linking these regions. In a driving context, this developmental feature could lead adolescent drivers to favor immediate rewards linked to risk taking over longer term life goals, such as remaining free from injury. Although adolescents in general appear to be more prone to risky decision making (Gardner & Steinberg, 2005), Casey et al. (2008) highlighted the importance of individual differences among adolescents with respect to risk taking, suggesting that genetic variation could underlie such differences, for example, relating to dopamine release to subcortical regions (O’Doherty, 2004; Steinberg, 2008b). Exploring the association between reward-related neural circuit activity and anticipating monetary rewards, Galvan et al. (2007) found a positive association between accumbens activity and the likelihood of engaging in risky activity across participants aged between 7 and 29 years. However, in individuals who perceived risk taking as leading to bad outcomes, the accumbens was less activated to reward. Impulsivity was related to age but not with accumbens activity. Casey et al. interpreted their findings as indicating that adolescent choices and behavior could not be explained by impulsivity or by PFC maturity following that of other brain regions alone. However, they could be accounted for by individual differences in risk taking as well as explaining how this aspect of adolescent behavior differed from that of both children and adults. In addition, social and situational influences such as peer pressure can play an important role in risk taking during driving. One possible mechanism for the transmission of risktaking behavior resulting from peer pressure during driving emerged in a study comparing an adolescent sample (aged 14e19 years) with older and younger groups. In this study, Cohen et al. (2010) determined that the striatum and angular gyrus were the two regions in which only the adolescent group had a hypersensitive response to prediction error. A driving example might involve approaching traffic lights showing green at speed from some distance away and seeing them turn to amber at the critical “go/nogo” decision pointdthe prediction in this case being that the lights would remain green. The ventral striatum was consistently sensitive to unexpected positive feedback, and of the three groups studied, only the adolescents responded more quickly to large rewards compared with small rewards. The “reward” in the traffic lights example would be successfully clearing the junction, irrespective of the color of the lights at the time of crossing. Building on the work of Ernst et al. (2005) and Galvan et al. (2006), which found that adolescents have a hypersensitive response to reward, Cohen et al. found that this was specific to prediction error rather than to valuation signals. Continuing the traffic lights example, valuation signals might involve reflecting on, or being alerted to by a passenger, the possibility that in different circumstances the outcome
Chapter | 9
Neuroscience and Young Drivers
could have been differentdfor example, involving a crash. Cohen et al. surmised that developmental differences in prediction error responses reflected differences in dopamine signaling, which could explain risky reward-seeking behavior typical of adolescents, for example, in driving. Cohen et al. considered that their data were consistent with a greater effect of positive outcomes, resulting in increased motivation for positive outcomes (e.g., peer approval of risky driving) and hence greater risk-taking propensity. They argued that adolescents’ overactive dopaminergenic predictive error response could result in increased reward seeking, particularly when combined with an immature cognitive control system. In a monetary decision-making task involving risk, compared with adults, adolescents showed lower levels of ventral PFC (VPFC) activity (Eshel et al., 2007). Although the strictures of ecological validity mean that the extent to which experimental studies of risk in decision making can be extrapolated to real-world driving risk is uncertain, such findings would be consistent with observed data, for example, of adolescents’ relatively high level of involvement in vehicle crashes. However, it remains to be determined whether the same brain circuitry that appears to guide non-life-threatening risk taking is also critical to the etiology of young drivers’ heightened risk-taking susceptibility. In a review, Ernst et al. (2006) explained the propensity during adolescence for reward/novelty seeking when confronted with uncertainty or potential harm by the combination of a strong reward system (nucleus accumbens), a weak harm-avoidant system (amygdala), and an inefficient supervisory system (medial/ventral PFC). This combination could be reflected in dangerous driving behavior. Notwithstanding individual differences, diminishing impulsivity throughout adolescence is associated with PFC development (Casey, Tottenham, et al., 2005; Casey et al., 2008; Galvan et al., 2007). Liston et al. (2006) showed that although white matter tracts continue to develop into adulthood, only those linking the PFC with the basal ganglia are associated with impulse control. Such findings emphasize the importance of the development of circuits as well as brain regions. Cognitive development throughout adolescence is characterized by increased efficiency of cognitive control and improved emotional regulation, primarily dependent on PFC maturation (Yurgelun-Todd, 2007). Casey et al. (2008) argued that limbic subcortical and prefrontal control regions should be considered as linked components of a developing system, with the limbic component developing before the prefrontal region. They suggested that adolescents’ behaviors are biased by their more functionally mature limbic regions. Developing functional connectivity between these components is critical to enhanced top-down control of subcortical regions. Cortical connections are enhanced as synapses are pruned
115
through a combination of development and experience. A key challenge for researchers and practitioners of young driver training is to capitalize on this developmental imperative while simultaneously moderating the inevitable downside of this critical maturational phase. This might be achieved by early and frequent reinforcement of the pathways that would develop naturally during this perioddfor example, by ensuring that adolescents are given opportunities to practice impulse control in simulated driving scenarios or during on-road driver training. This would contrast with the more typical approach to the practical component of driver training, which is to practice behaviors that accord with the relevant jurisdiction’s road rules until a certain standard can be reproduced during the driving examination. Whereas future orientation, impulse control, resistance to peer influence, punishment sensitivity, and planning showed linear increases during early adolescence, Steinberg (2008b) found a curvilinear pattern for sensation seeking, risk preference, and reward sensitivity. These distinct patterns might suggest that although those psychosocial phenomena that develop linearly might be open to “accelerated learning” strategies (e.g., during driver training), context-based approaches are likely to be required for those that develop nonlinearly. For example, because risk taking is more likely to be a group-based phenomenon for adolescents than for adults, restrictions on the number of peer passengers should be mandatory for younger drivers.
3.3. Processing Emotions Adolescence is known to be a time of greater vulnerability to an imbalance between emotion processing and top-down regulation. For example, exposure to both positive and negative information has been shown to result in heightened activity in subcortical limbic regions (e.g., ventral striatum and amygdala) in adolescents compared with adults (Ernst et al., 2005; Eshel et al., 2007; Galvan et al., 2006; Monk et al., 2003). Although later development of subcortical systems relative to top-down control systems generally biases adolescents’ behavior toward immediate over long-term goals (Galvan et al., 2006), this effect is moderated by considerable individual variability in emotional reactivity and its regulation. Basic emotions are generated mainly in the limbic system with neural circuits associated with each emotion, including happiness, fear, and anger (Panksepp, 1998, 2001). Panksepp maintained that each basic state was associated with an action category shaped by evolution to continue an activity, to sense danger and escape, or to prepare for confrontation. Limbic system circuits are relatively fixed and can powerfully affect our cognitions. The amygdala performs a rapid initial appraisal to detect
116
whether some external event presents a threat, and it induces feardthe “low road” (LeDoux, 1997). This can result in physiological responses that include blood pressure change, stress hormones release, and the startle reflex, as well as behaviors such as flight, fight, or immobilization. The longer route, via the cortex (the “high road”), results in more detailed analysis of an event. During adolescence, additional demands are placed on executive (cognitive) systems (e.g., memory, perception, decision making, and problem solving) and on the interplay between cognitions and emotional processesdfor example, in processing verbal and nonverbal environmental cues. These include the cognitions and emotions involved in interpersonal peer interactions (Watkins et al., 2002). Although basic aspects of face perception are in place soon after birth, quality and quantity of processing the meaning of facial expressions continue to improve throughout adolescence (Carey, 1992; McGivern, 2002; Taylor, McCarthy, Saliba, & Degiovanni, 1999). Of potential relevance to driving, and other social situations, is that processing the critical emotion of fear in others’ faces appears to be relatively weak in adolescents (Baird et al., 1999; Thomas et al., 2001). In a driving context, this effect could impair interpreting emotional feedback from peersdfor example, fear or anxiety expressed facially by passengers in a vehicle driven by an adolescent (Doherty, Andrey, & MacGregor, 1998). Hare and colleagues (Hare, Tottenham, Davidson, Glover, & Casey, 2005; Hare et al., 2008) showed that mean reaction times for acknowledging fearful facial expressions were positively correlated with amygdala activity. Hare et al. (2008) found elevated amygdala activity in an emotional context (fearful faces) in adolescents compared with children and adults. Individual differences in emotional expression could be accounted for by the strength of connectivity between top-down control regions, especially the VPFC and bottom-up emotional processing regions (amygdala). In adolescents, the strength of coupling between VPFC and amygdala was correlated with greater habituation of amygdala activity, showing that learning could occur. Less trait-anxious adolescents showed greater VPFC and less amygdala activity (habituation) after initial trials on a go/no-go task. Compared to adults, adolescents responded more slowly to fear targets and showed less prefrontal relative to amygdala activity, suggesting that they were likely to be more susceptible to emotional interference in decision making (“heat of the moment”) (Luna & Sweeney, 2004). Suppressing competing emotional responses involved adolescents engaging more prefrontal regions compared with adults (Luna & Sweeney, 2004). Although individual differences should not be ignored, learning in the context of specific stimuli can occur within a relatively short period of time (Hare et al., 2008). This knowledge could be put to
PART | II
Key Variables to Understand in Traffic Psychology
good use in training young drivers to enhance their level of control when faced with potentially emotionally arousing situations (e.g., perceived threats from other road users).
3.4. Sex Differences On average, the male brain is 10% larger than the female brain, with most structures displaying this difference (Durston et al., 2001). Females have proportionately more gray matter and less white matter; males have a higher ratio of white matter to brain volume (Luders et al., 2005; Nagel et al., 2006). Nagel et al. found a negative relationship between prefrontal white matter volume and age, particularly among an adolescent female subsample. In contrast with previous research, the nonlinear trend in their data suggested that at approximately 15 years of age, prefrontal white matter appeared to decline. Women have a higher concentration of gray matter in the neocortex (the phylogenetically newer part of the cerebral cortex), whereas men have proportionately more gray matter in the “older” entorhinal cortex. During childhood and adolescence, males have more prominent age-related gray matter decreases and white matter increases (De Bellis et al., 2001; Giedd et al., 1999). Male brains consistently show greater hemispheric asymmetry; in female brains, the two hemispheres are much more alike (Good et al., 2001). Total cerebral volume peaks at 11.5 years in girls and 14.5 years in boys (Giedd, 2008; Lenroot & Giedd, 2006; Lenroot et al., 2007). By age 6 years, brain volume is at approximately 95% of this peak (Giedd, 2008; Lenroot, & Giedd, 2006). Gray matter volumes follow inverted Ushaped trajectories, which are distinct for each lobe, and peak 1e3 years earlier in females (Lenroot & Giedd, 2006; Lenroot et al., 2007). Although mean total cerebral volume is 9 or 10% larger in boys (Goldstein et al., 2001; Lenroot & Giedd, 2006), total brain size differences do not imply any functional differentiation, and healthy children of the same age may show up to 50% difference in total brain volume (Giedd, 2008; Lenroot & Giedd, 2006). Brain morphology is highly variable between individuals, and although there are significant sex differences, there is considerable overlap between male and female distributions. Silveri et al. (2006) found that functionality in different brain regions accounted for some sex differences in impulse control. Although gender differences in human brain anatomy are well-established (Goldstein et al., 2001; Gro¨n et al., 2000; Gur, Gunning-Dixon, Bilker, & Gur, 2002; Nopoulos, Flaum, O’Leary, & Andreasen, 2000; Overman, 2004), the development of such differences is less well understood. Evidence from studies of adolescent brains has shed light on differential rates of maturation of the amygdala and higher cortical areas that control impulsive behavior. Although it has been known for some time that
Chapter | 9
Neuroscience and Young Drivers
girls’ brains develop faster than those of boys, research has revealed larger differences than previously suspected. However, the exact “developmental lag” of males vis-a`-vis females remains problematic. For example, Caviness, Kennedy, Richelme, Rademacher, and Filipek (1996) found that the typical brain of a 17-year-old boy resembled that of an 11-year-old girl, and a similar difference was found by Anokhin, Lutzenberger, Nikolaev, and Birbaumer (2000). Measuring brain myelination, Benes, Turtle, Khan, and Farol (1994) found that between ages 7 and 22 years, girls’ brains were 3 or 4 years ahead of boys’ brains, and that the men did not catch up with the women until age 29 years. Blanton et al. (2004) found significant gender differences in white matter development in speech-related brain areas, with boys but not girls showing a linear age-related increase in white matter volume. However, Giedd et al. (1999) found that gray matter development peaked at age 10 years in girls and at age 12 years in boys, after which ages there was a significant decrease in gray matter volume. De Bellis et al. (2001) found that between 6 and 18 years of age, males had a 19% reduction in gray matter compared with a less than 5% reduction in females. They also found that males had a 45% increase in white matter and a greater than 58% increase in CC area compared with females, for whom corresponding increases were 17 and 27%. The caudate nucleus and the hippocampus, which contain predominantly estrogen receptors, are proportionately larger in female brains, whereas the amygdala (containing predominantly androgen receptors) is proportionately smaller (Lenroot & Giedd, 2006). Longitudinal data showed that amygdala volume increased significantly with age only in males, whereas hippocampal volume increased significantly with age only in females (Giedd, Vaituzis, et al., 1996). During childhood and adolescence, myelination in the hippocampus occurred earlier in females than in males (Benes et al., 1994; Suzuki et al., 2005). Gur et al. (2002) reviewed studies on the generation and regulation of emotion. The amygdala is considered to be primarily involved in the excitatory aspects of emotional behavior, including aggression, whereas the orbitofrontal region has a modulating function. Gur et al. found that orbitofrontal brain regions were relatively larger in women than in men, and that compared with men, women had more brain tissue available for modulating amygdala input. One implication is that women have more available brain tissue that can moderate emotions underpinning behavioral displays such as are seen in aggression, which may play a key role in some forms of risk taking, such as when driving. Goldstein et al. (2001) confirmed that women had larger cortex (particularly frontal and medial paralimbic cortices) volume relative to cerebrum size, whereas men had larger volumes relative to cerebrum size in the amygdala and hypothalamus. However, reviewing evidence on
117
gender differences in risk taking, Steinberg (2008b) considered that sex differences in risk-taking behavior may be mediated more by context than by biology, and that the gender gap in real-world risk taking could be narrowing independently of sex differences in brain morphology. Studying how emotion is processed in the brains of children aged between 7 and 17 years, Killgore, Oki, and Yurgelun-Todd (2001) found that in young children emotional activity was localized in the amygdala and other older subcortical brain areas, and that at this age connections to higher brain centers (cerebral cortex) had not yet been made. During adolescence, brain activity associated with emotion moves to the cerebral cortex, but by age 17 years this change has occurred only in girls, whereas in 17year-old boys the locus of emotional control remains in the amygdala. Killgore et al. also found differences between male and female children and adolescents’ amygdala versus prefrontal activation while viewing faces expressing fear. With increasing age, females, but not males, progressively increased prefrontal activation relative to amygdala activation. Critical from a driving perspective (and other activities involving risk) would be knowing the extent to which these research findings reflected real differences between males’ and females’ respective PFC regulated responses to fear, for example, as reflected in passengers’ facial expressions or other potentially fearinducing stimuli, and their consequential impact on driving behavior.
4. DISCUSSION AND CONCLUSIONS Understanding the full implications of absolute and differential stages of brain development for driving remains in its early stages. Because behaviors emanate from integrated activity among distributed networks, a research priority is to link brain regions and circuits with relevant cognitions, including those relating to safety and risk perception and behaviors relevant to avoiding danger and managing risk. For example, Calhoun et al. (2002, 2004) identified seven separate brain networks involved in a simulated driving task in small samples of young drivers: (1) bilateral components of the parieto-occipital sulcus, including portions of cuneus, precuneus, and lingual gyrusdinvolved in visual monitoring; (2) mainly occipital areasdfor low-order visual processing; (3) bilateral visual association and parietal areasdfor high-order visual processing and visuomotor integration; (4) motor cortex and (5) cellebellar areasdfor gross motor control and motor planning; (6) orbitofrontal and cingulatedfor error monitoring and inhibition, including motivation, risk assessment, and “internal space”; and (7) medial frontal, parietal, and posterior cingulatedfor vigilance, including spatial attention, visual stream, monitoring, and “external space.” Comparing task performance of adolescent samples aged
118
15 and 16 and 18e20 years, Santesso and Segalowitz (2008) found a greater amplitude in anterior cingulate cortex activity among the older group. This brain region, associated with performance errors and self-monitoring, was found to be more mature in the older group. However, response times and accuracy were comparable across the groups, suggesting that the greater error rate in the younger sample was not due to performance differences. This study might suggest an age-related basis for driving-related errors that are not related to basic driving skill, particularly those associated with self-monitoring of performance. Giedd (2008) suggested that differences in developmental trajectories could be more informative than adult differences. For example, Shaw et al. (2006) found that at age 20 years, developmental trajectories were more predictive of IQ than were cortical thickness differences. However, as Steinberg (2008b) noted, current knowledge about neurophysiological changes during adolescence exceeds our understanding of how these relate to specific behaviors. Steinberg’s research indicated that although basic information processing showed no maturation after age 16 years, self-reported future orientation improved up to age 18 years, whereas planning and impulse control continued to increase through the early 20s. Respective contributions of these cognitive functions to driving might be coordinated with the delivery of relevant driving-related information within each of these age ranges. As our understanding of the role of brain function and developmental stages improves, it should be increasingly possible to link these with skill development and training for particular driving tasks (Eby et al., 2007). Anticipating probable intentions and behaviors of other road users is among many driving skills required. Psychologically, this is described as taking the perspective of another. Common brain areas (popularly known as “mirror neurons”) in the superior parietal and right inferior frontal cortices are activated both when an individual performs an activity and when they observe another performing that same activity (e.g., driving). Although social perspective taking develops during puberty, further research is required to establish when the perceptual, motor, and social functions required for mature adult performance become fully integrated (Blakemore & Choudhury, 2006). To assist in identifying links between brain function and development, on the one hand, and cognitive indicators relevant to safety and risk cognition and driving behaviors, on the other hand, triangulated methods are required. Because most research to date has been cross-sectional, more longitudinal studies are needed to enhance understanding of these links. Reflecting the cost of data collectiondfor example, using functional magnetic resonance imaging (fMRI) techniquesdsample sizes in several studies have been small. Resources are required to increase sample sizes so as to enable population norms, including
PART | II
Key Variables to Understand in Traffic Psychology
variances, for males and females at different ages with respect to brain development to be mapped in more detail. Age and sex differences in brain development are complex, and key individual differences may be difficult to reveal. As population norms for age and gender differences become clearer, it should be possible to identify individual differences. It is well recognized that brain systems controlling arousal, emotional experiences, and social information processing become much more active at puberty, and that these are associated with increased novelty-seeking, sensation-seeking, and risk-taking behaviors. With puberty for many young people beginning at 12 years or younger and the age at which the brain can be considered to be fully mature in both sexes now being assessed as the mid to late 20s, there is a considerable time period during which the brain is being driven by hormonal changes to undergo substantial transformations in structure and function. Improving our understanding of these processes and their implications for safety and risk-related behaviors is a key issue for future research. Young drivers are much more likely than older drivers to be influenced by their peers (Gardner & Steinberg, 2005; Simons-Morton, Lerner, & Singer, 2005; Steinberg, 2008b). Gardner and Steinberg found higher levels of risk taking, greater focus on benefits than potential costs of risk taking, and riskier decisions by adolescents when in peer groups than when alone. Steinberg reported fMRI data indicating that although brain regions activated in a driving task associated with cognitive control and reasoning (e.g., prefrontal and parietal association cortices) were active irrespective of driving condition, additional brain regions were activated (medial frontal cortex and left ventral striatumdprimarily the accumbens, left superior temporal sulcus, and left medial temporal structures) when peers were present. This socioemotional network led to more risky driving behavior, indicating that peer presence enhanced rewards from potentially risky driving behavior. Gardner and Steinberg (2005) found that although selfreported resistance to peer pressure continued to age 18 years, peer presence continued to be evident at age 20 years. According to Engstro¨m (2003), peer presence appears to influence risk taking in young people up to approximately the age of 25 years. Until this age, the brain’s prefrontal lobes, which govern an individual’s ability to plan, control impulses, and weigh risks and benefits, are still maturing. When age and maturation are taken into account, task experience remains important. There is thus a need to provide “safe” learning opportunities during the early period of learning to drive. To partly offset adverse effects of peer influence, alternative role models are important. Within a driving context, parents provide important role models, particularly the young driver’s same-sex parent (Glendon, 2005).
Chapter | 9
Neuroscience and Young Drivers
Kuhn (2006) argued that enhanced executive control is the major feature of cognitive development during adolescence. She focused on adolescents’ increasing ability for self-determinationdfor example, monitoring and managing their learning. During this period, improved information processing is achieved through a combination of faster speed (via increased myelination of nerve fibers), greater capacity (e.g., memory), and more effective inhibitiondboth resisting interfering stimuli and controlling one’s own responses. Luna et al. (2004) suggested that improved neuronal communication and increased myelination not only supported greater processing speed but also benefited response inhibition through more efficient synaptic organization. However, Kuhn noted that although experimental paradigms can provide evidence for the effects of instruction to inhibit responses, there is much less information about situations in which adolescents make their own decisions about inhibiting their cognitions or behaviors and their success in achieving this. In other words, adolescents’ self-regulatory processes under “reallife” conditions, such as driving, have been little studied, but it is particularly important to know about such situations. As Lerner (2002) noted, adolescents are already largely self-developing. Neurological development is partly experience driven, in that neuronal connections are strengthened by activities that are engaged in, whereas connections that are not developed through experience weaken. This presents a paradox with regard to an activity such as driving, given that a central feature of many adolescents’ self-image is their driving ability. The more (and the earlier) experience that an adolescent has of driving, the stronger will become neuronal connections that relate to this activity and the more skillful he or she will become. This means that a critical task for those involved in training young drivers is to ensure that their developing skill and self-confidence are appropriately matched with the maturity to make good decisions and to infuse their perception of their own driving competence with learned appropriate inhibitory responses to as wide a range of situations as possible. This argues for an extended period of learning, involving multiple inputs and testing to take account of large interindividual variability in the range of abilities required to master complex driving tasks (Eby et al., 2007). On the basis of evidence reviewed here, what are likely to be among the most promising strategies within a driving context for promoting harm avoidance and risk reduction for younger drivers? Given that multitasking skills develop over an extended period up to young adulthood, it is important to limit the number of tasks that a driver needs to perform. Banning the use of cell phones, in-vehicle entertainment (perhaps other than radio), and other distractions is one option. A strategy pursued by increasing numbers of jurisdictions worldwide consists of a graduated introduction to the road environment. The success of graduated
119
driver training programs in the United States has been extensively reviewed (Branche, Williams, & Feldman, 2002; Foss & Evenson, 1999; Hallmark, Veneziano, Falb, Pawlovich, & Witt, 2008; Hedlund, 2007; Hedlund, Shults, & Compton, 2006; Keating, 2007; Lin & Fearn, 2003; Neyens, Donmez, & Boyle, 2008; Shope & Molnar, 2003). These might include passenger restrictions (to limit peer influence) and limiting nighttime driving (a known high risk). Rather than focusing on purely punitive approaches, Blakemore and Choudhury (2006) recommended allocating more resources to educational and rehabilitation programs to take greater account of normal developmental changes in adolescents’ brains. One option might be a specified number of hours of practice on a controlled road network, which could also provide video feedback on a young driver’s performance, emphasizing safe driving. Because of their effects on brain and behavior, alcohol, drugs, and fatigue add substantially to risks faced by younger drivers. Stringent testing, controls, and sanctions should apply. Although the intellect of a young person might not be in doubt, to address the affective component of brain maturation and to mitigate what could otherwise constitute social pressure to engage in risk-taking behaviors, arrangements could be made to engage young drivers in dialogues with seriously affected victims of road crashes in which young drivers played a key role. These might include seriously injured victims, bereaved close family members, or contrite “perpetrators.” Such encounters need to be managed so that young drivers are able to identify with the “victim” in each case and are encouraged to imagine how they could be involved in an incident similar to that resulting in the victim’s situation (McKenna & Albery, 2001). Furthermore, young drivers should be provided with the necessary competencies to avoid such an outcome happening to them. As noted previously, brain development is partly shaped by experience; therefore, as one component of enhancing risk perception, adolescent drivers need to be exposed to some of the consequences of risk taking. An issue of likely continuing controversy is the age at which a young person should be eligible to obtain a driver’s license. From the neuroscience findings described in this chapter, when could a young person be deemed ready to become a driver? Waiting until all brain regions and neural networks have fully matured in both sexes, which does not occur until the late 20s, is unrealistic. Assuming that the starting age for driving on public roads is 16 years, given the evidence reviewed here, what developmentally defensible options might reasonably be imposed? Although complete evidence for a definitive answer to this question is still unavailable, general guidelines might be suggested, which happen to be broadly consistent with developing practice with respect to graduated licensing programs in several jurisdictions. During the stage at which major
120
PART | II
neuronal reorganization is occurring, from approximately ages 16 to 21 years (identified as adolescence in this chapter), strategies might reasonably be aimed at hastening those aspects of neural development that occur naturally and that are likely to be critical to safe driving performance. Psychologically, these include enhancing risk perception (e.g., by improving scanning techniques through simulator training), teaching impulse control (e.g., using mentoring), and moderating risk-taking behaviors (e.g., through exposure to road injury victims and training in judging probabilities of harm). These strategies, which might either be incorporated within standard novice driver training or delivered as additional modules, would need to be complemented by appropriate regulatory controls. These might include limiting peer passenger numbers, being accompanied by an experienced older driver, zero tolerance for alcohol and other drugs, nighttime driving restrictions, and
Key Variables to Understand in Traffic Psychology
eliminating in-vehicle distractions, particularly cell phones. These restrictions (other than cell phone use, which should be banned during driving) could be gradually relaxed between the ages of 22 and 25 years, after which the emerging adult could be eligible for an unrestricted driver’s license. If the scientific evidence becomes overwhelming in terms of enhanced driving safety for introducing further aspects of graduated driver licensing programs, then gradual introduction of measures based on the research could gain wider social acceptance, as happened with the generic safety benefits of seat belt use and enforcement of blood alcohol limits. Table 9.1 summarizes key neuroscience research findings to date, potential driving behavior outcomes, and possible ameliorative strategies, many of which are already used or under consideration in a number of jurisdictions. The suggested ameliorative strategies are broadly
TABLE 9.1 Summary of Key Developmental Neuroscience Findings, Potential Effects on Driving, and Possible Ameliorative Strategies Neuroscience findings on the adolescent developing brain
Potential cognitive/behavioral consequences for younger drivers
Possible ameliorative/ counter strategies
Shifting balance between limbic and cortical brain areas (e.g., right ventral striatum is less active in teenage brains).
Behaviors likely to be driven by more extreme incentives: reward seeking, novelty seeking, sensation seeking, risk taking, and “recklessness” (e.g., speeding and tailgating).
Provide opportunities for these motivated behaviors in “safe” (e.g., adequately supervised) environments for young drivers. Provide relevant feedback on potential consequences of behaviors to enhance learning about adverse consequences.
Differential development of cortical (e.g., prefrontal) and limbic (e.g., amygdala) systems.
Affects ability to use information to make good (e.g., safe) decisions. Processing critical emotions (e.g., fear) in others to adult levels still developing. Peer influence important up to age 25 years.
Provide opportunities for young drivers to practice using information in typical road environments (e.g., using case studies or simulations). Train young drivers to deal with driving scenarios that they could perceive as threatening. Role play scenarios involving potentially adverse impacts of peer influence.
Cerebellum development.
Some postures/movements could be adversely affected. Although younger drivers may appear to learn many skills rapidly, they remain prone to errors arising from coordination lapses.
Ensure that young drivers can practice in environments that are “forgiving” of postural lapsesdthat is, those that do not result in potentially fatal/other serious injury. Provide adequate feedback on performance.
Pineal gland and melatonin production.
Younger drivers may tend to “eveningness,” displaying a preference for activities late in the daily cycle.
Restricting nighttime driving for younger drivers runs counter to what may be a “natural” preference. Encourage supervised practice under nighttime conditions.
Amygdala and hippocampus development.
Integrating emotions and cognitions occurs over an extended period, which could put young drivers in danger (e.g., arising from an underdeveloped ability to handle stress).
Adopt expectations appropriate to what young drivers are able to undertake, particularly when driving under potentially “stressful” conditions.
Chapter | 9
Neuroscience and Young Drivers
121
TABLE 9.1 Summary of Key Developmental Neuroscience Findings, Potential Effects on Driving, and Possible Ameliorative Strategiesdcont’d Neuroscience findings on the adolescent developing brain
Potential cognitive/behavioral consequences for younger drivers
Possible ameliorative/ counter strategies
Corpus callosum, which links the right and left brain hemispheres, does not stabilize until early adulthood.
Young drivers who have “mixed handedness” could be particularly liable to adverse effects from hazards involving complex tasks requiring both hands and more than one sense modality.
Limit the range of allowable activities by young drivers, and enforce strict guidelines on what tasks and activities young drivers can safely perform.
Subcortical regions (e.g., nucleus accumbens) mature before and are more active in adolescents prior to decisions involving risk. Lower levels of VPFC activity and white matter linking circuits and later development of these regions.
Bias to immediate rewards over longer term goals. Decisions with possible risk outcomes more likely for younger drivers.
Test for individual differences in risk-taking propensity. Train younger drivers to understand a range of possible negative outcomes of risk taking during driving. Educate young drivers about the benefits of setting and seeking to meet longer term goals, especially relating to the safety of self and significant others.
Prefrontal cortex develops throughout adolescence; specifically, unneeded neuronal pathways are pruned as higher executive areas mature.
Liability to frustration and errorproneness in driving tasks involving decision making, with potential for “irrational” or disorganized thought patterns/behaviors. Response inhibition may be reduced.
Provide adequate support and guidance (e.g., mentoring) for younger drivers, particularly when undertaking maneuvers involving complex decision making.
Balance between white and gray matter in the frontal lobes of the cortex changes, with gray matter reducing in volume and white matter increasing in volume.
White matter is important in the speed and smoothness of information flows.
Young drivers are increasingly able to develop competence in maneuvers and in dealing with road hazards involving information processing. Appropriate cognitive ability and driving skills tests could be used at intervals to assess and provide feedback on this information-processing ability.
Visuospatial functions associated with parietal lobes mature earlier than executive functions of the frontal brain region (cortex).
Hazards may be perceived as in adults, but risk cognition (understanding the nature of hazards and their potential for harm) lags behind.
Training in risk perception is required to enhance young drivers’ understanding of risks associated with particular road hazards. It cannot be assumed that younger drivers have a complete understanding of the risks associated with driving merely because hazards are visible.
Areas of the brain responsible for creating mental imagery (e.g., insula and right fusiform face) are still developing.
Could result in delays in processing critical information about generically dangerous situations; younger drivers take longer to process such information.
Use danger perception scenarios to train younger drivers in particular aspects of driving to develop mental imagery of undesired outcomes. Develop younger drivers’ “metamemory,” for example, by getting them to give a running commentary on road conditions, particularly potential hazards, during supervised practice to enhance alertness and perceptual abilities.
Amygdala develops later in males; males have less brain tissue available to regulate their emotions.
Younger males are more prone to make aggressive responses to a wide range of situations, which could lead to risk taking.
Mentoring and support from older role models, particularly for young male drivers who are prone to aggressive responses, to ensure that they are unlikely to drive when poor control of aggression or risk taking could compromise safety.
122
consistent with Steinberg’s (2008b) maxim that efforts to reduce adolescent risk taking should focus on changing the context in which potentially risky behaviors occur rather than seeking to change adolescents’ knowledge and thinking. To provide useful further insights, future neuroscience research should be targeted toward driving behavior, in particular establishing more definitive links between brain development and specific components of driving behavior. To date, approximately 200 neuroscience studies have linked brain function with aspects of driving (Glendon, 2010). Future studies in this field would ideally focus on key implications for driving of adolescent brain development. These should be complemented by a more tailored approach to applications, particularly young driver education and training, that is based on relevant neuroscience research. Of particular relevance would be the study of developmental features in this age group and whether addressing these appropriately could attenuate risk-taking behaviors typically associated with younger drivers.
ACKNOWLEDGMENTS For their perceptive and expert comments on an earlier draft of this chapter, I thank my colleagues Graham Bradley, Trevor Hine, and David Shum as well as the volume editor, Bryan Porter.
REFERENCES Anokhin, A. P., Lutzenberger, W., Nikolaev, A., & Birbaumer, N. (2000). Complexity of electrocortical dynamics in children: Developmental aspects. Developmental Psychobiology, 36, 9e22. Arnett, J. J. (1992). Reckless behavior in adolescence: A developmental perspective. Developmental Review, 12, 339e373. Arnett, J. J. (2002). Developmental sources of crash risk in young drivers. Injury Prevention, 8, 17e23. Baird, A. A., & Fugelsang, J. A. (2004). The emergence of consequential thought: Evidence from neuroscience. Philosophical Transactions of the Royal Society: Biological Sciences, 359(1451), 1797e1804. Baird, A. A., Fugelsang, J. A., & Bennett, C. (2005). What were you thinking? An fMRI study of adolescent decision-making. New York: Paper presented at the Cognitive Neuroscience Society Conference. April. Baird, A. A., Gruber, S. A., Fein, D. A., Maas, L. C., Steingard, R. J., Renshaw, P. F., Cohen, B. M., & Yurgelun-Todd, D. A. (1999). Functional magnetic resonance imaging of facial affect recognition in children and adolescents. Journal of the American Academy of Child and Adolescent Psychiatry, 38, 195e199. Barnea-Goraly, N., Menon, V., Eckert, M., Tamm, L., Bammer, R., Karchemskiy, A., Dant, C. C., & Reiss, A. L. (2005). White matter development during childhood and adolescence: A cross-sectional diffusion tensor imaging study. Cerebral Cortex, 15, 1848e1854. Baxter, M. G., Parker, A., Linder, C. C. C., Izquierdo, A. D., & Murray, E. A. (2000). Control of response selection by reinforcer values requires interaction of amygdala and orbital prefrontal cortex. Journal of Neuroscience, 20, 4311e4319.
PART | II
Key Variables to Understand in Traffic Psychology
Benes, F. M. (1994). Development of the corticolimbic system. In G. Dawson, & K. W. Fischer (Eds.), Human behavior and the developing brain (pp. 176e206). New York: Guilford. Benes, F. M., Turtle, M., Khan, Y., & Farol, P. (1994). Myelination of a key zone in the hippocampal formation occurs in the human brain during childhood, adolescence, and adulthood. Archives of General Psychiatry, 51, 477e484. Bjork, J. M., Knutson, B., Fong, G. W., Caggiano, D. M., Bennett, S. M., & Hommer, D. W. (2004). Incentive-elicited brain activation in adolescents: Similarities and differences from young adults. Journal of Neuroscience, 24, 1793e1802. Blakemore, S.-J., & Choudhury, S. (2006). Development of the adolescent brain: Implications for executive function and social cognition. Journal of Child Psychology and Psychiatry, 47, 296e312. Blanton, R. E., Levitt, J. G., Peterson, J. R., Fadale, D., Sporty, M. L., Lee, M., To, D., Mormino, E. C., Thompson, P. M., McCracken, J. T., & Toga, A. W. (2004). Gender differences in the left inferior frontal gyrus in normal children. NeuroImage, 22, 626e636. Branche, C., Williams, A. F., & Feldman, D. D. (2002). Graduated licensing for teens: Why everybody’s doing it. Journal of Law, Medicine & Ethics, 30 (Suppl.3), 146e149. Brown, I. D., & Groeger, J. A. (1988). Risk perception and decision taking during the transition between novice and experienced driver status. Ergonomics, 31, 585e597. Calhoun, V. D., Pekar, J. J., McGinty, V. B., Adali, T., Watson, T. D., & Pearlson, G. D. (2002). Different activation dynamics in multiple neural systems during simulated driving. Human Brain Mapping, 16, 158e167. Calhoun, V. D., Pekar, J. J., & Pearlson, G. D. (2004). Alcohol intoxication effects on simulated driving: Exploring alcohol-dose effects on brain activation using functional MRI. Neuropsychopharmacology, 29, 2097e2107. Carey, S. (1992). Becoming a face expert. Philosophical Transactions of the Royal Society of London B: Biological Sciences, 335, 95e102. Carskadon, M., Acebo, C., Richardson, G., Tate, B., & Seifer, R. (1997). Long nights protocol: Access to circadian parameters in adolescents. Journal of Biological Rhythms, 12, 278e289. Casey, B. J., Galvan, A., & Hare, T. A. (2005). Changes in cerebral functional brain development and its relation to cognitive development. Current Opinion in Neurobiology, 15, 239e244. Casey, B. J., Getz, S., & Galvan, A. (2008). The adolescent brain. Developmental Review, 28, 62e77. Casey, B. J., Tottenham, N., Liston, C., & Durston, S. (2005). Imaging the developing brain: What have we learned about cognitive development? Trends in Cognitive Science, 9, 104e110. Casey, B. J., Trainor, R. J., Orendi, J. L., Schubert, A. B., Nystrom, L. E., Giedd, J. N., Castellanos, F. X., Haxby, J. V., Noll, D. C., Cohen, J. D., Forman, S. D., Dahl, R. E., & Rapoport, J. L. (1997). A developmental functional MRI study of prefrontal activation during performance of goeno-go task. Journal of Cognitive Neuroscience, 9, 835e847. Caviness, V. S., Kennedy, D. N., Richelme, C., Rademacher, J., & Filipek, P. A. (1996). The human brain age 7e11 years: Avolumetric analysis based on magnetic resonance images. Cerebral Cortex, 6, 726e736. Cohen, J. R., Asarnow, R. F., Sabb, F. W., Bilber, R. M., Bookheimer, S. Y., Knowlton, B. J., & Poldrack, R. A. (2010). A unique adolescent response to reward prediction errors. Nature Neuroscience, 13, 669e671.
Chapter | 9
Neuroscience and Young Drivers
Corley, T. (1999). Becoming an employer of choice for Generation X: The elements of the deal. Journal of Career Planning and Employment, 59, 21e26. De Bellis, M. D., Keshavan, M. S., Beers, S. R., Hall, J., Frustaci, K., Masalehdan, A., Noll, J., & Boring, A. M. (2001). Sex differences in brain maturation during childhood and adolescence. Cerebral Cortex, 11, 552e557. Deery, H. A. (1999). Hazard and risk perception among young novice drivers. Journal of Safety Research, 30, 225e236. Deery, H. A., & Fildes, B. N. (1999). Young novice driver subtypes: Relationship to high-risk behavior, traffic accident record, and simulator driving performance. Human Factors, 41, 628e644. Dicks, L. (2005, March 5). The original rebels. New Scientist 39e43. Doherty, S. T., Andrey, J. C., & MacGregor, C. (1998). The situational risks of young drivers: The influence of passengers, time of day and day of week on accident rates. Accident Analysis and Prevention, 30, 45e52. Draganski, B., Gaser, C., Busch, V., Schuierer, G., Bogdahn, U., & May, A. (2004). Neuroplasticity: Changes in grey matter induced by training. Nature, 427(6972), 311e312. Dryfoos, J. G. (1991). Preventing high-risk behavior. American Journal of Public Health, 81, 157e158. Duangpatra, N. K., Bradley, G. L., & Glendon, A. I. (2009). Variables affecting emerging adults’ self-reported risk and reckless behaviors. Journal of Applied Developmental Psychology, 30, 289e309. Durston, S., Hulshoff Pol, H. E., Casey, B. J., Giedd, J. N., Buitelaar, J. K., & van Engeland, H. (2001). Anatomical MRI of the developing human brain: What have we learned? Journal of the American Academy of Child and Adolescent Psychiatry, 40, 1012e1020. Eby, D. W., Molnar, L. J., Kartje, P., St. Louis, R. M., Parow, J. E., Vivoda, J. M., & Neumeyer, A. L. (2007). Older driver self-screening based on health concerns. Volume I: Technical report. Ann Arbor: University of Michigan Transportation Research Institute. Engstro¨m, I. (2003). Passenger influence on young drivers. In L. Dorn (Ed.), Driver behaviour and training (pp. 191e199). Aldershot, UK: Ashgate. Ernst, M., Nelson, E. E., Jazbec, S. P., McClure, E. B., Monk, C. S., Leibenluft, E., Blair, J., & Pine, D. S. (2005). Amygdala and nucleus accumbens in response to receipt and omission of gains in adults and adolescents. NeuroImage, 25, 1279e1291. Ernst, M., Pine, D. S., & Hardin, M. (2006). Triadic model of the neurobiology of motivated behavior in adolescence. Psychological Medicine, 36, 299e312. Eshel, N., Nelson, E. E., Blair, R. J., Pine, D. S., & Ernst, M. (2007). Neural substrates of choice selection in adults and adolescents: Development of the ventrolateral prefrontal and anterior cingulate cortices. Neuropsychologia, 45, 1270e1279. Fields, R. D., & Stevens-Graham, B. (2002). New insights into neurone glia communication. Science, 298, 700e712. Foss, R. D., & Evenson, K. R. (1999). Effectiveness of graduated driver licensing in reducing motor vehicle crashes. American Journal of Preventive Medicine, 16, 47e56. Galvan, A., Hare, T. A., Parra, C. E., Penn, J., Voss, H., Glover, G., & Casey, B. J. (2006). Earlier development of the accumbens relative to orbitofrontal cortex might underlie risk-taking behavior in adolescents. Journal of Neuroscience, 26, 6885e6892. Galvan, A., Hare, T. A., Voss, H., Glover, G., & Casey, B. J. (2007). Risktaking and the adolescent brain: Who is at risk? Developmental Science, 10, F8eF14.
123
Gardner, M., & Steinberg, L. (2005). Peer influence on risk taking, risk preference, and risky decision making in adolescence and adulthood: An experimental study. Developmental Psychology, 41, 625e635. Giedd, J. N. (2004). Structural magnetic resonance imaging of the adolescent brain. Annals of the New York Academy of Sciences, 1021, 77e85. Giedd, J. N. (2008). The teen brain: Insights from neuroimaging. Journal of Adolescent Health, 42, 335e343. Giedd, J. N., Blumenthal, J., Jeffries, N. O., Castellanos, F. X., Liu, H., Zijdenbos, A., Paus, T., Evans, A. C., & Rapoport, J. L. (1999). Brain development during childhood and adolescence: A longitudinal MRI study. Nature Neuroscience, 2, 861e863. Giedd, J. N., Snell, J. W., Lange, N., Rajpakse, J. C., Casey, B. J., Kozuch, P. L., et al. (1996). Quantitative magnetic resonance imaging of human brain development: Ages 4e18. Cerebral Cortex, 6, 551e560. Giedd, J. N., Vaituzis, A. C., Hamburger, S. D., Lange, N., Rajapakse, J. C., Kaysen, D., Vauss, Y. C., & Rapoport, J. L. (1996). Quantitative MRI of the temporal lobes, amygdala, and hippocampus in normal human development: Ages 4-18 years. Journal of Comparative Neurology, 366, 223e230. Giorgio, A., Watkins, K. E., Douaud, G., James, A. C., James, S., De Stefano, N., Matthews, P. M., Smith, S. M., & Johansen-Berg, H. (2008). Changes in white matter microstructure during adolescence. NeuroImage, 39, 52e61. Glendon, A. I. (2005). Young drivers’ attitudes towards risks arising from hazardous driving behaviours. In L. Dorn (Ed.), Driver behaviour and training (pp. 193e206). Aldershot, UK: Ashgate. Glendon, A. I. (2010). Traffic psychology: A state-of-the-art review. Melbourne: Invited keynote address, 27th International Congress of Applied Psychology. July, 12e16. Gogtay, N., Giedd, J. N., Lusk, L., Hayashi, K. L., Greenstein, D., Vaituzis, A. C., Nugent, T. F., III, Herman, D. H., Clasen, L. S., Toga, A. W., Rapoport, J. L., & Thompson, P. M. (2004). Dynamic mapping of human cortical development during childhood through early adulthood. Proceedings of the National Academy of Sciences USA, 101, 8174e8179. Goldstein, J. M., Seidman, L. J., Horton, N. J., Makris, N., Kennedy, D. N., Caviness, V. S., Jr., Faraone, S. V., & Tsuang, M. T. (2001). Normal sexual dimorphism of the adult human brain assessed by in vivo magnetic resonance imaging. Cerebral Cortex, 11, 490e497. Good, C. D., Johnsrude, I., Ashburner, J., Henson, R. N. A., Friston, K. J., & Frackowiak, R. S. J. (2001). Cerebral asymmetry and the effects of sex and handedness on brain structure: A voxel-based morphometric analysis of 465 normal adult human brains. NeuroImage, 14, 685e700. Gro¨n, G., Wunderlich, A. P., Spitzer, M., Tomczak, R., & Riepe, M. W. (2000). Brain activation during human navigation: Gender-different neural networks as substrate of performance. Nature Neuroscience, 3, 404e408. Gur, R. C., Gunning-Dixon, F., Bilker, W. B., & Gur, R. E. (2002). Sex differences in temporo-limbic and frontal brain volumes of healthy adults. Cerebral Cortex, 12, 998e1003. Hallmark, S. L., Veneziano, D. A., Falb, S., Pawlovich, M., & Witt, D. (2008). Evaluation of Iowa’s graduated driver’s licensing program. Accident Analysis and Prevention, 40, 1401e1405. Hare, T. A., Tottenham, N., Davidson, M. C., Glover, G. H., & Casey, B. J. (2005). Contributions of amygdala and striatal activity in emotion regulation. Biological Psychiatry, 57, 624e632.
124
Hare, T. A., Tottenham, N., Galvan, A., Voss, H. U., Glover, G. H., & Casey, B. J. (2008). Biological substrates of emotional reactivity and regulation in adolescence during an emotional goeno go task. Biological Psychiatry, 63, 927e934. Harre´, N. (2000). Risk evaluation, driving and adolescents: A typology. Developmental Review, 20, 206e226. Hedlund, J. (2007). Novice teen driving: GDL and beyond. Journal of Safety Research, 38, 259e266. Hedlund, J., Shults, R. A., & Compton, R. (2006). Graduated driver licensing and teenage driver research in 2006. Journal of Safety Research, 37, 107e121. Hooper, C. J., Luciana, M., Conklin, H. M., & Yarger, R. S. (2004). Adolescents’ performance on the development of decision making and ventromedial prefrontal cortex. Developmental Psychology, 40, 1148e1158. Igra, V., & Irwin, C. E., Jr. (1996). Theories of adolescent risk-taking behavior. In R. J. DiClemente, W. B. Hansen, & L. E. Ponton (Eds.), Handbook of adolescent health (pp. 35e51). New York: Plenum. Janssen, S., Murre, J., & Meeter, M. (2007). Reminiscence bump in memory for public events. European Journal of Cognitive Psychology, 20, 738e764. Johansen-Berg, H., Della-Maggiore, V., Behrens, T. E., Smith, S. M., & Paus, T. (2007). Integrity of white matter in the corpus callosum correlates with bimanual co-ordination skills. NeuroImage, 36(Suppl. 2), T16eT21. Jonah, B. A. (1997). Sensation seeking and risk driving: A review and synthesis of the literature. Accident Analysis and Prevention, 29, 651e665. Jurkiewicz, C. L. (2000). Generation X and the public employee. Public Personnel Management, 29, 55e74. Keating, D. P. (2007). Understanding adolescent development: Implications for driving safely. Journal of Safety Research, 38, 147e157. Keshavan, M. S., Diwadkar, V. A., De Bellis, M., Dick, E., Kotwal, R., Rosenberg, D. R., Sweeney, J. A., Minshew, N., & Pettegrew, J. W. (2002). Development of the corpus callosum in childhood, adolescence and early adulthood. Life Sciences, 70, 1909e1922. Killgore, W., Oki, M., & Yurgelun-Todd, D. A. (2001). Sex-specific developmental changes in amygdala responses to affective faces. NeuroReport, 12, 427e433. Kuhn, D. (2006). Do cognitive changes accompany developments in the adolescent brain? Perspectives on Psychological Science, 1, 59e67. Kuhnen, C. M., & Knutson, B. (2005). The neural basis of financial risk taking. Neuron, 47, 763e770. LeDoux, J. (1997). The emotional brain: The mysterious underpinnings of emotional life. New York: Simon & Schuster. Lenroot, R. K., & Giedd, J. N. (2006). Brain development in children and adolescents: Insights from anatomical magnetic resonance imaging. Neuroscience and Biobehavioral Reviews, 30, 718e729. Lenroot, R. K., Gogtay, N., Greenstein, D. K., Wells, E. M., Wallace, G. L., Clasen, L. S., Blumenthal, J. D., Lerch, J., Zijdenbos, A. P., Evans, A. C., Thompson, P. M., & Giedd, J. N. (2007). Sexual dimorphism of brain developmental trajectories during childhood and adolescence. NeuroImage, 36, 1065e1073. Leon-Carrion, J., Garcia-Orza, J., & Perez-Santamaria, F. J. (2004). The development of the inhibitory component of the executive functions in children and adolescents. International Journal of Neuroscience, 114, 1291e1311. Lerner, R. (2002). Concepts and theories of human development (3rd ed.). Mahwah, NJ: Erlbaum.
PART | II
Key Variables to Understand in Traffic Psychology
Lin, M.-L., & Fearn, K. T. (2003). The provisional license: Nighttime and passenger restrictionsdA literature review. Journal of Safety Research, 34, 51e61. Liston, C., Watts, R., Tottenham, N., Davidson, M. C., Niogi, S., Ulug, A. M., & Casey, B. J. (2006). Frontostriatal microstructure modulates efficient recruitment of cognitive control. Cerebral Cortex, 16, 553e560. Luciana, M., Conklin, H. M., Cooper, C. J., & Yarger, R. S. (2005). The development of nonverbal working memory and executive control processes in adolescents. Child Development, 76, 697e712. Luders, E., Narr, K. L., Thompson, P. M., Woods, R. P., Rex, D. E., Jancke, L., Steinmetz, H., & Toga, A. W. (2005). Mapping cortical gray matter in the young adult brain: Effects of gender. NeuroImage, 26, 493e501. Luna, B., Garver, K. E., Urban, T. A., Lazar, N. A., & Sweeney, J. A. (2004). Maturation of cognitive processes from late childhood to adulthood. Child Development, 75, 1357e1372. Luna, B., & Sweeney, J. A. (2004). The emergence of collaborative brain function: fMRI studies of the development of response inhibition. Annals of the New York Academy of Sciences, 1021, 296e309. Mackie, S., Shaw, P., Lenroot, R., Pierson, R., Greenstein, D. K., Nugent, T. F., Sharp, W. S., Giedd, J. N., & Rapoport, J. L. (2007). Cerebellar development and clinical outcome in attention deficit hyperactivity disorder. American Journal of Psychiatry, 164, 647e655. Marin, P. S., & Brown, B. V. (2005). Are teens driving safer? Child Trends, 4, 1e10. Matthews, S. C., Simmons, A. N., Lane, S. D., & Paulus, M. P. (2004). Selective activation of the nucleus accumbens during risk-taking decision making. NeuroReport, 15, 2123e2127. May, J. C., Delgado, M. R., Dahl, R. E., Stenger, V. A., Ryan, N. D., Fiez, J. A., & Carter, C. S. (2004). Event-related functional magnetic resonance imaging of reward-related brain circuitry in children and adolescents. Biological Psychiatry, 55, 359e366. Maycock, G. (2002). Novice driver accidents and the driving test (Report No. 527). Crowthorne, UK: Transport Research Laboratory. Mayhew, D. R., Simpson, H. M., & Pak, A. (2003). Changes in collision rates among novice drivers during the first months of driving. Accident Analysis and Prevention, 35, 683e691. McGivern, R. F. (2002). Cognitive efficiency on a match to sample task decreases at the onset of puberty in children. Brain and Cognition, 50, 73e89. McKenna, F. P., & Albery, I. P. (2001). Does unrealistic optimism change following a negative experience? Journal of Applied Social Psychology, 31, 1146e1157. McKnight, A. J., & McKnight, A. S. (2003). Young novice drivers: Careless or clueless? Accident Analysis and Prevention, 35, 921e925. Moffitt, T. E. (1993). Adolescence-limited and life-course-persistent antisocial behavior: A developmental taxonomy. Psychological Review, 200, 674e701. Monk, C. S., McClure, E. B., Nelson, E. E., Zarahn, E., Bilder, R. M., Leibenluft, E., Charney, D. S., Ernst, M., & Pine, D. S. (2003). Adolescent immaturity in attention-related brain engagement to emotional facial expressions. NeuroImage, 20, 420e428. Montague, P. R., & Berns, G. (2002). Neural economics and biological substrates of valuation. Neuron, 36, 265e284.
Chapter | 9
Neuroscience and Young Drivers
Nagel, B. J., Medina, K. L., Yoshii, J., Schweinsburg, A. D., Moadab, I., & Tapert, S. (2006). Age-related changes in prefrontal white matter volume across adolescence. NeuroReport, 17, 1427e1431. Neyens, D. M., Donmez, B., & Boyle, L. N. (2008). The Iowa graduated driver licensing program: Effectiveness in reducing crashes of teenage drivers. Journal of Safety Research, 39, 383e390. Nopoulos, P., Flaum, M., O’Leary, D., & Andreasen, N. C. (2000). Sexual dimorphism in the human brain: Evaluation of tissue volume, tissue composition and surface anatomy using magnetic resonance imaging. Psychiatry Research: Neuroimaging Section, 98, 1e13. O’Doherty, J. P. (2004). Reward representation and reward-related learning in the human brain: Insights from neuroimaging. Current Opinions in Neurobiology, 14, 769e776. Overman, W. H. (2004). Sex differences in early childhood, adolescence, and adulthood on cognitive tasks that rely on orbital prefrontal cortex. Brain and Cognition, 55, 134e147. Panksepp, J. (1998). Affective neuroscience: The foundation of human and animal emotions. Oxford: Oxford University Press. Panksepp, J. (2001). The neuro-evolutionary cusp between emotions and cognitions: Implications for understanding consciousness and the emergence of a unified mind science. Evolution and Cognition, 7, 141e163. Paus, T. (2005). Mapping brain maturation and cognitive development during adolescence. Trends in Cognitive Sciences, 9, 60e68. Pujol, J., Lo´pez, A., Deus, J., Cardoner, N., Vallejo, J., Capdevila, A., & Paus, T. (2002). Anatomical variability of the anterior cingulate gyrus and basic dimensions of human personality. NeuroImage, 15, 847e855. Reyna, V. F., & Farley, F. (2006). Risk and rationality in adolescent decision making: Implications for theory, practice, and public policy. Psychological Science in the Public Interest, 7, 1e44. Santesso, D. L., & Segalowitz, S. J. (2008). Developmental differences in error-related ERPs in middle- to late-adolescent males. Developmental Psychology, 44, 205e217. Shaw, P., Greenstein, D., Lerch, J., Clasen, L., Lenroot, R., Gogtay, N., Evans, A., Rapoport, J., & Giedd, J. (2006). Intellectual ability and cortical development in children and adolescents. Nature, 440, 676e679. Shope, J. T., & Molnar, L. J. (2003). Graduated driver licensing in the United States: Evaluation results from the early programs. Journal of Safety Research, 34, 63e69. Silveri, M. M., Rohan, M. L., Pimentel, P. J., Gruber, S. A., Rosso, I. M., & Yurgelun-Todd, D. A. (2006). Sex differences in the relationship between white matter microstructure and impulsivity in adolescents. Magnetic Resonance Imaging, 24, 833e841. Simons-Morton, B. G. (2002). Reducing young driver crash risk. Injury Prevention, 8(Suppl. 2), ii1eii38. Simons-Morton, B. G., Lerner, N., & Singer, J. (2005). The observed effects of teenage passengers on the risky driving behavior of teenage drivers. Accident Analysis and Prevention, 37, 973e982. Smith, S., Horswill, M., Chambers, B., & Wetton, M. (2009). Sleepiness and hazard perception while driving (Road Safety Grant Report No. 2009-001). Canberra: Australian Government, Department of Infrastructure, Transport, Regional Development and Local Government. Sowell, E. R., Peterson, B. S., Thompson, P. M., Welcome, S. E., Henkenius, A. L., & Toga, A. W. (2003). Mapping cortical change across the life span. Nature Neuroscience, 6, 309e315.
125
Sowell, E. R., Thompson, P. M., Holmes, C. J., Jernigan, T. L., & Toga, A. W. (1999). In vivo evidence for post-adolescent brain maturation in frontal and striatal regions. Nature Neuroscience, 2, 859e861. Sowell, E. R., Thompson, P. M., Tessner, K. D., & Toga, A. W. (2001). Mapping continued brain growth and gray matter density reduction in dorsal frontal cortex: Inverse relationships during postadolescent brain maturation. Journal of Neuroscience, 21, 8819e8829. Sowell, E. R., Thompson, P. M., & Toga, A. W. (2004). Mapping changes in the human cortex throughout the span of life. Neuroscientist, 10, 372e392. Spear, L. P. (2000). The adolescent brain and age-related behavioral manifestations. Neuroscience and Biobehavioral Reviews, 24, 417e463. Steinberg, L. (2008a). Adolescence (8th ed.). Boston: McGraw-Hill. Steinberg, L. (2008b). A social neuroscience perspective on adolescent risk-taking. Developmental Review, 28, 78e106. Steinberg, L., Cauffman, E., Woolard, J., Graham, S., & Banich, M. (2009). Are adolescents less mature than adults? Minors’ access to abortion, the juvenile death penalty, and the alleged APA “flip-flop.”. American Psychologist, 64, 583e594. Steinberg, L., & Scott, E. S. (2003). Less guilty by reason of adolescence: Developmental immaturity, diminished responsibility and the juvenile death penalty. American Psychologist, 58, 1009e1018. Suzuki, M., Hagino, H., Nohara, S., Zhou, S.-Y., Kawasaki, Y., Takahashi, T., Matsui, M., Seto, H., Ono, T., & Kurachi, M. (2005). Male-specific volume expansion of the human hippocampus during adolescence. Cerebral Cortex, 15, 187e195. Taylor, M. J., McCarthy, G., Saliba, E., & Degiovanni, E. (1999). ERP evidence of developmental changes in processing of faces. Clinical Neurophysiology, 110, 910e915. Thomas, K. M., Drevets, W. C., Whalen, P. J., Eccard, C. H., Dahl, R. E., Ryan, N. D., & Casey, B. J. (2001). Amygdala response to facial expressions in children and adults. Biological Psychiatry, 49, 309e316. Toga, A. W., Thompson, P. M., & Sowell, E. R. (2006). Mapping brain maturation. Trends in Neuroscience, 29, 148e159. Watkins, K., Paus, T., Mangin, J.-F., Zijdenbos, A., Collins, L., Lerch, J., Worsley, K., Blumenthal, J., Giedd, J. N., & Rapoport, J. (2002). Maturation of white matter tracts during adolescence: A longitudinal MRI study. NeuroImage, 16(Suppl.), 1718e1719. Whitford, T. J., Rennie, C. J., Grieve, S. M., Clark, C. R., Gordon, E., & Williams, L. M. (2007). Brain maturation in adolescence: Concurrent changes in neuroanatomy and neurophysiology. Human Brain Mapping, 28, 228e237. Williams, A. F. (2003). Teenage drivers: Patterns of risk. Journal of Safety Research, 34, 5e15. Wolman, D. (2005). A left-hand turn around the world: Chasing the mystery and meaning of all things southpaw. Cambridge, MA: Da Capo Press. Yurgelun-Todd, D. A. (2007). Emotional and cognitive changes during adolescence. Current Opinion in Neurobiology, 17, 251e257. Zemke, R., Raines, C., & Filipczak, B. (2000). Generations at work: Managing the clash of veterans, boomers, Xers, and Nexters in your workplace. Washington, DC: American Management Association.
This page intentionally left blank
Chapter 10
Neuroscience and Older Drivers Maria T. Schultheis and Kevin J. Manning Drexel University, Philadelphia, PA, USA
1. NEUROSCIENCE AND OLDER DRIVERS Cognitive neuroscience of aging is a multifaceted discipline that encompasses clinical neuropsychology, cognitive neuroscience, and cognitive aging (Cabeza, Nyberg, & Park, 2005; Grady, 2008). Advances in technology, such as structural and functional neuroimaging, have contributed a growing body of literature that has better defined the overall changes in the aging brain. Specifically, research has been extremely informative regarding the age-related differences (i.e., cross-sectional research) and age-related changes (i.e., longitudinal research) in brain structure and function.
1.1. The Effect of Aging on Neuroanatomy and Cognition As individuals age, physiological changes to the brain occur. In general, brain volume is reduced through atrophy and subsequent enlargement of the cerebral ventricles, or ventriculomegaly (Raz, Gunning-Dixon, Head, Dupuis, & Acker, 1998). Specifically, cross-sectional studies of normative cohorts reveal that the majority of brain structures demonstrate reduced volume, including the cerebral gray (i.e., neurons) and white (i.e., axons) matter, as well as subcortical structures, such as the hippocampus and major elements of the basal ganglia (Raz & Rodrigue, 2006). The rate of ventricular expansion and shrinkage of the total brain parenchyma (atrophy) accelerates with age and appears to follow a linear course. The volume of the white matter, especially in the prefrontal regions (Raz et al., 2005), follows a nonlinear longitudinal course, with linear increase until young adulthood, plateau during middle age, and decline in later years. This overall cerebral atrophy of gray and white matter is thought to explain much of the cognitive changes seen in all adults as they age. For example, it is hypothesized that atrophy of the brain’s frontal lobes’ gray matter and its surrounding white matter connections may be the underlying neuropathology of observed mild memory difficulties in aging. Behaviorally, this mild impairment may be most Handbook of Traffic Psychology. DOI: 10.1016/B978-0-12-381984-0.10010-4 Copyright Ó 2011 Elsevier Inc. All rights reserved.
apparent on tasks demanding high levels of attention and executive functioning. An overall “slowing” of behavior is a universal description associated with aging (Salthouse, 1996), which has been related to white matter changes (Gunning-Dixon & Raz, 2000) in the brains of healthy adults. In particular, among normal aging adults, atrophy of the frontal lobes has been a robust and consistent finding (Haug & Eggers, 1991; Raz, et al., 1998). Researchers have suggested that this change in frontal lobe physiology may lead to subtle changes in inhibitory control, leading to observed declines in performance on tests of executive function (i.e., problem solving and decision making). In addition to general “executive functioning decline,” working memory, another important cognitive construct associated with the frontal lobes, has also been shown to decline among normal aging adults. Of note, working memory serves as an important link to slowed processing speed (Gunning-Dixon & Raz, 2000). From a broader and simpler perspective, cognitive functions can be conceptualized and categorized into two general aspects: (1) crystallized intelligence, which includes overlearned, familiar skills accumulated through education and practice, and (2) fluid intelligence, which includes nonverbal reasoning, motor tasks, and problemsolving abilities that evolve and change as a result of physiologic maturation. It has been theorized that crystallized abilities, such as knowledge of general facts and vocabulary, sharply increase during the early years of formal education and then stabilize or gradually improve throughout adulthood. By contrast, fluid abilities are theorized to improve throughout childhood and then gradually decline in adult years, with more rapid deterioration in old age due to neuronal loss, changes in physiologic brain function, and increased rates of disease and injury. Research employing both cross-sectional and longitudinal designs has supported the relative stability of verbal abilities with advancing age (i.e., crystallized) and the decline in tasks requiring perceptual speed, selective attention, and complex problem solving (i.e., fluid) (Tucker-Drobb & Salthouse, 2008). 127
128
1.2. Age-Related Changes in Cognition Relevant to Driving A clinical priority of research on driving with older adults is the identification of deficits associated with specific diagnosis (e.g., Alzheimer’s disease). However, it can be argued that diagnosis alone does not render an individual unfit to drive. Certainly, the driving of individuals with a neurodegenerative illness should be monitored because their cognitive abilities will change and/or decline to the level that may prohibit driving. However, as evidenced by studies examining healthy older adults, cognitive abilities necessary for safe driving can be disrupted in older adults without dementia as well. In an analysis of moderately cognitively impaired adults with an average age of 76 years (Average Mini-Mental State Exam ¼ 25; range, 14e30), performance on a test of clock drawing highly correlated with total number of driving errors using a driving simulator (r ¼ 0.68) (Freund, Gravenstein, Ferris, Burke, & Shaheen, 2005). The authors hypothesized that this may be “because executive functioning is a critical component of safe driving, in the presence of executive dysfunction, the automatized and procedural skills learned over decades of daily living do not protect the older driver from errors” (p. 243). Hypothetically, subtle executive changes in a more cognitively intact group may show a stronger relationship with cognitively demanding driving tasks. Other authors have reported data suggesting a relationship between executive functioning and driving. Whelihan, DiCarlo, and Paul (2005), using a mixed sample of older adults with questionable dementia and brain injury, found that out of a comprehensive neuropsychological battery, only performance on a maze navigation test, time to complete Trail Making Test Part B, and the Useful Field of View (a measure of visual attention) correlated with driving ability as measured by a road test. Together, these three measures explained 46% of the variance in a total composite of the road test (Whelihan et al., 2005). Ott et al. (2003) found maze performance to be predictive of driving ability; this was the sole measure from a comprehensive battery of tests to be associated with caregiver ratings of driving performance in individuals with Alzheimer’s disease (AD). Finally, Daigneault, Joly, and Frigon (2002) found that older adults with a history of accidents were more impaired on four measures of executive functioning: Trail Making Test, Wisconsin Card Sorting, Stroop Color Word, and Tower of London. Taken together, these findings suggest that normally occurring changes in cognitive status may be relevant to driving performance. In particular, cognitive functions often grouped as “executive functions” consistently appear to be important; these include cognitive domains such as information process speed, working memory, decision
PART | II
Key Variables to Understand in Traffic Psychology
making, and visual problem solving. Although medically diagnosed older adults may demonstrate more significantly impaired levels of these domains, the fact remains that driving capacity appears to be affected when changes in these domains occur. The increasing amount of evidence supporting the relationship between aging and cognitive deterioration has led to policy discussions about mandatory aged-based and disorder-based assessments. To further explore reception for this, Adler and Rottunda (2010) investigated the attitudes, beliefs, and preferences of older adults, law enforcement officers, and licensing authorities toward reexamination of driving skills for people with AD and Parkinson’s disease (PD) and at varying ages. The results indicated strongest support across all groups for retesting for those with AD but only moderate endorsement of retesting for those with PD. Moderate support was also given for re-testing of 90-year-old drivers, and the least support was given for reassessment of 70-year-old drivers (Adler & Rottunda, 2010).
1.3. What are the Characteristics of Older Drivers? In the literature, driving performances have been defined using a variety of measures, including traffic crashes and the behind-the-wheel (BTW) exam. Population-based studies of driving in older adults have typically used number of traffic crashes as the outcome variable representing driving performance (Langford, 2008). McGwin and Brown (1999) compared the characteristics of crashes among young, middle-aged, and drivers older than age 55 years from all of the police-reported traffic crashes in the state of Alabama during 1996. Compared to young and middle-aged drivers, older drivers were more likely to be involved in crashes at intersections, fail to yield the right of way, and fail to heed stop signs or signals. Crashes occurring while turning and changing lanes were also more common among older drivers. By contrast, older drivers were less likely to have crashes during adverse weather and while traveling at high speeds. Other studies have reported similar crash characteristics among older adults (Cooper, 1990; Hakamies-Blomqvist, 1993; Ryan, Legge, & Rosman, 1998). A plethora of research demonstrates the significant association between crash risk in older adults and cognitive test performance on tests of attention, memory, executive functioning, and processing speed (Ball, Owsley, Sloane, Roenker, & Bruni, 1993; Lafont, Laumon, Helmer, Dartigues, & Fabrigoule, 2008; Staplin, Gish, & Wagner, 2003; Stutts, Stewart, & Martell, 1998). Despite statistical significance, the clinical significance of the association between cognitive test performance and crash risk is poor. For example, Carr, Duchek, and Morris (2000) compared
Chapter | 10
Neuroscience and Older Drivers
63 healthy older adults and 58 individuals with clinically diagnosed AD and found no group difference in the number of accidents 5 years prior to the onset of the study. In other words, despite impairment in various domains of functional capacity severe enough to warrant a diagnosis of AD, adults with the disease would show no differences in functional capacity if functional capacity were measured using only driving accidents. The point is that although the statistical association between cognitive test performance and crash risk may be beyond a chance possibility, it provides little clinical value in determining individuals who should limit or relinquish their driving privileges (Bedard, Weaver, Darzins, & Porter, 2008). This is likely due to the rarity of accidents and the fact that they represent a heterogeneous collection of incidents occurring in a multitude of circumstances (McGwin & Brown, 1999). More sensitive measures of driving capacity are needed to distinguish older adults with obvious gross functional and cognitive impairment from healthy controls. Currently, the clinical “gold standard” of driving ability is the BTW driving evaluation. The BTW consists of route following, or the examiner directing the examinee where to drive next, and is similar to the “road test” that most individuals undergo to receive their driver’s license. However, when considering its utility as an outcome measure, similar to “number of traffic crashes,” the BTW lacks sensitivity to detect subtle changes in driving performance. Kay, Bundy, Clemson, and Jolly (2008) made this point in their investigation of the psychometric properties of a standardized BTW with 100 cognitively intact older drivers aged 60e86 years. Although the authors found total driving errors and overall ratings of performance (i.e., pass/fail) to be valid and reliable indicators of driving safety, these scoring systems were not sensitive enough to determine different levels of driving ability or even discriminate “safe” versus “unsafe” drivers (Kay et al., 2008). Similar findings have been found for older adults with dementia. Ott et al. (2008) conducted a longitudinal study of drivers with AD spanning 3 years using the BTW. Greater severity of dementia, increased age, and lower education were associated with higher rates of BTW failure at follow-up. However, only 22% of individuals with mild AD failed the BTW at followup. The failure rate was even less in the group of individuals considered to have questionable dementia or mild cognitive impairment. On a practical level, in an attempt to harness this research information for clinical applications, the American Medical Association (2010) compiled physician guidelines for the assessment and counseling of older drivers. These guidelines include recommendations for specific tests for the Assessment of Driving-Related Skills (ADReS), including visual measures (i.e., visual acuity and visual fields) and basic cognition (Trail Making A and B Test and Clock Drawing Test).
129
2. MEDICAL ISSUES AND OLDER DRIVERS 2.1. Driving and the Dementias Without doubt, the dementing conditions are among the most problematic faced by aging drivers. In addition to affecting cognitive and visuomotor abilities that can impact everyday activities such as driving, the dementing illnesses can also deprive individuals of the judgment and insight needed to accurately assess their own declining abilities and increase risk due to these deficits. At the same time, many dementing conditions are progressive and tend to be insidious, making detection more difficult for patients, families, and health care professionals. Alzheimer’s disease is the most common cause of dementia and the sixth leading cause of death. Women are more likely than men to have AD, and it has been estimated to affect 14% of all people aged 71 years or older. AD is a steadily progressive disorder that is characterized by a variety of cognitive function abnormalities. By definition, diagnosis of dementia requires the objective measurement of impairments in memory and one other cognitive domain that is resulting in a negative impact on occupational or social functioning (American Psychiatric Association, 2000). One of the common activities of daily functioning that is included is driving an automobile. Competence in driving a motor vehicle has implications both for the safety of the individual affected by dementia and for other road users. Therefore, it is not uncommon for health care providers to have to determine whether individuals with dementia are able to continue to drive and/or when they should stop driving. Retrospective surveys concerning driving and dementia suggest that many patients diagnosed with dementia do continue to drive and may be reluctant to give up driving (Friedland et al., 1988; Gilley et al., 1991). In 2000, Dubinsky, Stein, and Lyons reported an eightfold increase in the crash rate for a group with AD, implying a greater risk of crashes for drivers with AD compared to other drivers. It is also noteworthy that two early retrospective studies found that only 50% of drivers with AD had ceased driving within a 3-year period of the onset of dementia (Drachman & Swearer, 1993; Friedland et al., 1988), after which time crash risk increases substantially. Tuokko, Tallman, Beattie, Cooper, and Weir (1995) examined driving records (insurance claims) of 165 drivers with dementia and found that they had an approximately 2.5 times higher crash rate than that of the matched control sample. In contrast, in a study using state records, road crash and violation rates among AD patients did not differ significantly from those of matched controls (Trobe, Waller, Cook-Flannagan, Teshima, & Bieliauskas, 1996). However, this study did not control for mileage driven, and reduced driving exposure of AD patients may be the reason why their crash rate was equal to that of control subjects. Carr and
130
colleagues (2000) reported that a sample of 63 drivers with very mild or mild Clinical Dementia Rating (Hughes, Berg, Danziger, Coben, & Martin, 1982) showed no difference in state recorded crash rate for the previous 5-year period compared to nondemented, older control drivers, even after adjusting for exposure. Carr et al. noted that the drivers with dementia in their study may have been only mildly impaired in their driving skills, with little, if any, impairment in driving skills evident in the preceding 5-year period. It has been recommended that limitation of driving privileges should be based on demonstration of impaired driving competence rather than on a clinical diagnosis such as AD. As early as 1988, Drachman and colleagues argued that individuals with an AD diagnosis should not be excluded from driving based on the possibility of minimal functional decline in early AD. They also suggested that the likelihood of the loss of driving privileges may result in many people with mild or potentially treatable cognitive impairments refraining from seeking medical advice about continuing to drive. O’Neill et al. (1992) based their argument on findings of studies that demonstrated that a substantial percentage of patients with AD at the time of driving assessment had suffered no deterioration in driving skills, thus supporting the view that a diagnosis of AD alone is not sufficient to preclude driving. Indeed, researchers continue to attempt to define the rate of AD progression and nature of disease manifestation, particularly in the earlier stages; therefore, using a diagnosis of AD as a basis for a decision regarding driving is not recommended. In 2000, the American Academy of Neurology published a practice parameter regarding AD and driving (Dubinsky et al., 2000). Two recommendations were made. First, drivers with AD who have a Clinical Dementia Rating (CDR) of 1.0 or greater should not drive because of driving performance errors and a substantially increased accident rate. Second, drivers with possible AD with a severity of CDR of 0.5 should be considered for referral for driving performance evaluation. Furthermore, because of the high likelihood of disease progression, it was recommended that dementia severity and appropriateness of continued driving be reassessed every 6 months. This follow-up recommendation was evaluated in a prospective longitudinal study of 58 healthy controls, 21 individuals with very mild AD, and 29 individuals with mild AD. In this study, participants underwent a standardized on-road test approximately every 6 months for a 3-year period. Analysis of the survival curves generated for each group supported the recommendation to conduct driver evaluations every 6 months for people with very mild and mild dementia of the Alzheimer’s type (Duchek et al., 2003). Although helpful, there remains limited follow-up to the application of these initial guidelines, and additional longitudinal studies are needed to better describe the progression of AD and its subsequent effect on driving ability.
PART | II
Key Variables to Understand in Traffic Psychology
A 2010 update of these practice parameters attempted to provide more specific guidelines for clinicians (Iverson et al., 2010). The authors recommended that for individuals with dementia, (1) consideration of the CDR scale, (2) a caregiver’s rating of a patient’s driving ability as marginal or unsafe, (3) a history of crashes or traffic citations, (4) reduced driving mileage or self-reported situational avoidance, (5) Mini-Mental State Examination scores of 24 or less, and (6) aggressive or impulsive personality characteristics are useful for identifying patients at increased risk for unsafe driving. Although informative, the study excluded much of the work with neuropsychological testing and subsequently did not support or refute the contributions of cognitive testingdan important limitation that minimizes the relevance of cognitive status in this population. Although these studies serve to provide clinicians with some guidelines, potential difficulties with implementation of these parameters have been noted. For example, the AD patient or his or her family may not accept the physician’s recommendation for discontinuation of driving. Thus, although physicians have a significant responsibility to determine “medical” competence to drive, in practice such a clinical decision is difficult because of lack of standards and effective guidelines.
2.2. Other Dementias Other less frequently occurring dementias include illnesses such as PD and Huntington’s disease (HD). Much less is known about the relationship between these disorders and driving capacity. However, given some commonality (particularly in the cognitive domains), the use of similar strategies for evaluating driving that are employed for AD have been recommended.
2.2.1. Parkinson’s Disease This common disease, known since ancient times, was first clinically described by James Parkinson in 1817. The disease generally begins at 40e70 years of age, with the peak age of onset in the sixth decade. It is infrequent before 30 years of age, and most series contain a somewhat larger proportion of men. The core syndrome is one of expressionless face, poverty and slowness of voluntary movement, “resting” tremor, stooped posture, axial instability, rigidity, and festinating gait. Although most are familiar with the motor effects of the disease, cognitive decline may also be seen. Patients therefore not only experience a progressive loss of motor control but also eventually are at risk for cognitive and emotional deterioration. Cognitive symptoms include slowed information processing, executive dysfunction, memory loss, and associated personality changes (Aarsland, Bronnick, Larsen, Tysnes, & Alves, 2009; Rodriguez-Oroz et al., 2009).
Chapter | 10
Neuroscience and Older Drivers
Work by Uc and colleagues has provided a greater understanding of the driving ability of individuals with PD (Rizzo, Uc, Dawson, Anderson, & Rodnitzky, 2010). In 2009, Uc and colleagues compared the performance of licensed drivers with PD with that of an age-matched control group. They found that overall, drivers with PD had poorer road safety compared to controls, but there was considerable variability among the drivers with PD, and some performed normally. Familiarity with the driving environment was a mitigating factor against unsafe driving in PD. Impairments in visual perception and cognition (attention, visuospatial, and visual memory) were associated with road safety errors in drivers with PD (Uc, Rizzo, Johnson, et al., 2009). Another study examining driving ability found that drivers with PD made more safety errors than did neurologically normal drivers during a route-following task. The authors concluded that the PD group driver safety was degraded due to an increase in the cognitive load in patients with limited reserves. Driving errors and lower driver safety were also associated more with impairments in cognitive and visual function than with the motor severity of the disease in drivers with PD (Uc et al., 2006). This group of researchers also employed the use of driving simulation to better delineate the driving errors seen in PD. Using this methodology, they concluded that under low-contrast visibility conditions, drivers with PD had poorer vehicle control and were at higher risk for crashes (Uc, Rizzo, Anderson, et al., 2009), and that the quantitative effect of an auditoryeverbal distracter task on driving performance was not significantly different between PD and control groups. However, a significantly larger subset of drivers with PD had worsening of their driving safety errors during distraction (Uc et al., 2006). Across the various studies, cognitive predictors of driving performance included visual processing speed and attention, motion perception, contrast sensitivity, visuospatial construction, motor speed, and Activities of Daily Living score.
2.2.2. Huntington’s Disease This disease, distinguished by the triad of dominant inheritance, choreoathetosis, and dementia, derives its eponym from George Huntington (1872). Although relatively rare, in university hospital centers this is one of the most frequently observed types of hereditary nervous system disease. The usual age of onset is in the fourth and fifth decades, but 3e5% of cases begin before the 15th year and some even in childhood. In 28% of cases, symptoms become apparent after 50 years. The progression of the disease is slower in older patients. Once begun, the disease progresses relentlessly. The personality and psychiatric changes associated with HD assume several subtle forms long before the deterioration of cognitive functions becomes evident. In approximately
131
half of the cases, alterations of character are the first symptoms. Patients begin to find fault with everything; they may be suspicious, irritable, impulsive, eccentric, untidy, or excessively religious; or they may exhibit a false sense of superiority. Poor self-control may be reflected in outbursts of temper, fits of despondency, alcoholism, or sexual promiscuity. Disturbances of mood, particularly depression, are common and may constitute the most prominent symptoms early in the disease. Eventually, other cognitive functions deteriorate, and the patient becomes less communicative and more socially withdrawn. Diminished work performance, inability to manage household responsibilities, disturbances of sleep, difficulty in maintaining attention, impaired concentration, deficits in learning new material, and mental rigidity become apparent, along with loss of fine manual skills. Because memory performance benefits from cues to help with retrieval of information, HD has been characterized as a “subcortical dementia.” Increased deterioration of motor functions and chorea (a relatively ceaseless occurrence of a wide variety of rapid, highly complex, jerky movements that appear to be coordinated but are in fact involuntary) usually follow. To date, only one study has empirically assessed the influence of the neurological and cognitive impairments of HD on automobile driving (Rebok, Bylsma, Keyl, Brandt, & Folstein, 1995). These authors found that HD patients performed significantly worse than control subjects on driving-simulator tasks and were more likely to have been involved in a collision in the preceding 2 years (58% of HD patients vs. 11% of control subjects). Patients with collisions were less functionally impaired but had slower simple reaction time scores than did those without collisions. Although additional research is needed, to date there is a presumption that such patients will eventually cease driving as this terminal disease progresses. It is remarkable that other systematic studies examining driving performance in this population are lacking, despite the fact that this is a progressive disease with known cognitive impairments. In particular, difficulties with divided attention, executive functioning, and awareness have all been identified as potential cognitive contributors to driving difficulties in this population. As is the case with other dementias, there is no uniform national law about driving with HD, but several support organizations directly address the topic of driving and offer recommendations for modifying driving behaviors (http://hopes.stanford.edu/ n3547/managing-hd/lifestyle-and-hd/driving-andhuntingtons-disease).
2.3. Cerebral Vascular Accidents or Stroke Cerebral vascular accidents or strokes are the third leading cause of death in the United States. By definition, a stroke
132
occurs as a result of blockage or hemorrhage of a blood vessel leading to the brain. The resulting lack of oxygen supply to the brain results in damage that can manifest in a variety of physical, cognitive, and behavioral deficits for the individual. Common difficulties can include hemiparegia or paralysis of upper and lower extremities, speech difficulties, visual perception and visual spatial difficulties, and changes in cognition (e.g., memory and attention). Not surprisingly, these deficits can have a significant impact on an individual’s activities of daily living and overall quality of life. Given the high value placed on individual transportation in the United States, it is not surprising that many individuals seek to return to driving after experiencing a stroke. In fact, it is estimated that approximately 30e50% of stroke survivors return to driving (Fisk, Owsley, & Mennemeier, 2002; Fisk, Owsley, & Pulley, 1997; Heikkila, Korpelainen, Turkka, Kallanranta, & Summala, 1999). However, it has also been reported that many stroke survivors do not go through any formal evaluation of their driving ability or receive advice before returning to the road. Therefore, the challenge remains in determining how the various sensorimotor and cognitive impairments resulting from stroke may or may not impact the individual’s performance on the road. To date, although no single measurement can be used to definitively calculate an individual’s driving capacity, much has been learned about driving after stroke. It is well documented that stroke can result from different etiologies and can present in significantly varying degrees of severity. As a result, stroke is a major cause of disability, affecting approximately 500,000 individuals annually. Although the highest incidence is reported in older adults, work has identified an increasing number of younger adults who suffer from strokes (Bjorkdahl & Sunnerhagen, 2007). Given this fact, it is not surprising that stroke survivors (both young and older) find that driving cessation interferes with activities related to independent living (e.g., working) and consider the resumption of driving after stroke an important step in their recovery. Long-standing evidence supporting this finding first came from studies that demonstrated that stroke survivors who did not resume driving participated in fewer social activities and were more likely to be depressed (Legh-Smith, Wade, & Hewer, 1986). In addition, one study that focused on driving resumption after mild stroke found that 50% of individuals returned to driving within the first month after experiencing a stroke (Lee, Tracy, Bohannon, & Ahlquist, 2003), further underscoring the need for early assessment. Indeed, accuracy in measuring driving safety after stroke is crucial for ensuring that individuals who are safe drivers are not prevented from maintaining their independent mode of transportation and for preventing individuals who are unsafe drivers from posing a danger to themselves and others. Several studies have documented that of individuals who drove before their stroke, approximately 30e59%
PART | II
Key Variables to Understand in Traffic Psychology
return to driving after their stroke (Fisk et al., 1997; Heikkila et al., 1999). Of those individuals returning to driving, almost one-third report high driving exposure, driving 6 or 7 days per week and/or 100e200 miles per week (Fisk et al., 1997). However, other findings indicate that stroke survivors drive less compared to a nonstroke cohort (Fisk et al., 2002). Specifically, although no differences in days per week of driving were seen, nonstroke drivers drove to more places, took more trips, and drove more miles (Fisk et al., 2002). Drivers who returned to driving also acknowledged difficulties in varying driving situations, such as making left turns, driving on the interstate, and driving in heavy traffic. Despite this, the stroke drivers did not differ from the nonstroke drivers in occurrences of self-reported crashes or citations (Fisk et al., 2002). Overall, stroke drivers appear to be self-regulating their driving behaviors and exposure.
2.3.1. Right Versus Left One of the most common areas of stroke research is evaluating differences in impairment resulting from strokes in the two hemispheres. Several studies have examined the lesion location and the extent of brain damage incurred to better determine the impact of the resulting impairments on driving performance. Cortical damage in the area of the temporoparietal lobe of the right hemisphere often results in impairments in spatial and perceptual abilities and also attentional and visual skills deficits such as visual neglect. Physically, a right-hemisphere stroke can often lead to paralysis of the left side of the body, known as left hemiplegia. In contrast, cortical damage to the left hemisphere often results in language and speech difficulties and paralysis of the right side of the body, known as right hemiplegia. More global cognitive deficits, such as changes in memory and attention, cannot be exclusively associated with one or the other hemisphere. In relation to driving difficulties, several studies have indicated poorer performance in individuals who have sustained a right-hemisphere stroke (Fisk et al., 2002; Korner-Bitensky et al., 2000; Quigley & DeLisa, 1983). These researchers have noted the impact of visual spatial and perceptual deficits on driving capacity. Although physical impairments can lead to problems with motor reaction time, which can be crucial in driving (e.g., braking) and safe maneuvering (e.g., steering), in many cases, adaptive driving equipment can be used to minimize the impact of physical limitations. For example, an adaptive spinner knob can be attached to the steering wheel to allow controlled steering with the use of only one hand, or a left-foot gas or pedal may be used if the individual is unable to use his or her right foot to push the accelerator or brake. In fact, Smith-Arena, Edelstein, and Rabadi (2006) found that individuals in an acute rehabilitation setting with higher Motricity Index scores and intact
Chapter | 10
Neuroscience and Older Drivers
visual fields were more likely to pass an in-clinic driver evaluation. The researchers concluded that physicians could safely identify post-stroke patients most appropriate for driver evaluation when mild physical impairments, normal visual fields, and mild cognitive impairments were present. In summary, although physical challenges resulting from stroke can impact driving performance, cognitive and visual impairments pose a greater challenge for returning to driving.
2.3.2. Cognition and Perception Stroke can produce a variety of cognitive difficulties that can affect an individual’s ability to return to driving, including slowed information processing speed, visuospatial and perceptual deficits, visual inattention, decreased ability to concentrate, and reasoning difficulties. As a result, cognition and driving after stroke has been extensively studied, with the main goals being to better define this relationship and to identify potential cognitive predictors of driving performance. To date, research has not defined a specific cognitive impairment pattern that is predictive of driving performance, but the results of these studies have identified specific cognitive domains relevant to driving, and some new computerized and noncomputerized driving assessment tasks designed to assess the cognitive domains of driving after stroke have been generated. One of the most common problems associated with stroke and a cognitive domain identified early on as relevant to driving concerns perceptual abilities (Quigley & DeLisa, 1983; Sivak, Olson, Kewman, Won, & Henson, 1981). Early studies examining perceptual/cognitive abilities among right- and left-hemisphere stroke survivors indicated that individuals with right-hemisphere strokes demonstrated the most severe perceptual difficulties. Of those who returned to driving, when self-reported traffic difficulties (e.g., accident involvement) were examined 1 year later, the predictive validity of the perceptual assessment procedure held true for approximately 80% of the sample (Simms, 1985) . Another study used a factor analysis approach to better define the perceptual/cognitive constructs of driving performance by conducting a comprehensive neuropsychological battery on 72 consecutively referred patients who had suffered a stroke (Sundet, Goffeng, & Hofft, 1995). The test battery was factor analyzed into four valid principal components: visual perception, spatial attention, visuospatial processing, and language/praxis. The researchers reported greater visual neglect in right-hemisphere strokes compared to left-hemispheres strokes, but they did not find overall group differences in the number of patients denied driving after a stroke. They concluded that in addition to hemianopia, measures of neglect, speed of mental processing, and emotional disturbances such as denial of illness
133
were the most potent subject characteristics in assessing patients for return to driving (Sundet et al., 1995). Mazer, Korner-Bitensky, and Sofer (1998) examined the use of perceptual tests to predict driving performance in individuals with stroke. Driving performance was quantified as pass or fail outcome of an on-road driving evaluation that was conducted by an occupational therapist and was based on observed driving behaviors. Their results indicated that a test of visual perception skills (Motor Free Visual Perception Test (MVPT)) was the most predictive of on-road performance (positive predictive value ¼ 86.1%; negative predictive value ¼ 58.3%), and that the combination of the MVPT and a measure of task switching (Trail Making Test Part B) represented the most predictive and parsimonious model for predicting on-road performance. Other researchers have reported that a neuropsychological assessment including tests measuring dynamic cognitive processing and complex speed can be useful in assessing driving skills after stroke. For example, Lundqvist, Gerdle, and Ronnberg (2000) reported that complex reaction time the Stroop Color and Word Test, the Listening Span task, and a computerized administration of the K-test were most associated with driving skills, as defined by both on-road and simulated driving performance. Similarly, other researchers have found that although the MVPT is believed to be a strong predictor of on-road evaluation failure, its predictive validity is not sufficiently high to warrant its use as the sole screening tool in identifying those who are unfit to undergo an on-road evaluation (Korner-Bitensky et al., 2000). As the challenge of determining driving capacity following a stroke has been acknowledged, it is not uncommon for many settings to rely on a team of clinicians who evaluate varying aspects of an individual’s ability (e.g., medical and cognitive). One retrospective study, which attempted to better define the contributing factors to a team’s decision on driving ability, examined 104 individuals who had suffered a first stroke (Akinwuntan et al., 2002). The researchers administered both a comprehensive predriving assessment and an on-road test. The predriving assessment included specific measures of vision (monocular vision, binocular vision, stereoscopy, and kinetic vision) and a neuropsychological assessment consisting of eight different tests: the Rey Complex Figure Test, UFOV, divide attention, flexibility, visual scanning, incompatibility, visual field, and neglect (Akinwuntan et al., 2002). Using logistic regression, the researchers found that a model including the side of lesion, kinetic vision, visual scanning, and a road test was the predictor of the team decision; within this model, the road test was the most important determinant. A combination of visual acuity and the Rey Figure test was the best subset for predicting on-road test performance (Akinwuntan et al., 2002). In follow-up prospective studies, these researchers found that a combination of visual neglect,
134
Rey Figure, and on-road test was the best predictor of fitness to drive (as defined by clinicians’ ratings) (Akinwuntan et al., 2006). The accuracy of this short battery was confirmed in another study demonstrating an 86% predictive value of these three tests, which are both sensitive (77%) and specific (92%) in their prediction (Akinwuntan et al., 2007). Taken together, these studies clearly indicate that the determination of driving after stroke cannot be limited to a single cognitive domain.
2.3.3. Aging and Stroke In addition to coping with residual deficits of a stroke, many older adults must cope with ongoing cognitive and physical changes that are commonly seen in aging adults, such as decreased physical mobility, changes in vision, and changes in cognition (e.g., memory problems). Older individuals are also at risk for other neurological involvement; they may be at risk for additional strokes, other cardiovascular disorders, and/or injuries.
3. OTHER CONSIDERATIONS FOR OLDER DRIVERS 3.1. Medication A peripheral challenge in considering the driving ability of older adults is the issue of polypharmacy or the use of multiple types of medications. Not surprisingly, as individuals age, there are higher risks for multiple types of medical issues, including systemic (e.g., diabetes), focal (e.g., cardiac and stroke), or emotional (e.g., depression). Treatment of these often coexisting diagnoses often results in the individual taking multiple medications. Subsequently, the older adult is at risk for taking concomitant medications, which can lead to serious drug interactions.
3.1.1. Emotional Impairment and SelfAwareness Individuals who are unable to drive may suffer increased isolation, which may contribute to depression (Martolli et al., 2000). In addition to the physical limitations that can affect one’s life after stroke, one’s professional and personal lives are also tested and stressed. Many people who have suffered from stroke are unable to return to work in the same capacity as that prior to their disability. The disability not only affects the inflicted person but also the person’s close inner circle. For example, there is more dependence on spouses/family members/close friends for basic everyday activities, such as eating, personal hygiene, and dressing. It is also well-known that stroke patients may have problems recognizing their own cognitive or psychomotor disorders, and they may have serious impairment of functions
PART | II
Key Variables to Understand in Traffic Psychology
that are crucial for safe driving. Particularly, damage to the nondominant hemisphere often causes anosognogia and neglect syndrome and, hence, lowered awareness. Heikkila et al. (1999) found that both patients and their spouses demonstrated a clear tendency to overestimate driving ability compared to the estimates of the neurologist and psychologist.
3.1.2. Education (for the Patient and for the Clinician) As reported in even the earliest study by Quigley and DeLisa (1983), part of the rehabilitation team’s efforts is directed toward providing the candidate with information about the policies and restrictions of the department of motor vehicles. However, Kelly, Warke, and Steele (1999), who investigated the awareness of patients and doctors of medical restrictions to driving, found that educating the client may be difficult. In addition to patients having difficulty knowing if they should drive based on their medical condition, Kelly et al. found that doctors had very poor knowledge of the current licensing policy and action to be taken if a patient was not eligible to drive. Medical staff does not seem to be able to provide this guidance. To increase doctors’ awareness of medical restrictions to driving, greater emphasis must be placed on this aspect of patient care during both undergraduate and postgraduate training (Kelly et al., 1999). With the growing need to improve older adult driving safety, different training strategies are beginning to emerge that focus on changing driving behaviors and knowledge. In 2007, Tuokko, McGee, Gabriel, and Rhodes examined the perceptions of risk, beliefs and attitudes, and openness to change of 86 older participants who voluntarily attended a driver education program. The authors reported that most people attending these sessions were not necessarily concerned about their own driving, safety, or abilities but were interested in maintaining mobility. They were conservative and reasonably consistent in their attitudes toward traffic regulations and safe driving practices. Some gender differences emerged, with more men than women being resistant to changing their driving habits and reporting that they drive after consuming alcohol, and more women than men identifying a role for their families in decision making regarding driving cessation. This suggests that educational material may need to be targeted differently for men and women and that psychosocial factors related to driving, such as driver perception, beliefs, and openness to change, will be useful for maximizing the fit between education program content and outcomes (Tuokko et al., 2007).
4. CONCLUSION This chapter provided an introduction to the current literature on older drivers with and without neurological compromise.
Chapter | 10
Neuroscience and Older Drivers
A main focus was providing information relevant to common clinical diagnoses in older drivers (e.g., dementia and stroke). Although much has been accomplished in this area, much work is needed. In particular, as new technologies (e.g., functional neuroimaging) provide greater insight into the neuroanatomy of aging and neurological compromises, we will be better able to understand the relationship between behavior and brain changes in older drivers. Given the growing number of older adults (and the anticipated continued growth), the need to increase our knowledge is clear. The ability to drive is often synonymous with autonomy, and older drivers in our fast-paced society will likely continue to have a high reliance on automobiles in order to maintain their independence in their communities. The challenge posed to clinicians and driving experts is the ability to balance between an individual’s autonomy and safety (both for the individual and for others). Given the known consequences of neurological compromise, the challenge that remains before us is to identify and refine the best methods to accurately make driving recommendations. In doing so, we can keep our roads and older drivers safe.
REFERENCES Aarsland, D., Bronnick, K., Larsen, J. P., Tysnes, O. B., & Alves, G. (2009). Cognitive impairment in incident, untreated Parkinson disease: The Norwegian Park West study. Neurology, 72, 1121e1126. Adler, G., & Rottunda, S. J. (2010). Mandatory testing of drivers on the basis of age and degenerative diseases: Stakeholder opinions. Journal of Aging and Social Policy, 22, 304e319. Akinwuntan, A. E., Devos, H., Feys, H., Verheyden, G., Baten, G., Kiekens, C., et al. (2007). Confirmation of the accuracy of a short battery to predict fitness-to-drive of stroke survivors without severe deficits. Journal of Rehabilitation Medicine, 39, 698e702. Akinwuntan, A. E., Feys, H., De Weerdt, W., Baten, G., Arno, P., & Kiekens, C. (2006). Prediction of driving after stroke: A prospective study. Neurorehabilitation Neural Repair, 20, 417e423. Akinwuntan, A. E., Feys, H., De Weerdt, W., Pauwels, J., Baten, G., & Strypstein, E. (2002). Determinants of driving after stroke. Archives of Physical Medicine and Rehabilitation, 83, 334e341. American Medical Association. (2010). Physician’s guide to assessing and counseling older drivers. Chicago. American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC. Ball, K., Owsley, C., Sloane, M. E., Roenker, D. L., & Bruni, J. R. (1993). Visual attention problems as a predictor of vehicle crashes in older drivers. Investigative Ophthalmology & Visual Science, 34, 3110e3123. Bedard, M., Weaver, B., Darzins, P., & Porter, M. M. (2008). Predicting driving performance in older adults: We are not there yet!. Traffic Injury Prevention, 9, 336e341. Bjorkdahl, A., & Sunnerhagen, K. S. (2007). Process skill rather than motor skill seems to be a predictor of costs for rehabilitation after a stroke in working age; A longitudinal study with a 1 year follow up post discharge. Biomed Central Health Services Research, 7, 209.
135
Cabeza, R., Nyberg, L., & Park, D. (2005). Cognitive neuroscience of aging: Linking cognitive and cerebral aging. New York: Oxford University Press. Carr, D. B., Duchek, J., & Morris, J. C. (2000). Characteristics of motor vehicle crashes of drivers with dementia of the Alzheimer type. Journal of the American Geriatric Society, 48, 18e22. Cooper, P. J. (1990). Differences in accident characteristics among elderly drivers and between elderly and middle-aged drivers. Accident Analysis and Prevention, 22, 499e508. Daigneault, G., Joly, P., & Frigon, J. (2002). Executive functions in the evaluation of accident risk of older drivers. Journal of Clinical and Experimental Neuropsychology, 24, 221e238. Drachman, D. A., & Swearer, J. M. (1993). Driving and Alzheimer’s disease: The risk of crashes. Neurology, 43, 2448e2456. Dubinsky, R. M., Stein, A. C., & Lyons, K. (2000). Practice parameter: Risk of driving and Alzheimer’s disease (an evidence-based review): Report of the quality standards subcommittee of the American Academy of Neurology. Neurology, 54, 2205e2211. Duchek, J. M., Carr, D. B., Hunt, L., Roe, C. M., Xiong, C., Shah, K., et al. (2003). Longitudinal driving performance in early-stage dementia of the Alzheimer type. Journal of the American Geriatric Society, 51, 1342e1347. Fisk, G. D., Owsley, C., & Mennemeier, M. (2002). Vision, attention, and self-reported driving behaviors in community-dwelling stroke survivors. Archives of Physical Medicine and Rehabilitation, 83, 469e477. Fisk, G. D., Owsley, C., & Pulley, L. V. (1997). Driving after stroke: Driving exposure, advice, and evaluations. Archives of Physical Medicine and Rehabilitation, 78, 1338e1345. Freund, B., Gravenstein, S., Ferris, R., Burke, B. L., & Shaheen, E. (2005). Drawing clocks and driving cars. Journal of General Internal Medicine, 20, 240e244. Friedland, R. P., Koss, E., Kumar, A., Gaine, S., Metzler, D., Haxby, J. V., et al. (1988). Motor vehicle crashes in dementia of the Alzheimer type. Annals of Neurology, 24, 782e786. Gilley, D. W., Wilson, R. S., Bennett, D. A., Stebbins, G. T., Bernard, B. A., Whalen, M. E., et al. (1991). Cessation of driving and unsafe motor vehicle operation by dementia patients. Archives of Internal Medicine, 151, 941e946. Grady, C. L. (2008). Cognitive neuroscience of aging. Annals of the New York Academy of Sciences, 1124, 127e144. Gunning-Dixon, F. M., & Raz, N. (2000). The cognitive correlates of white matter abnormalities in normal aging: A quantitative review. Neuropsychology, 14, 224e232. Hakamies-Blomqvist, L. E. (1993). Fatal accidents of older drivers. Accident Analysis and Prevention, 25, 19e27. Haug, H., & Eggers, R. (1991). Morphometry of the human cortex cerebri and corpus striatum during aging. Neurobiology of Aging, 12, 336e338. Heikkila, V. M., Korpelainen, J., Turkka, J., Kallanranta, T., & Summala, H. (1999). Clinical evaluation of the driving ability in stroke patients. Acta Neurologica Scandinavia, 99, 349e355. Hughes, C. P., Berg, L., Danziger, W. L., Coben, L. A., & Martin, R. L. (1982). A new clinical scale for the staging of dementia. British Journal of Psychiatry, 140, 566e572. Iverson, D. J., Gronseth, G. S., Reger, M. A., Classen, S., Dubinsky, R. M., & Rizzo, M. (2010). Practice parameter update: Evaluation and management of driving risk in dementia. Neurology, 74, 1316e1324.
136
Kay, L., Bundy, A., Clemson, L., & Jolly, N. (2008). Validity and reliability of the on-road driving assessment with senior drivers. Accident Analysis and Prevention, 40, 751e759. Kelly, R., Warke, T., & Steele, I. (1999). Medical restrictions to driving: The awareness of patients and doctors. Postgraduate Medical Journal, 75, 537e539. Korner-Bitensky, N. A., Mazer, B. L., Sofer, S., Gelina, I., Meyer, M. B., Morrison, C., et al. (2000). Visual testing for readiness to drive after stroke: A multicenter study. Archives of Physical Medicine and Rehabilitation, 79, 253e259. Lafont, S., Laumon, B., Helmer, C., Dartigues, J. F., & Fabrigoule, C. (2008). Driving cessation and self-reported car crashes in older drivers: The impact of cognitive impairment and dementia in a population-based study. Journal of Geriatric Psychiatry and Neurology, 21, 171e182. Langford, J. (2008). Usefulness of off-road screening tests to licensing authorities when assessing older driver fitness to drive. Traffic Injury and Prevention, 9, 328e335. Lee, N., Tracy, J., Bohannon, R. W., & Ahlquist, M. (2003). Driving resumption and its predictors after stroke. Connecticut Medicine, 67, 387e391. Legh-Smith, J., Wade, D. T., & Hewer, R. L. (1986). Driving after a stroke. Journal of the Royal Society of Medicine, 79, 200e203. Lundqvist, A., Gerdle, B., & Ronnberg, J. (2000). Neuropsychological aspects of driving after a strokedin the simulator and on the road. Applied Cognitive Psychology, 14(2), 135e150. Martolli, R. A., de Leon, C. F. M., Glass, T. A., Williams, C. S., Cooney, L. M., & Berkman, L. F. (2000). Consequences of driving cessation: Decreased out-of-home activity levels. Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 55, S334eS340. Mazer, B. L., Korner-Bitensky, N. A., & Sofer, S. (1998). Predicting ability to drive after stroke. Archives of Physical Medicine and Rehabilitation, 79, 743e750. McGwin, G., Jr., & Brown, D. (1999). Characteristics of traffic crashes among young, middle-aged, and older drivers. Accident Analysis and Prevention, 31, 181e198. O’Neill, D., Neubauer, K., Boyle, M., Gerrard, J., Surmon, D., & Wilcock, G. K. (1992). Dementia and driving. Journal of the Royal Society of Medicine, 85, 199e202. Ott, B. R., Heindel, W. C., Papandonatos, G. D., Festa, E. K., Davis, J. D., Daiello, L. A., et al. (2008). A longitudinal study of drivers with Alzheimer’s disease. Neurology, 70, 1171e1178. Ott, B. R., Heindel, W. C., Whelihan, W. M., Caron, M. D., Piatt, A. L., & DiCarlo, M. A. (2003). Maze test performance and reported driving ability in early dementia. Journal of Geriatric Psychiatry and Neurology, 16, 151e155. Quigley, F. L., & DeLisa, J. A. (1983). Assessing the driving potential of cerebral vascular accident patients. American Journal of Occupational Therapy, 37, 474e478. Raz, N., Gunning-Dixon, F. M., Head, D., Dupuis, J. H., & Acker, J. D. (1998). Neuroanatomical correlates of cognitive aging: Evidence from structural magnetic resonance imaging. Neuropsychology, 12, 95e114. Raz, N., Lindenberger, U., Rodrigue, K. M., Kennedy, K. M., Head, D., Williamson, A., et al. (2005). Regional brain changes in aging healthy adults: General trends, individual differences and modifiers. Cerebral Cortex, 15, 1676e1689. Raz, N., & Rodrigue, K. M. (2006). Differential aging of the brain: Patterns, cognitive correlates and modifiers. Neuroscience and Biobehavioral Reviews, 30, 730e748.
PART | II
Key Variables to Understand in Traffic Psychology
Rebok, G. W., Bylsma, F. W., Keyl, P. M., Brandt, J., & Folstein, S. E. (1995). Automobile driving in Huntington’s disease. Movement Disorders, 10, 778e787. Rizzo, M., Uc, E. Y., Dawson, J., Anderson, S., & Rodnitzky, R. (2010). Driving difficulties in Parkinson’s disease. Movement Disorders, 25(Suppl. 1), S136eS140. Rodriguez-Oroz, M. C., Jahanshahi, M., Krack, P., Litvan, I., Macias, R., Bezard, E., et al. (2009). Initial clinical manifestations of Parkinson’s disease: Features and pathophysiological mechanisms. Lancet Neurology, 8, 1128e1139. Ryan, G. A., Legge, M., & Rosman, D. (1998). Age related changes in drivers’ crash risk and crash type. Accident Analysis and Prevention, 30, 379e387. Salthouse, T. A. (1996). The processing-speed theory of adult age differences in cognition. Psychological Review, 103, 403e428. Simms, B. (1985). The assessment of the disabled for driving: A preliminary report. International Rehabilitation Medicine, 7, 187e192. Sivak, M., Olson, P. L., Kewman, D. G., Won, H., & Henson, D. L. (1981). Driving and perceptual/cognitive skills: Behavioral consequences of brain damage. Archives of Physical Medicine and Rehabilitation, 62, 476e483. Smith-Arena, L., Edelstein, L., & Rabadi, M. H. (2006). Predictors of a successful driver evaluation in stroke patients after discharge based on an acute rehabilitation hospital evaluation. Archives of Physical Medicine and Rehabilitation, 85, 44e52. Staplin, L., Gish, K. W., & Wagner, E. K. (2003). MaryPODS revisited: Updated crash analysis and implications for screening program implementation. Journal of Safety Research, 34, 389e397. Stutts, J. C., Stewart, J. R., & Martell, C. (1998). Cognitive test performance and crash risk in an older driver population. Accident Analysis and Prevention, 30, 337e346. Sundet, K., Goffeng, L., & Hofft, E. (1995). To drive or not to drive: Neuropsychological assessment for driver’s license among stroke patients. Scandinavian Journal of Psychology, 36, 47e58. Trobe, J. D., Waller, P. F., Cook-Flannagan, C. A., Teshima, S. M., & Bieliauskas, L. A. (1996). Crashes and violations among drivers with Alzheimer disease. Archives of Neurology, 53, 411e416. Tucker-Drobb, E. M., & Salthouse, T. (2008). Adult age trends in the relations among cognitive abilities. Psychology of Aging, 23(2), 453e460. Tuokko, H., Tallman, K., Beattie, B. L., Cooper, P., & Weir, J. (1995). An examination of driving records in a dementia clinic. Journal of Gerontology Series B: Psychological Sciences and Social Sciences, 50, S173eS181. Tuokko, H. A., McGee, P., Gabriel, G., & Rhodes, R. E. (2007). Perception, attitudes and beliefs, and openness to change: Implications for older driver education. Accident Analysis and Prevention, 39, 812e817. Uc, E. Y., Rizzo, M., Anderson, S. W., Dastrup, E., Sparks, J. D., & Dawson, J. D. (2009). Driving under low-contrast visibility conditions in Parkinson disease. Neurology, 73, 1103e1110. Uc, E. Y., Rizzo, M., Anderson, S. W., Sparks, J. D., Rodnitzky, R. L., & Dawson, J. D. (2006). Driving with distraction in Parkinson disease. Neurology, 67, 1774e1780. Uc, E. Y., Rizzo, M., Johnson, A. M., Dastrup, E., Anderson, S. W., & Dawson, J. D. (2009). Road safety in drivers with Parkinson disease. Neurology, 73, 2112e2119. Whelihan, W. M., DiCarlo, M. A., & Paul, R. H. (2005). The relationship of neuropsychological functioning to driving competence in older persons with early cognitive decline. Archives of Clinical Neuropsychology, 20, 217e228.
Chapter 11
Visual Attention While Driving Measures of Eye Movements Used in Driving Research David Crundall and Geoffrey Underwood University of Nottingham, Nottingham, UK
1. INTRODUCTION Van Gompel, Fischer, Murray, and Hill (2007) introduced their book on eye movements with a historic perspective on the idea that the eyes provide access to the inner workings of the mind and brain. They quote De Laurens (1596) as referring to the eyes as “windowes (sic) of the mind” (p. 3), which presents an opportunity to indirectly observe what is being processed in the brain on the basis of what the eye is looking at. This link between the location of the eye in the visual world and the concomitant processing in the brain is most formally stated in Just and Carpenter’s (1980) eyeemind assumption, which states that the eye remains fixated on an object until the brain has finished processing it. Various ancillary assumptions can be appended to this, such as the argument that the brain should not be processing any visual information that the eye is not looking at, and that whenever the eye is fixating something, that particular object must be being processed. If these assumptions are met, it is easy to see how valuable it would be for a psychologist to monitor the eye movements of individuals engaging in various tasks, including driving, which is predominantly dependent on the processing of visual information. Unfortunately, the story is not so simple. Numerous studies demonstrate the ability of readers to process words parafoveallydthat is, to process them without looking at them directly (Underwood & Everatt, 1992). Conversely, there is evidence that looking directly at an object or area of a scene does not guarantee that the viewer will process the information. Studies of change blindness have demonstrated that viewers may not notice a change made to an object in a visual scene even though they are looking at the object when the change occurs (Caplovitz, Fendrich, & Hughes, 2008). Indeed, whenever the mind wanders while reading a book, it is common to feel that one has read a sentence without actually processing what it meant, and drivers sometimes report not being aware of familiar sections of roadway that they have successfully negotiated. Despite the refutation of the strong version of the Handbook of Traffic Psychology. DOI: 10.1016/B978-0-12-381984-0.10011-6 Copyright Ó 2011 Elsevier Inc. All rights reserved.
eyeemind assumption, the link between what the eye is looking at and what the viewer is thinking about is still very robust. Although Underwood and Everatt (1992) have presented several challenges to the eyeemind assumption, they acknowledge that these are all special cases in which the assumption can be shown to fail, and in general it is a safe working assumption that if someone is looking at something, then he or she is processing it. Modern reviews of eye tracking (Van Gompel et al., 2007) have demonstrated the appeal of this methodology as a means of better understanding how people approach and engage in a variety of tasks and situations. We believe that the driving task is eminently suited to the application of eye tracking methodologies. The information that a driver uses is predominantly visual (Sivak, 1996), and a wide range of specific driving behaviors, from navigation to anticipation of hazardous events, are primarily dependent on the optimum deployment of attention through overt eye movements. Classic studies of road collision statistics have identified perceptual problems to be a leading cause of traffic crashes (Lestina & Miller, 1994; Sabey & Staughton, 1975; Treat et al.,1979), and in-car observation of driver behavior preceding an actual crash supports the causal role of distraction and inattention (Klauer, Dingus, Neale, Sudweeks, & Ramsey, 2006). Several reviews of driving research have all reached the same conclusion: When and where drivers look is of vital importance to driver safety (Lee, 2008; Underwood, 2007), and we need to record and interpret these eye movements in order to decrease death and injury on our roads (Shinar, 2008). Recording and interpreting eye movements has indeed been a valuable tool in driving research during the past 40 years, and we review much of the findings in this chapter. This methodology is likely to become even more important in the future with the development of transport simulators that allow the recording of eye movements in near-naturalistic situations while maintaining a high degree of experimental control over the environment. Considering 137
138
the interest that has been and will be shown in recording drivers’ eye movements, we believe that it is important to provide a review of the various measures that can be and have been captured in previous studies and to discuss how these different measures can be used to address different hypotheses. This chapter is not concerned with the benefits of one eye tracker over another (for a review of eye tracking methods, see Duchowski (2007)) but is restricted to measures we might record. Essentially, eye movements consist of two primary events: fixations and saccades. Fixations are periods of relative stability, during which the eyes focus on something in the visual scene. Such fixations most often reflect the fact that the brain is processing the fixated information. Saccades are rapid, ballistic jumps of the eye that separate the fixations and serve to orient the focus of the eyes from one point of interest to another. No visual information is taken in during these rapid movements. Although we can reduce eye movements to these two components, there are numerous exceptions and different methods of capturing, combining, averaging, and analyzing these processes. This chapter provides an overview of some of these methods and also reviews the various studies that have employed these methods in their search for greater insight into the task of driving. We start with the most obvious of measuresd assessing whether drivers actually look at elements of the road scene that might help prevent a collision.
2. DO DRIVERS LOOK AT CRITICAL INFORMATION? Lee (2008) reviewed 50 years of research and concluded that collisions occur because drivers “fail to look at the right thing at the right time” (p. 525). We first consider how we should measure whether drivers “look at the right thing” before later considering how to measure whether they look “at the right time.” The simplest conception of whether an individual has looked at a certain object in a scene is whether the eye coordinates recorded by an eye tracker are coincident with the world coordinates of the object. This can be calculated automatically for eye movements to static images where the precise coordinates of an object are easily defined and related to the eye position coordinates. Many eye tracking software packages allow areas of interest (AOIs) to be generated for particular pictures, which allow automatic calculation of when individuals look at specific objects. These AOIs are regions of a visual image that are defined by two-dimensional (2-D) coordinates in the viewing plane and thus allow software to identify fixations that fall within their boundaries. However, there are two particular problems with this approach. First, the AOI is typically a symmetric shape drawn on top of the image or stimulus
PART | II
Key Variables to Understand in Traffic Psychology
(most often a rectangle). Unfortunately, real-world objects rarely fit into such shapes. Objects may have irregular outlines, or their 2-D shape may be distorted by the 3-D representation. Objects in real scenes also tend to be partially obscured or may themselves partially obscure other interesting stimuli. It is impossible to determine from 2-D data whether the participant is genuinely looking at the car ahead or whether he or she is looking through the windows of the car to identify any further traffic that might be obscured. The second, more pertinent, problem is that AOIs cannot be practically defined for unpredictable interactions on the road or in a driving simulator. Thus, if eye movements are recorded from an on-road vehicle, it is impractical to define the coordinates of a particular vehicle in the road ahead because the position of both the target car and the participant’s car would require the AOI coordinates to be updated constantly. Improvements in the software of at least one eye tracking system have extended the application of AOIs to video-based stimuli (where one can specify an AOI at several points during a video, and the software will interpolate the coordinates in between, creating a dynamic AOI); however, this procedure is limited by the predictability of the dynamic object that one wishes to track. For instance, it is relatively easy to interpolate the coordinates of an approaching car if the vehicle maintains its heading and speed: Drawing an AOI around the car when it first appears in the video and when it is last visible will allow the software to estimate the rate and direction of the AOI expansion as the vehicle approaches. Less predictable patterns require more experimenterdefined AOIs to allow more accurate prediction of how the AOIs change. At least with video-based stimuli, even if a large number of experimenter-defined AOIs were required to identify a single object, the calculations could then be applied across all participants watching the same video clips. With simulation and on-road eye tracking, however, this benefit is absent because the position and dynamic nature of visual targets will vary across participants. A requirement to define a large number of AOIs for every participant simply to assess whether drivers tend to look at a particular object would rapidly become impractical. Although it is theoretically feasible that a simulator could record coordinates of objects as they move through the virtual world, providing a personalized dynamic AOI (several research groups are pursuing this goal), we are unaware of any published articles that have used this methodology. Instead, researchers who are faced with eye movement data from dynamic stimuli (especially on-road eye tracking) must often perform a frame-by-frame analysis of video footage containing the dynamic stimuli (e.g., the external world in on-road tests, often recorded through a windscreen-mounted or head-mounted camera) and an overlaid cursor depicting where the eye tracker thinks they were looking.
Chapter | 11
Visual Attention While Driving
A study that employed this methodology was conducted by Pradhan et al. (2005). Three groups of drivers, of varying age and experience, drove through a series of simulated scenarios in which potential hazards might (but did not) occur. For instance, one scenario contained a line of bushes obscuring the entry to a pedestrian crossing. It was feasible that a pedestrian could have emerged from behind the hedge and entered the crossing in front of the participant’s vehicle. Participants were assessed in regard to whether they looked at this hedge on approach to the pedestrian crossing, with the associated assumption that a glance at the hedge reflected the driver’s concern that it might conceal a pedestrian. The results of the study demonstrated that novices “often completely fail to look at elements of a scenario that clearly need to be scanned in order to acquire information relevant to the assessment of a potential risk” (p. 851). In one particular scenario, only 10% of novices looked in the appropriate direction to check whether pedestrians might emerge from behind a parked truck onto a crossing. The more experienced drivers, by contrast, were significantly more likely to look at these a priori areas of the visual scene, which was considered indicative of safe behavior by the experimenters. Similar results were found by Borowsky, Shinar, and Oron-Gilad (2010), who compared young, inexperienced drivers to more experienced drivers. They noted instances in which the experienced drivers fixated areas in the scene that they considered important to hazard detection (e.g., vehicles merging from an adjoining road). In this particular instance, the young, inexperienced drivers tended to focus on the road ahead, apparently disregarding the hazard posed by merging vehicles. These results are important because they offer a suggestion as to why novice drivers are consistently overrepresented in crash statistics (Clarke, Ward, Bartle, & Truman, 2006; Organisation for Economic Co-operation and Development/European Conference of Ministers of Transport, 2006; Underwood, 2007; Underwood, Chapman, & Crundall, 2009; Underwood, Crundall, & Chapman, 2007). Several researchers have argued that novices have poor hazard perception skills (Horswill & McKenna, 2004), and the results of Pradhan et al. (2005) and Borrowsky et al. (2010) ostensibly demonstrate that a failure to anticipate locations from where hazards may emerge and then prioritize these locations for visual search may be a major contributor to failures in hazard perception. There is, however, a potential confound in the interpretation of simple glance measures that record only where the driver has looked. Often, they do not take into account any measure of duration, simply assuming that to look at an object is to process that object. However, extremely short fixations may not reflect sufficient processing time for a particular fixated object to be identified. It is common practice in eye movement research to filter out extremely short fixations from subsequent analyses because they are
139
unlikely to reflect object processing (often researchers define a fixation as at least 100 ms of eye stability). Fixations shorter than 100 ms do occur, but they are more likely to reflect attempts to reorient visual search on the basis of global features in the whole scene rather that processing what is at the point of fixation. However, the 100-ms cutoff is not a psychophysical threshold, after which any fixations must have accessed the identity of the fixated object, but is instead a heuristic for accepting and rejecting data. We know that some objects take longer to process than others because they have a higher threshold for identification. This is most obvious in the research literature on eye movements during reading. Results have consistently shown that lower frequency words receive longer fixations (Liversedge & Findlay, 2000; Rayner, 1998), and this appears to translate to objects in scenes, with unexpected or inconsistent objects receiving longer fixations (Henderson & Hollingworth, 1999; Underwood, Templeman, Lamming, & Foulsham, 2008). Even task-irrelevant objects may evoke longer fixation durations due to their novel or unexpected nature (Brockmole & Boot, 2009). This, in itself, does not pose a problem for the typical glance analyses seen in many driving studies (Pradhan et al., 2005), providing that all glances to target objects meet the required threshold duration. The problem arises when fixations are curtailed before the threshold is met. This issue is detailed in the E-Z Reader model of reading (Reichle, Rayner, & Pollatsek, 2003). Reichle et al. presume two levels of lexical access when looking at a word in a sentence. The first level is the familiarity check, which is undertaken when the eye first lands on the word. If the word is identified as familiar (and therefore to have a low threshold for complete identification), then the oculomotor system begins to plan the next saccade in parallel with the second level of word processing, secure in the knowledge that the full identity of the word will be accessed before the saccade is triggered. If, however, the word does not pass the familiarity check, then the subsequent plan for the next saccade may be delayed. Even if the system has already begun to plan the next saccade, it can be canceled if the order is made quickly enough (during the labile stage of the saccadic planning process). If the reader realizes that the word is a difficult one to process only once the saccadic plan has reached the nonlabile stage, the eye movement will go ahead even though the reader has realized that more attention needs to be devoted to the troublesome word. Thus, the eye will move away from the word before it is identified. This will often lead to a regressive saccade, where the eye jumps back in the text in order to reprocess a tricky word. It is highly probable that something similar occurs with drivers’ glances when on the road. Consider, for instance, the most common cause of motorcycle collisions in the United Kingdomdwhen a car driver pulls out from a side road into the path on an oncoming motorcycle (Clarke,
140
Ward, Bartle, & Truman, 2007). The driver will check down the road to see if there is any conflicting traffic. If the driver’s eyes land directly upon an approaching car, the familiarity check will probably be completed successfully and allow the driver to plan the next saccade during level 2 processing. Information about this approaching vehicle (trajectory, speed, etc.) could then be integrated with information about traffic coming from the opposite direction, which may then identify a suitable gap in which to pull out. Even if the approaching car was displaying behavior that required closer attention, this should be flagged at least in the labile stage of saccadic programming, allowing the subsequent saccade to be canceled and more attention to be given to the car. The situation becomes more precarious, however, if the approaching vehicle is a motorcycle. Because motorcycles comprise only 1% of UK traffic (Department for Transport, 2010a, 2010b), they are novel, low-frequency items that should therefore be associated with higher thresholds for target identification (and thus, just as with low-frequency words, would require longer fixations before identification is achieved). Even assuming that the driver looks directly at the approaching motorcycle, the low cognitive and physical conspicuity associated with the target may result in the familiarity check wrongly assuming that the road is empty. Thus, the eye may move away from the motorcycle before the perceptual system has had time to process and identify the threat. Even if this happens, in fortunate cases the driver may at least realize that there is an approaching motorcycle in the nonlabile stage of the saccade program. Although the eyes may move off the motorcycle before fully processing it, the driver may then have enough information to realize that this was an error and then re-fixate on the motorcycle, which may be enough to enable a second look. In particularly unfortunate cases, however, the driver may have no contradictory information to challenge the initial assumption that the road is empty, and the maneuver may proceed with dire consequences. This is a classic case of a “look but fail to see” accident, in which drivers often report that they had looked in the appropriate direction but had completely failed to see the approaching motorcycle (Brown, 2002). There is even evidence that initial gaze durations upon approaching motorcycles at T-junctions can, in certain situations, be shorter than the corresponding gazes devoted to approaching cars. Considering that cars and motorcycles are the equivalent of high- and low-frequency words, respectively, we would expect this significant effect to be reversed. The short initial gazes on motorcycles were therefore argued to be indicative of initial fixations that were not long enough for drivers to realize exactly what they were looking at (Crundall, Crundall, Clarke, & Shahar, in press). To further pursue the analogy of driving and reading, a study of eye movements during reading found
PART | II
Key Variables to Understand in Traffic Psychology
the equivalent of “look but fail to see” events (Ehrlich & Rayner, 1981). Readers were set a task of detecting misspellings in a paragraph of text, and quite often they would look at a misspelling but not report it. Interestingly, this was more likely to happen if the word was predictable: The readers saw what they expected to see, and perhaps drivers sometimes do the same. In light of the possibility of “look but fail to see” errors, one can understand how perilous it might be to rely on a binary measure of whether a glance occurred or not. Instead, we argue that glance probability or frequency needs to be paired with a measure of glance duration. The following section considers the different measures of duration that can be used and what previous studies have shown regarding the sensitivity of these durations to different driving conditions.
3. MEASURES OF GLANCE DURATION Several measures of glance duration are typically used in studies of reading, scene perception, and driving. The smallest unit of duration is the first fixation duration (FFD). This represents the time that the eye dwells in one place for the first time. This is only applicable in relation to specific objects. For instance, it makes sense to consider the FFD upon a pedestrian who steps out from behind a truck, but it is less useful to record the FFD on more general categories such as the “road ahead.” This initial fixation and subsequent fixations are often averaged to create a mean fixation duration (MDF), which can apply to both specific objects and general categories. Gazes differ slightly from fixations in that they are concerned with multiple fixations on specific objects. Depending on how large an object of interest is, it may be possible to have two separate fixations within the same object without having saccaded away (i.e., the eye moves to another location within the object). Only when a fixation occurs outside the boundary of the object does the gaze end. Total dwell time (TDT) is simply the summation of all fixations on a specific object. Figure 11.1 shows a pictorial representation of how these measures are calculated. Typically, all of these measures reflect various levels of processing of the stimuli, although there are arguments for and against different measures. For instance, TDT represents the most stable measure of overt attention, presented either in absolute terms (providing all participants had the same duration of opportunity to fixate the object) or in percentage or ratio terms. Individual fixations, however, represent the most sensitive measure of processing demand, although they are more susceptible to slight variations in a magnitude of potentially confounding factors. Henderson (1992) argues that the more global measures of fixation time, such as TDT, on an object will include not just the time spent processing the target but also many other
Chapter | 11
Visual Attention While Driving
141
FIGURE 11.1 Hypothetical fixations while driving a car. There are three fixations on the rearview mirror (b, c, and e). There are two gazes on the rearview mirror (b þ c and e). The total dwell time on the rearview mirror is b þ c þ e. Mean fixation duration on the rearview mirror is the total dwell time divided by the number of fixations [(b þ c þ e)/3]. Mean gaze duration is ((b þ c) þ e)/2.
post-identification processes, such as integrating the object into a situational model. On this basis, Henderson argues in favor of the FFD as the cleanest measure of initial processing demands, although within the driving domain this could give undue prominence to “look but fail to see” fixations, potentially leading researchers to underestimate the processing time required to successfully identify an object in the first fixation. Some researchers have gone further than Henderson: Kotowicz, Rutishauser, and Kock (2010) suggest that in simple visual search experiments, a fixation on the target might not be necessary to identify it. They found that participants could accurately report a target location with as little as 10 ms of fixation on it before it offset. They argue that the target is actually identified extrafoveally during the visual search, which then results in the saccade to the target location. The amount of time spent fixating the target does not increase target identification accuracy, but it does increase confidence in reporting the item as the target. We argue, however, that Kotowicz et al.’s data are unlikely to transfer from a simple visual search task for a target among distracters to a complex driving situation. In simple visual search tasks, the targets are usually highly constrained in both appearance and location, the task tends not to require any secondary response, and the distracters are welldefined. In driving, however, there are multiple task goals that need to be monitored, with a nonuniform background to interrogate. Furthermore, it might be understandable why participants in some simple visual search experiments attempt to maximize their use of extrafoveal vision: Without any peripheral cues, participants are faced with using a random search strategy (or at least a strategy that is unrelated to the likely location of subsequent targets). When driving, however, people often use specific search strategies to scan the road ahead in anticipation of potential hazards that might occur. As discussed later, these strategies are built up from experience and from learning where in the driving scene certain information is available or where certain hazards might be likely to appear. Thus, drivers are unlikely to mobilize as many extrafoveal resources to direct their saccades as were
participants in Kotowicz et al.’s (2010) simple visual search experiments. The use of fixation and gaze durations in driving research has revealed some very consistent patterns. For instance, Chapman and Underwood (1998) and Underwood, Phelps, Wright, van Loon, and Galpin (2005) demonstrated that fixation durations tend to increase in the presence of specific hazards (e.g., the car ahead suddenly braking). We interpret this in the same way that lowfrequency words or incongruent objects in pictures evoke longer fixation durations: Because hazards are relatively novel events, which are more difficult to predict and require additional processing, they require longer fixation durations. This is not to say that all long fixation durations on a hazard are indicative of high initial processing demands, but (as Henderson (1992) would argue) it is likely that longer fixations also reflect ongoing monitoring of the hazard, attempts to integrate it into a situational model, and concomitant memory processes. Regardless of the precise reason for the increased fixation lengths during hazards, it is clear that this is a form of attentional capture, in which the saccade to the next fixation location is delayed for longer than usual due to the additional processing tasks that are inherent with hazards. Interestingly, there is an additional parallel to the reading literature: Not only do low-frequency, novel, and complex hazards (or words) demand longer fixations but also those individuals who are considered better at the primary task tend to have overall shorter fixations. In studies of eye movements in reading, it has been shown that reading age correlates with a reduction in fixation length on words (Rayner, 1998), and greater exposure to typically low-frequency words will lower the thresholds for those words relative to those of other readers. For instance, lawyers are likely to have shorter fixations on certain Latin phrases compared to readers from other professions. In the same way, it appears that greater experience in driving tends to lead to a reduction in the duration of certain types of fixations. Chapman and Underwood (1998) noted this when recording the eye movements of novice and experienced drivers while watching hazard perception video clips.
142
Although the appearance of a hazard tended to increase the fixation durations of all participants, the experienced drivers had consistently shorter fixations than the novice drivers in all conditions, especially in the presence of a hazard. They argued that this was because the hazardous events are less novel to the experienced drivers: If they have previously been exposed to similar situations, their threshold for understanding the threat posed by a particular object should be lower. This experiential effect has been replicated in a simulator. Konstantopoulos, Crundall, and Chapman (2010) measured the eye movements of driving instructors and learner drivers while navigating a hazardous route through a virtual city in a medium fidelity simulator. They found the driving instructors to have shorter, more frequent fixations than the learner drivers, which supports the suggestion that experience and expertise improve one’s ability to extract relevant driving information from a single fixation. A second finding of interest from the Konstantopoulos study was an increase in fixation durations during nighttime driving and driving through rain in the simulator. They argue that the nighttime and rain conditions decreased visibility, thereby increasing the difficulty of extracting information during individual fixations. Again, this parallels the reading research, which suggests that words that are more difficult to perceive will require longer fixations (Reingold & Rayner, 2006). However, there are certain paradoxes in the fixation duration literature on driving that echo Henderson’s (1992) concerns that anything beyond the FFD is likely to reflect a host of post-identification processes. For instance, whereas hazards tend to produce longer fixations, more complex driving scenes tend to reduce overall fixation length. For instance, Chapman and Underwood (1998) noted that video clips of complex urban settings tended to evoke significantly shorter fixations than did clips of rural settings. Similarly, on-road data suggest that more visually complex roadways tend to produce shorter, more frequent fixations than more sparsely populated roads, such as dual carriageways or single-carriageway rural roads (Crundall & Underwood, 1998). Typically, the urban and suburban roads that produce the shortest fixation durations have a greater number of potential distracters and potential hazards: Shop fronts, pedestrians, parked vehicles, road and informational signs, and roadside advertising all vie for the attention of drivers. This increase in visual complexity requires a higher sampling rate of visual search. In contrast, empty undulating rural roads provide little to distract or interest the attention of the driver beyond the immediate road ahead. Although participants might occasionally search hedgerows for gates and emerging vehicles, the majority of their time will be spent looking as far down the road as possible (although how far one can comfortably look down the road varies from person to person for a variety of reasons, including driving experience and the
PART | II
Key Variables to Understand in Traffic Psychology
extent of extrafoveal region from which ambient information may be processed) (Underwood et al., 2007). Thus, the long fixation durations that are seen during rural driving are not predominantly due to object processing or identification but instead related to vigilance and monitoring. In summary, the length of fixation durations provides an important addition to a simple analysis of binary or frequency-based glances analyses. Unfortunately, the length of a fixation is affected by many factors (Henderson, 1992), and without a careful understanding of what is reflected in these duration measures, it is possible to draw erroneous conclusions. However, a number of clear patterns have emerged from a decade or more of studies. First, it seems that experience in the driving domain does tend to reduce fixation durations on average, most likely through a mixture of reduced thresholds for object or event identification and through an increase in processing speed. Second, localized increases in demand, such as the appearance of a hazardous pedestrian stepping out from behind a parked vehicle, tend to increase fixation durations. Considering the priority given to these stimuli, this may be understandable. However, the fact that novice driver fixations are proportionally more affected by the appearance of a hazard raises the potential problem of attentional capture beyond what is required to process the hazard. This may have a negative impact on the driver’s ability to process other objects or events that appear soon after the hazard. Third, a dispersed increase in general demand (in regard to more visually complex road scenes) provokes the opposite reaction to a localized increased in demand. Whereas hazards capture attention, urban and suburban scenes promote shorter and more frequent fixations in order to cope with the greater number of points of potential interest.
4. MEASURES OF SPREAD Although the process of identifying what drivers look at, and for how long, has resulted in a number of insights into driver vision and behavior, these measures provide no indication of whether drivers adopt general scanning strategies when driving. Certainly, a number of experts in the field of driver training expound the view that wide and constant scanning is important for safe driving (Coyne, 1997; Mills, 2005) and warn against the “disastrous habit of fixating [in one place for too long]” (Haley, 2006, p. 112). Some of the earliest research also noted that novice drivers scan a smaller area of the visual scene (Mourant & Rockwell, 1972). Later research confirmed that new drivers scan the road in a curiously maladaptive way, tending to look straight ahead of them and not showing any sensitivity to changes in driving conditions (Crundall & Underwood, 1998; Konstantopoulos et al., 2010). The measure of greatest interest in the study by Crundall and Underwood was the variance of fixation locations, with high variance
Chapter | 11
Visual Attention While Driving
indicating greater scanning. This measure is calculated from the location of fixations in one axis (in the eye tracking system’s reference frame). For instance, a sample of 100 fixations will provide a sequence of xy coordinates. The variance, or standard deviation, of the x coordinates reflects the extent of search activity in the horizontal axis. Likewise, using the y coordinates would produce a measure of the spread of search in the vertical axis. In the Crundall and Underwood (1998) study, new and experienced drivers traveled along a range of roads through countryside, suburban housing areas, and along a demanding section of a multilane highway. Their eye movements were monitored and recorded during this drive. On simple rural roads with few hazards, all drivers tended to look at the roadway ahead, but on a tricky multilane highway with traffic joining from both left and right, the experienced drivers increased their scanning, whereas the novices continued to look straight ahead. This is maladaptive, or insensitive, because merging traffic requires an adjustment to the driver’s own speed and a preparedness to take avoiding acting (and is in accordance with the work of Borowsky et al., 2010). The experienced drivers showed situation awareness in that they looked around them to determine the trajectories of the proximal traffic. Why should the novice drivers fail to scan for hazards on the roads most likely to present them with dangers? Three explanations considered here are that (1) novices need to look at markings on the road (white lines, curbs, and barriers) in order to steer their vehicles, (2) novices have not yet automatized the steering and speed controldthe subskills required for the coordination of the vehicledand are thus are unable to allocate mental resources to the task of monitoring other traffic, or (3) novices may opt not to look around them because they have a poor idea of the dangers present on these roadsdthey have inadequate situation awareness (Gugerty, 1997; Horswill & McKenna, 2004; Underwood, 2007). Of course, these three explanations are not mutually exclusive. It is plausible that new drivers have difficulties in maintaining their lane position, in thinking about much else other than controlling the vehicle, and in forming a mental model of what the other road users are doing and what they are likely to do next. As drivers become more skilled in handling their vehicles, cognitive resources are released and can be allocated to other tasks such as hazard surveillance. With increased experience, novice drivers no longer need to concentrate on their engine speed when deciding on the moment to change gear or on the coordinated sequence of accelerator pedal release and clutch pedal depression when doing so. They are increasingly able to think about the behavior of the traffic around them while maintaining vehicle speed and position seemingly without thinking about these relatively low-level actions. The three hypotheses emphasize steering control demands, vehicle control demands, and the driver’s situation
143
awareness. The first two hypotheses are closely related, with steering control being a special case of the demands of vehicle control. Mourant and Rockwell (1972) demonstrated that novices tend to look at the roadway closer to the vehicle than do experienced drivers, perhaps suggesting that they have not yet learned to use peripheral vision for steering control (Land & Horwood, 1995) or that the dynamics of perceptual-motor coordination are still being learned. If they need to look at road markers in order to keep their vehicle in the center of their lane, then they will have limited scope for looking at other objects in the roadway. The second hypothesis extends this view of the demands of vehicle control and sees central cognitive resources being occupied more generally, to the extent that the novice driver does not have the resources available for scanning the road scene and thereby acquiring new information about potential hazards. It has been established that varying the demand of the driving task will cause variations in the acquisition of information. Recarte and Nunes (2000) reported that mirror checking is reduced as the mental load on the driver is increased, and Underwood, Crundall, and Chapman (2002) also found that mirror checking varied with driving experience, with greater selectivity of the choice of mirror used by experienced drivers during lane changing. Similarly, as driving demands increase, fewer fixations on mirrors and other nonessential objects are reported (Schweigert & Bubb, 2001). This evidence from studies of the inspection of the information available in the mirrors suggests that as driving demands increase, experienced drivers re-allocate their cognitive resources and modify their intake of information about traffic in the roadway. The third suggestion is that perhaps novices stereotypically look straight ahead when driving because they have inadequate situation awareness. Differences in search patterns associated with driving experiencedspecifically, increased variance of fixation locations in more experienced driversdwould then be explained as a product of the knowledge base developed through previous traffic encounters. As drivers interact with other drivers and observe the behavior of other road users, they accumulate memories of events that happen on different kinds of roads, and they develop an awareness of their probability of happening. These situation-specific probabilities can help guide drivers through newly encountered environments if they are sufficiently similar to earlier circumstances for drivers to generalize their behavior (Shinoda, Hayhoe, & Shrivastava, 2001). Because of his or her limited exposure to varying roadway conditions, a new driver necessarily has an impoverished catalog compared to an experienced driver. Perhaps, when novices scanned a multilane highway to a lesser extent than the experienced drivers in Crundall and Underwood’s (1998) eye tracking study, they behaved like this because they were unaware of the special dangers associated with this particular type of road. They perhaps
144
had insufficient exposure to this kind of road with which to build a mental model of the behavior that might be shown by other vehicles. They would then be unable to predict where other vehicles would be a few seconds later, and they would not recognize the demands of negotiating interweaving lanes of traffic and the need to monitor not only the traffic ahead but also the lane-changing activity of traffic immediately to the rear. We sought evidence to discriminate between the vehicle control and situation awareness hypotheses by recording the scanning behavior of drivers in a laboratory task that eliminated the need for vehicle control (Underwood, Chapman, Bowden, & Crundall, 2002). Drivers sat in the laboratory and watched film clips recorded from a car as it traveled along the roads used by Crundall and Underwood (1998). One advantage of this approach is that each driver saw the same traffic conditions, whereas in the original study traffic conditions inevitably varied from moment to moment and from driver to driver. The laboratory task was essentially one of observation and prediction, with the task being to make a key-press response if they saw an event that would cause a driver to take evasive actiondessentially a hazard detection task that gave a reason for monitoring the video recordings carefully. The scanning behavior of new and experienced drivers was the principal interest, and while they watched the films, their eye movements were recorded. If new drivers have restricted search patterns because their resources are allocated to vehicle control, then eliminating the vehicle-control element of driving should result in a visual search pattern in the laboratory that is similar to that of an experienced driver. Take away the demands of maintaining vehicle speed and lane position and resources should become available for scanning, but only if the need for scanning is understood. If the search patterns of new drivers result from a mental model that does not inform them of the particular hazards associated with multilane highways, then they would continue to restrict their scanning while watching the roadway video recordings in the viewing-only task. The results indicated that the two groups of drivers were thinking about the scene differently, even when their resources were not occupied by the demands of vehicle control. Experienced drivers exhibited more extensive scanning when they watched more demanding sections of the roadway, whereas new drivers showed less sensitivity to changing traffic conditions. The eye tracking data indicated differences between new and experienced drivers that support the hypothesis that their inspection of the roadway varies not because they have differences in their mental resources residual from the task of vehicle control but, rather, because the novice drivers have an impoverished mental model of what other drivers might do on demanding roadways. Other research supports the hypothesis of limited situational awareness in novice
PART | II
Key Variables to Understand in Traffic Psychology
drivers. Jackson, Chapman, and Crundall (2009) asked drivers to predict events from driving video clips that were stopped at critical moments and obscured. They found that more experienced drivers were better able to predict the subsequent events, which they argued was akin to greater level 3 situational awareness (Endsley, 1995, 1999). Although these measures of the spread of search have provided some consistent and replicable findings, they need to be used with caution. For instance, in Figure 11.2, panels a and b reflect the typical scan path that one might expect from an experienced driver and novice driver, respectively. Calculating the variance or standard deviation of the x-axis fixation coordinates would reveal a clear difference between the two. However, the two scan paths shown in panels c and d would be indistinguishable on the basis of this simple calculation of spread. Imagine two pedestrians on either side of the roadway: In panel c, the driver extensively scans one of the pedestrians before shifting to the other, whereas in panel d the driver constantly switches between the two pedestrians. From a commonsense viewpoint, we might consider panel d to reflect greater spread of search. Certainly, most driver-training experts would consider panel d to represent a safer search strategy than panel c (Mills, 2005). However, the calculation of the variance of locations will not discriminate between these two strategies because it cannot take into account the sequential nature of the fixations. A further issue with this measure is that it will not discriminate between a decrease in the eccentricity of fixations to the left and right of the road ahead and a decrease in the frequency of these eccentric fixations. Thus, a reduction in the actual area of the visual scene that is scanned is indistinguishable from a reduction in the number of fixations away from the road ahead. In essence, a measure of spread is still informative about the extent to which drivers sample the visual scene, but it cannot be used to separate out the more subtle differences in visual strategies. However, these spread measures can be used in conjunction with range measures (e.g., mean saccade length, or how far the visual search extends in the scene) (Crundall et al., in press) to provide a more detailed picture. The failure of measures of spread to take into account the sequential nature of the fixations can also be overcome through scan path analysis. This type of analysis searches for statistical regularity in sequences of fixation locations using a transition matrix. Underwood, Chapman, Brocklehurst, Underwood, and Crundall (2003) recorded where drivers looked as a function of where they had looked immediately beforehand. Video recordings of fixations made in the Crundall and Underwood (1998) study were used as the input to the process of identifying fixation scan paths, and driver differences in the inspection of different roadways were again the focus of interest. The most interesting differences between drivers were again seen when
Chapter | 11
Visual Attention While Driving
145
FIGURE 11.2 Hypothetical scan patterns while driving a car. Panel a reflects a typical experienced driver, panel b reflects a typical novice, and panels c and d represent two scan paths that pose problems for simple calculations of spread of search.
they traveled along the multilane highway with merging trafficdthe most demanding section of the drive. The increased variance of experienced drivers from the earlier analysis was reflected in the scan paths, which showed very little consistency. There were few two-fixation scan paths that appeared regularly for these drivers, indicating that their fixation behavior was unpredictable statistically, and this can be explained by variations in traffic conditions from moment to moment prompting changes in fixations. As other vehicles appeared in the roadway or in the vehicle’s mirrors, the experienced drivers inspected them and evaluated their trajectories. There was no consistency in the location of one fixation according to where they had looked previously. The new drivers, on the other hand, showed a remarkable consistency that can be summarized by
a simple generalization: Wherever they had looked previously, the next place they looked was at the roadway straight ahead of them. Their fixation behavior was stereotyped and not sensitive to the variations in traffic behavior seen on a highway with fast-moving vehicles that are regularly changing lanes, both ahead and to the rear. The low variance of fixations recorded by Crundall and Underwood was a product of the new drivers repeatedly moving their eyes to inspect the roadway directly before them.
5. CONCLUSIONS The use of eye tracking measures has greatly increased our understanding of how driving skills develop and what strategies drivers employ to ensure a safe journey. Eye
146
movement analyses are now being applied to help understand specific accident types (e.g., spotting overtaking motorcycles; Shahar, van Loon, Clarke, & Crundall, in press) and are forming the basis of training interventions to decrease accident rates (Chapman, Underwood, & Roberts, 2002; Pradhan, Pollatsek, Knodler, & Fisher, 2009). It is in these areas that the most exciting advances are being made. However, there are caveats that need to be mentioned. First, we have noted in this chapter that eye movement measures, when taken in isolation, are open to errors of interpretation. Simply inferring safe driving on the basis of whether one has looked at a particular area of the scene (Pradhan et al., 2005) may not be sufficient because extremely short fixations may not be indicative of full processing (as with “look but fail to see” errors; Crundall et al., in press). Similarly, we noted that measures of spread may be unable to identify certain visual strategies. Thus, it seems that multiple measures of visual behavior should be taken to ensure that the potential confounds associated with one particular measure do not dominate the conclusions. Second, we need to relate all measures back to the context in which they were collected. It is useful to relate measures of vision in driving to other fields of research such as reading, although it must always be borne in mind that the complex context of real driving is unlikely to be paralleled in an analogous laboratory. Thus, although reading research provides us with a framework with which to interpret increased fixation durations in terms of processing difficulty, this analogy does not necessarily transfer to the rural road, where fixations on the focus of expansion can be extremely large. The number of variables that can influence the patterns of eye movements during driving seem too many to document, but that has not stopped attempts to do so. Indeed, during approximately the past 10 years, great strides have been made in our understanding of how various factors interact, although we must remember that there are other potential aspects of the context that we have not accounted for when drawing conclusions. Finally, when considering the potential for designing training interventions to encourage eye movements, particularly in young and novice drivers, we must be aware of the developmental limitations on eye movement strategies. For instance, Mourant and Rockwell (1972) found that novice drivers look more at lane markings than do more experienced drivers; Land and Horwood (1995) demonstrated that experienced drivers still use the information provided by lane markers but do so through peripheral vision; and Crundall, Underwood, and Chapman (1999, 2002) demonstrated that inexperienced drivers have fewer resources devoted to peripheral vision than do more experienced drivers. Taken together, this body of work suggests that although lane markings are of vital importance to maintaining the lateral position of the vehicle,
PART | II
Key Variables to Understand in Traffic Psychology
inexperienced drivers do not necessarily have the available resources to devote to peripheral vision in order to extract lane marker information without foveating them. Thus, an intervention strategy that directly or indirectly trains inexperienced drivers to focus less on lane markings may have the unintended effect of impairing lane maintenance. Despite these caveats, it is clear that studies of eye movements have provided considerable insights into the driving process and have achieved moderate success in rudimentary training situations (Chapman et al., 2002; Pradhan et al., 2005). As eye tracking technology continues to improve and costs are reduced, the use of these systems in future research will increase, and we hope that this brief discussion provides some ideas on how to best employ this technology for current and future uses.
REFERENCES Borowsky, A., Shinar, D., & Oron-Gilad, T. (2010). Age, skill and hazard perception in driving. Accident Analysis and Prevention, 42, 1240e1249. Brockmole, J. R., & Boot, W. R. (2009). Should I stay or should I go? Attentional disengagement from visually unique and unexpected items at fixation. Journal of Experimental Psychology: Human Perception and Performance, 35, 808e815. Brown, I. D. (2002). A review of the “look but failed to see” accident causation factor. In Behavioural Research in Road Safety XI. London: Department for Transport, Local Government and the Regions. Caplovitz, G. P., Fendrich, R., & Hughes, H. C. (2008). Failures to see: Attentive blank stares revealed by change blindness. Consciousness and Cognition, 17, 877e886. Chapman, P., & Underwood, G. (1998). Visual search of driving situations: Danger and experience. Perception, 27, 951e964. Chapman, P., Underwood, G., & Roberts, K. (2002). Visual search patterns in trained and untrained novice drivers. Transportation Research Part F: Traffic Psychology and Behaviour, 5, 157e167. Clarke, D. D., Ward, P., Bartle, C., & Truman, W. (2006). Young driver accidents in the UK: The influence of age, experience, and time of day. Accident Analysis and Prevention, 38, 871e878. Clarke, D. D., Ward, P., Bartle, C., & Truman, W. (2007). The role of motorcyclist and other driver behaviour in two types of serious accident in the UK. Accident Analysis and Prevention, 39, 974e981. Coyne, P. (1997). Roadcraft: The essential police driver’s handbook. London: The Stationary Office. Crundall, D., Crundall, E., Clarke, D., and Shahar, A. (in press). Why do car drivers violate the right-of-way of motorcycles at t-junctions? Accident Analysis and Prevention. Crundall, D., & Underwood, G. (1998). Effects of experience and processing demands on visual information acquisition in drivers. Ergonomics, 41, 448e458. Crundall, D., Underwood, G., & Chapman, P. (1999). Driving experience and the functional field of view. Perception, 28, 1075e1087. Crundall, D., Underwood, G., & Chapman, P. (2002). Attending to the peripheral world while driving. Applied Cognitive Psychology, 16, 459e475.
Chapter | 11
Visual Attention While Driving
Department for Transport. (2010a). Road traffic by vehicle type, Great Britain: 1950e2009 (miles). http://www.dft.gov.uk/pgr/statistics/ datatablespublications/roads/traffic/annual-volm/tra0101.xls. Accessed November 11, 2010. Department for Transport. (2010b). Road traffic and speed statisticsd2009. http://www.dft.gov.uk/pgr/statistics/datatablespublications/roads/traffic. Accessed November 11, 2010. Duchowski, A. T. (2007). Eye tracking methodology: Theory and practice (2nd ed.). New York: Springer-Verlag. Ehrlich, S. E., & Rayner, K. (1981). Contextual effects on word perception and eye movements during reading. Journal of Verbal Learning and Verbal Behavior, 20, 641e655. Endsley, M. R. (1995). Toward a theory of situation awareness in dynamic systems. Human Factors, 37(1), 32e64. Endsley, M. R. (1999). Situation awareness and human error: Designing to support human performance. In Proceedings of the high consequences system surety conference, Albuquerque, NM. Gugerty, L. (1997). Situation awareness during driving: Explicit and implicit knowledge in dynamic spatial memory. Journal of Experimental Psychology: Applied, 3, 42e66. Haley, S. (2006). Mind driving. Croydon, UK: Safety House. Henderson, J. M. (1992). Identifying objects across saccades: Effects of extrafoveal preview and flanker object context. Journal of Experimental Psychology: Learning, Memory, and Cognition, 18, 521e530. Henderson, J. M., & Hollingworth, A. (1999). The role of fixation position in detecting scene changes across saccades. Psychological Science, 10, 438e443. Horswill, M. S., & McKenna, F. P. (2004). Drivers’ hazard perception ability: Situation awareness on the road. In S. Banbury, & S. Tremblay (Eds.), A cognitive approach to situation awareness. Aldershot, UK: Ashgate. Jackson, A. L., Chapman, P., & Crundall, D. (2009). What happens next? Predicting other road users’ behaviour as a function of driving experience and processing time. Ergonomics, 52(2), 154e164. Just, M. A., & Carpenter, P. A. (1980). A theory of reading: From eye fixations to comprehension. Psychological Review, 87(4), 329e354. Klauer, S. G., Dingus, T. A., Neale, V. L., Sudweeks, J. D., & Ramsey, D. J. (2006). The impact of driver inattention on near-crash/crash risk: An analysis using the 100-Car Naturalistic Driving Study data. Washington, DC: National Highway Traffic Safety Administration. Konstantopoulos, P., Crundall, D., & Chapman, P. (2010). Driver’s visual attention as a function of driving experience and visibility. Using a driving simulator to explore visual search in day, night and rain driving. Accident Analysis and Prevention, 42(Special issue), 827e834. Kotowicz, A., Rutishauser, U., & Kock, C. (2010). Time course of target recognition in visual search. Frontier in Human Neuroscience, 4, 31. Land, M. F., & Horwood, J. (1995). Which parts of the road guide steering? Nature, 377, 339e340. Lee, J. D. (2008). Fifty years of driving safety research. Human Factors, 50, 521e528. Lestina, D. C., & Miller, T. R. (1994). Characteristics of crash-involved younger drivers. In 38th Annual proceedings of the Association for the Advancement of Automotive Medicine (pp. 425e437). Des Plaines, IL: Association for the Advancement of Automotive Medicine. Liversedge, S. P., & Findlay, J. M. (2000). Saccadic eye movements and cognition. Trends in Cognitive Sciences, 4, 6e14.
147
Mills, K. C. (2005). Disciplined attention: How to improve your visual attention when you drive. Chapel Hill, NC: Profile Press. Mourant, R. R., & Rockwell, T. H. (1972). Strategies of visual search by novice and experienced drivers. Human Factors, 14, 325e335. Organisation for Economic Co-operation and Development/European Conference of Ministers of Transport. (2006). Young drivers: The road to safety. Paris: Organisation for Economic Co-operation and Development. Pradhan, A. K., Hammel, K. R., DeRamus, R., Pollatsek, A., Noyce, D. A., & Fisher, D. L. (2005). Using eye movements to evaluate effects of driver age on risk perception in a driving simulator. Human Factors, 47, 840e852. Pradhan, A. K., Pollatsek, A., Knodler, M., & Fisher, D. L. (2009). Can younger drivers be trained to scan for information that will reduce their risk in roadway traffic scenarios that are hard to identify as hazardous? Ergonomics, 52, 657e673. Rayner, K. (1998). Eye movements in reading and information processing: Twenty years of research. Psychological Bulletin, 124, 372e422. Recarte, M. A., & Nunes, L. M. (2000). Effects of verbal and spatialimagery tasks on eye fixations while driving. Journal of Experimental Psychology: Applied, 6, 31e43. Reichle, E. D., Rayner, K., & Pollatsek, A. (2003). The E-Z Reader model of eye movement control in reading: Comparisons to other models. Behavioral and Brain Sciences, 26, 445e476. Reingold, E. M., & Rayner, K. (2006). Examining the word identification stages hypothesized by the E-Z Reader model. Psychological Science, 17, 742e746. Sabey, B.E., & Staughton, G. C. (1975). Interacting roles of road environment, vehicle and road user in accidents. In Proceedings of the fifth international conference of the International Association for Accident and Traffic Medicine. Schweigert, M., & Bubb, H. (2001, August). Eye movements, performance and interference when driving a car and performing secondary tasks. Brisbane, Australia: Paper presented at the Vision in Vehicles 9 conference. Shahar, A., van Loon, E., Clarke, D., & Crundall, D. (in press). Attending overtaking cars and motorcycles through the mirrors before changing lanes. Accident Analysis and Prevention. Shinar, D. (2008). Looks are (almost) everything: Where drivers look to get information. Human Factors, 50, 380e384. Shinoda, H., Hayhoe, M. M., & Shrivastava, A. (2001). What controls attention in natural environments? Vision Research, 41, 3535e3545. Sivak, M. (1996). The information that drivers use: Is it indeed 90% visual? Perception, 25, 1081e1089. Treat, J. R., Tumbas, N. S., McDonald, S. T., Shinar, D., Hume, R. D., Mayer, R. E., Stansifer, R. L., & Castellan, N. J. (1979). TRI-level study of the causes of traffic accidents: Final report. (Report No. DOT HS 805-085). Washington, DC: U.S. Department of Transportation, National Highway Traffic Safety Administration. Underwood, G. (2007). Visual attention and the transition from novice to advanced driver. Ergonomics, 50, 1235e1249. Underwood, G., Chapman, P., Bowden, K., & Crundall, D. (2002). Visual search while driving: Skill and awareness during inspection of the scene. Transportation Research Part F: Traffic Psychology and Behaviour, 5, 87e97. Underwood, G., Chapman, P., Brocklehurst, N., Underwood, J., & Crundall, D. (2003). Visual attention while driving: Sequences of eye
148
fixations made by experienced and novice drivers. Ergonomics, 46, 629e646. Underwood, G., Chapman, P., & Crundall, D. (2009). Experience and visual attention in driving. In C. Castro (Ed.), Human factors of visual and cognitive performance in driving (pp. 89e116). Boca Raton, FL: CRC Press. Underwood, G., Crundall, D., & Chapman, P. (2002). Selective searching while driving: The role of experience in hazard detection and general surveillance. Ergonomics, 45, 1e12. Underwood, G., Crundall, D., & Chapman, P. (2007). Cognition and driving. In F. Durso (Ed.), Handbook of applied cognition (2nd ed). (pp. 391e414). New York: Wiley. Underwood, G., & Everatt, J. (1992). The role of eye movements in reading: Some limitations of the eyeemind assumption. In
PART | II
Key Variables to Understand in Traffic Psychology
E. Chekaluk, & K. R. Llewellyn (Eds.), The role of eye movements in perception. Amsterdam: North-Holland. Underwood, G., Phelps, N., Wright, C., van Loon, E., & Galpin, A. (2005). Eye fixation scanpaths of younger and older drivers in a hazard perception task. Ophthalmic and Physiological Optics, 25, 346e356. Underwood, G., Templeman, E., Lamming, L., & Foulsham, T. (2008). Is attention necessary for object identification? Evidence from eye movements during the inspection of real-world scenes. Consciousness & Cognition, 17, 159e170. Van Gompel, R. P. G., Fischer, M. H., Murray, W. S., & Hill, R. L. (2007). Eye-movement research: An overview of current and past developments. In R. P. G. van Gompel, M. H. Fischer, W. S. Murray, & R. L. Hill (Eds.), Eye movements: A window on mind and brain (pp. 1e28). Oxford: Elsevier.
Chapter 12
Social, Personality, and Affective Constructs in Driving Dwight Hennessy Buffalo State College, Buffalo, NY, USA
1. INTRODUCTION Driving is more than the mechanical operation of a vehicle, as a means of movement between destinations. Rather, it is a complex process involving individual factors expressed within a social exchange among drivers, passengers, and pedestrians, which is ultimately impacted by contextual and environmental stimuli found inside and outside the vehicle. Rotton, Gregory, and Van Rooy (2005) argued that, traditionally, the focal point of most traffic research has been the individual in this system, often at the expense of situational determinants. This is perhaps due to the fundamental attribution error, which is the tendency to explain the actions of others in terms of personal causes even when situational factors are evident. The driver is a central component of this system, but only one component, whose thoughts, feelings, and actions are shaped and directed by the micro and macro context. In many respects, the traffic environment is a distinct and intriguing setting. There is a degree of speed and anonymity not found in other contexts, with huge discrepancies in history, experience, or skill level among drivers; subtle forms and means of communication that often have multiple interpretations; a unique blend of written and unwritten rules that can vary across locations; and ultimately a high degree of danger. This uniqueness, in addition to the widespread application potential and relevance to the general public, has made the traffic environment an attractive context for social research. This chapter focuses on the components of this personesituation system and how they have been shown, individually and in combination, to impact driving behavior (e.g., rule violations and collision), personal outcomes (e.g., health, mood, stress, and fatigue), and interpersonal interactions (e.g., aggression and judgments of others).
Handbook of Traffic Psychology. DOI: 10.1016/B978-0-12-381984-0.10012-8 Copyright Ó 2011 Elsevier Inc. All rights reserved.
2. PERSONAL FACTORS AND DRIVING OUTCOMES Given that it is “people” who drive, it would seem rational that any discussion of factors that impact driving outcomes would include personal factors. It would also seem logical to assume that drivers are all ultimately unique, and that at any given moment on the roadway, there is immense variability in driving styles, learning experiences, collision history, and expectations/judgments. However, this does not preclude the search for patterns that identify categories of drivers who might be more or less at risk for negative outcomes on the roadway over time. Although there are numerous categories of individual differences that can impact driving, traffic psychology has often focused on personality variables. In fact, personality has been used to qualify many demographic and affective factors, such as gender (often associated with a masculinity trait) and driver anger (often examined as a trait disposition). Although there are many definitions of personality, most share the notion that it involves the consistent pattern of thoughts, feelings, and actions that emerge with some level of stability across time and context. Traffic psychology has typically relied on the “trait” approach, in which the focus has been on individual characteristics that combine or cluster to determine the overall expression of personality. As a result, certain traits are believed to be inherently more dangerous than others in the traffic environment. Those who possess more of these or in a more dangerous balance are believed to be a greater risk to self and others. However, it is also a matter of “degree” in that each trait is identified along a continuum of its “strength,” where some characteristics may have a much stronger impact on the overall personality, and those that possess dangerous traits to a higher degree are most problematic. 149
150
2.1. A Case for Personality in Understanding Negative Driving Outcomes A number of researchers have attempted to link personality to negative driving outcomes, particularly collisions, with mixed results. One reason for such discrepancies may be the selection and number of personality factors used in prior traffic research. Personality and its behavioral outcomes may be most accurately reflected by combinations or clusters of traits rather than by individual components. In this respect, personality and its expression are multidimensional. Thus, selection of unitary constructs or perhaps the “wrong” blend of personality characteristics may give the impression that it is not predictive of an outcome as rare and complex as collisions. Another reason may be that personality represents an indirect rather than direct link with collisions. Beirness (1993) argued that personality on its own is a poor predictor of collisions but instead interacts with more “proximal” factors that often involve current, state, or pressing drivingrelated factors. Su¨mer (2003) provided an excellent review of this research and proposed a model in which personality represents one of several possible distal factors (in addition to other enduring personal, cognitive, and situational factors, such as culture, vehicle condition, and attributions) that impact more immediate proximal factors composed of driving style and transitory factors (e.g., violations, errors, and safety skills) that then influence collisions. In this respect, personality is an important focal point in traffic research given its impact on the more macro, persistent, or ambient distal level (e.g., the approach to driving, personal tendencies, and beliefs about other drivers), as well as on the immediate and transitory proximal influences (e.g., altering state interpretations of actual driving behavior, state emotional experiences, and negative driving behaviors). The following sections provide a brief examination of personality factors that have been linked to negative driving outcomes. Although this list is far from comprehensive, it represents factors commonly observed in traffic psychology.
2.2. Sensation Seeking One of the most widely studied personality predictors of negative driving behavior is sensation seeking, which is defined as a trait characterized by the pursuit of novel, diverse, and extreme experiences. To achieve these goals, high sensation seekers often display a willingness to take disproportionate physical and social risks (Zuckerman, 1994). Driving provides an excellent opportunity for high sensation seekers to satisfy the desire for sensation given the inherent potential for arousal, excitement, danger, speed, and competition. Jonah (1997) reviewed 40 studies on sensation seeking in drivers and concluded that it can
PART | II
Key Variables to Understand in Traffic Psychology
increase dangerous driving in a number of areas, including impaired driving, speeding, and seat belt use. In fact, a relationship in the range of 0.3 to 0.4 was found in all but 4 studies he examined. Sensation seeking is typically measured using Zuckerman’s Sensation Seeking Scale (SSS) Form V, which is further divided into four subscales identifying boredom susceptibility (BS), thrill and adventure seeking (TAS), disinhibition (DIS), and experience seeking (ES). However, comparisons across studies applying sensation seeking to driving behavior are difficult due to the fact that researchers often do not use the entire SSS, instead opting for selective subscales (thrill seeking being the most typical), truncation of items from subscales, combinations of items with selfgenerated or other unrelated items, or a total sensation seeking score that is undifferentiated across subscales. Another tool that is gaining in popularity is the Arnett Inventory of Sensation Seeking (AISS), which was designed in response to perceived limitations in Zuckerman’s SSS. Most notably, Arnett (1994) believed that several items were culturally irrelevant and also contained risky behavior-based items that were confounded with many of the activities that were typically used as outcome variables in research, including drinking and taking drugs. The AISS contains two subscales based on a need for novelty and intensity of stimulation. Previous research has established that the AISS does measure a similar dimension of sensation seeking as that measured by the SSS and predicts dangerous and risky behavior (Andrew & Cronin, 1997; Arnett, 1994). Despite their differences, both the AISS and the SSS have been linked to negative driving outcomes. According to Zuckerman (2007), the SSS predicts reduced risk appraisal and elevated risk taking, including reckless driving. Jonah, Thiessen, and Au-Yeung (2001) argued that sensation seeking does appear to increase dangerous driving patterns, including aggressive (e.g., swearing, yelling, and horn honking) and high-risk activities (e.g., weaving and speeding). Interestingly, this trend is evident even among predrivers (Waylen & McKenna, 2008). Specifically, using eight items from the AISS, high sensation seeking among boys aged 11e16 years was related to favorable attitudes toward risky road use. Other negative driving activities that have been linked to sensation seeking include tailgating and speeding (using TAS & BS subscales; Harris & Houston, 2010), lack of seat belt use and unsafe passing (using TAS & DIS subscales; Dahlen & White, 2006), and ignorance of traffic rules (using newly generated items; Iversen & Rundmo, 2002). Sensation seeking has also been found to predict elevated convictions (using TAS & BS subscales; Matthews, Tsuda, Xin, & Ozeki, 1999), self-report traffic violations (using total SSS score; Schwerdtfeger, Heims, & Heer, 2010), drinking and driving (using all four SSS factors; Greene,
Chapter | 12
Social, Personality, and Affective Constructs in Driving
Krcmar, Walters, Rubin, & Hale, 2000), and driving under the influence of cannabis (using total SSS; Richer & Bergeron, 2009). Furthermore, using TAS and DIS ˚ berg (1999) determined that the subscales, Rimmo¨ and A violations subscale of the DBQ (which measures a dispositional tendency to willfully commit violations) mediated the relationship between sensation seeking and self-report violations and collisions. Interestingly, Wong, Chung, and Huang (2010) advocated a positive effect of sensation seeking in that high sensation seekers (using self-derived items) were involved in fewer motorcycle collisions due to the regulating effect of confidence in their riding ability. However, the collisions they experienced were more severe than those of low sensation seekers. Some critics have argued that caution must be exercised when interpreting reported associations between sensation seeking and risky behaviors due to a high degree of conceptual overlap between the two constructs (Arnett & Balle-Jensen, 1993). A risk-taking lifestyle may naturally predispose one to seek stimulating activities; hence, the psychological phenomenon of sensation seeking may overlap with the behavioral predictor of risk taking. However, Burns and Wilde (1995) contended that sensation seekers may, in fact, seek activities that are not inherently risky. For example, sensation seekers may drive fast and experience heightened arousal yet remain within the bounds of safety by not taking excessive risks (e.g., driving on higher speed highways). Another consideration is the fact that sensation seeking is strongly associated with age and developmental processes, which peak in late adolescence and begin to decline thereafter (Arnett & Balle-Jensen, 1993). This is especially important in traffic research because most studies that focus on dangerous or risky outcomes of sensation seeking tend to concentrate on younger drivers, for whom sensation seeking, risky tendencies, and dangerous driving are concurrently heightened. In addition, younger drivers are typically the least experienced and may engage in more dangerous activity not because they are sensation seekers but, rather, because they lack the practical knowledge, speed of processing, or sensitivity to potential danger that come with practice. In this sense, the link between sensation seeking and risky or dangerous driving may be overstated. However, the fact that sensation seeking is still linked to negative driving outcomes among older drivers may provide impetus for its continued use in identifying a category of at-risk drivers.
2.3. Trait Aggression The issue of aggressive driving has received a great deal of attention from both scientific and public communities. However, there are differing conceptualizations of the term “aggressive,” which has made it difficult to interpret and
151
compare conclusions in this area. Following the motivational approach to aggression (Berkowitz, 1993), many have treated traffic aggression as actions intended to physically, psychologically, or emotionally harm another within the driving environment, including drivers, passengers, and pedestrians (Hennessy & Wiesenthal, 1999; Shinar, 1998). Examples include yelling, swearing, purposely tailgating, leaning on the horn, and roadside confrontations. In contrast, others have approached aggressive driving as any dangerous driving behavior regardless of intent, such as speeding, weaving through traffic, and using the shoulder to pass. However, several studies have found that these latter actions, which more closely resemble highway code violations or assertiveness, are distinct compared to intentionally harmful “aggressive violations.” Drivers who intentionally cause other motorists harm are different in important ways from those who break traffic rules, even though these latter actions are sometimes selfish, illegal, or dangerous to others (Hennessy & Wiesenthal, 2005; Lawton, Parker, Manstead, & Stradling, 1997). For the purpose of this chapter, discussions of driver aggression focus on the motivational definition emphasizing intentional harm. One important limitation to this approach is that intention is often difficult to determine, especially when self-report measures are used. This is exaggerated by the fact that several actions that are often considered as “aggressive” can also have different meanings between cultures or could be initiated for reasons other than harm. For example, Lajunen, Parker, and Summala (2004) noted that horn honking is typically considered aggressive in Scandinavia but as a message in southern Europe. Nonetheless, research has found that aggressive drivers typically show patterns of increased frequency and duration, as well as decreased latency of such actions (e.g., honk more often, for longer, and quicker in response instigation), and that they concurrently engage in a pattern of multiple forms of aggression, all of which would be more indicative of aggression than signaling behavior (Doob & Gross, 1968; Gulian, Matthews, Glendon, & Davies, 1989; Hennessy & Wiesenthal, 2005). Although aggression is predicted by numerous personal, social, cognitive, and environmental factors (Berkowitz, 1993), there are some individuals who demonstrate a “trait”-like proclivity toward more frequent and severe acts of aggression. From this perspective, some drivers develop trait driver aggression tendencies (Hennessy & Wiesenthal, 1999), which represents a unique personal quality from those who seek to take risks or gain excitement from driving. Rather, trait driver aggression appears to involve an elevated tendency to misperceive the actions and intentions of other drivers as hostile and threatening; to become frustrated or irritated by the actions of others; and to be willing to “pay back” offending drivers
152
through physical, emotional, and psychological harm (Gulian, Matthews, et al., 1989; Hennessy & Wiesenthal, 2001a). One explanation for the development of trait driver aggression is grounded in learning theories. Specifically, acting in an aggressive manner, when viewed as “successful,” can be negatively reinforcing due to the removal of the source of irritation, frustration, and conflict and/or positively reinforcing through the addition of feelings of control, power, and dominance. Over time, these actions are then considered as more acceptable, beneficial, and potentially successful during similar subsequent situations. As a result, aggression becomes more customary for that driver as such responses increase in his or her response hierarchy. With repeated reinforcement, and concurrent lack of negative repercussions, driver aggression becomes acceptable or normative to that driver’s behavioral repertoire. Trait aggression has been linked to a number of dangerous outcomes, including state aggressiveness and violations. Britt and Garrity (2006) had participants recall past driving events and indicate how they would react, and the authors found that dispositional aggression predicted angry and aggressive responses. Using selfreport measures, Hennessy and Wiesenthal (2002) found that trait driver aggression was related to traffic (highway code) violations, which is consistent with the results of King and Parker (2008), who reported that trait aggressive drivers were more likely to commit traffic violations and to falsely underestimate the frequency of their violations in comparison to others. Similarly, Maxwell, Grant, and Lipkin (2005) found a relationship between trait driver aggression and traffic violations in a sample of UK drivers, whereas Kontogiannis, Kossiavelou, and Marmaras (2002) also noted that speeding convictions and general law breaking were predicted by a tendency to commit aggressive violations among Greek drivers. The link between trait aggression and collisions may be more complicated, perhaps due to the relative rarity of collisions. Li, Li, Long, Zhan, and Hennessy (2004) found that self-reported trait driver aggression was related to active collisions, as well as traffic violations, among Chinese drivers. However, they cautioned that their sample was predominantly males, which may overestimate this ¨ zkan, and Lajunen (2008) also relationship. Bener, O revealed that driver aggression predicted collisions in Qatar, although their aggressive construct also contained elements of speeding competition. Gulliver and Begg (2007) utilized face-to-face interviews with drivers and identified a relationship between trait aggression and crashes, but only among men and after controlling for exposure. According to Su¨mer (2003), the link between trait aggression and collisions may, in fact, be indirect
PART | II
Key Variables to Understand in Traffic Psychology
because it is mediated by an aberrant driving style (i.e., elevated errors and violations).
2.4. Negative Emotions and Trait Anger Negative emotions experienced while driving may alter cognitions of traffic-related stimuli, narrow attentional focus or processing, and modify interpretations of other drivers and their activities, thus eventually increasing the potential for risky, harmful, or unsafe driving (Mesken, Haganzieker, Rothengatter, & de Waard, 2007). Shahar (2009) found that anxiety can have a negative impact on the number of errors, lapses, and ordinary violations, largely due to distraction and attention deficits. Similarly, Oltedal and Rundmo (2006) demonstrated that anxiety may have an indirect relationship to collisions through its link with risky driving behavior. However, others have argued that anxiety can serve a positive role in driving through increased cautiousness and alertness (Stephens & Groeger, 2009). Other emotions, such as sadness, may have an indirect impact on driving outcomes as well. According to Peˆcher, Lemercier, and Cellier (2009), drivers who listened to sad music in a simulator felt calmer but were unable to focus as much attention on the driving task. Similarly, Bulmash et al. (2006) found slower steering reaction times and a higher number of accidents among depressed participants in a simulator. The most commonly examined emotion in driving research has been anger. Using a diary methodology, Underwood, Chapman, Wright, and Crundall (1999) revealed that 85% of their sample of UK drivers (104 participants recruited from the general driving population through media advertisements and from driving test centers) reported experiencing anger at least once during their routine daily driving excursions during a 2-week period. According to Speilberger (1988), there is a clear distinction between state and trait anger. State anger represents an emotional experience, from mild annoyance to extreme fury, in a given context. Trait anger characterizes an enduring tendency to experience state anger more frequently, intensely, and with greater duration. Deffenbacher and colleagues have advocated a similar distinction in the driving environment where trait driver anger represents a special context-specific form of trait anger that predicts more frequent and intense anger in actual driving conditions (Deffenbacher, Huff, Lynch, Oetting, & Salvatore, 2000). The most widely used measurement tool to date has been the Driving Anger Scale (DAS; Deffenbacher, Oetting, & Lynch, 1994), in which participants are asked to rate their likely degree of anger to common anger-provoking driving situations, leading to an average score believed to differentiate drivers on a continuum of trait driver anger.
Chapter | 12
Social, Personality, and Affective Constructs in Driving
The DAS has been translated into several languages and used in a number of different countries, with a general degree of success. Despite variability in factor structure and unique predictive ability, there has been general acceptance in its ability to measure a “trait” anger characteristic. For example, Deffenbacher et al. (1994) and Yasak and Esiyok (2009) found six factors using a U.S. and Turkish sample, respectively, whereas Lajunen, Corry, Summala, and Hartley (1998) identified three factors in a UK sample, Bjo¨rklund (2008) found a similar three-factor version using a Swedish translation, and Sullman (2006) uncovered four factors in a New Zealand sample. Also, Sullman found age and gender differences in DAS scores, whereas Yasak and Esiyok did not. Despite discrepancies, trait driver anger has been linked to a number of dangerous driving behaviors, including violations (Lajunen et al., 1998), lost concentration, poor driving control, and close calls (Dahlen, Martin, Ragan, & Kuhlman, 2005). Some studies have linked driving anger to collision, but others have not. Deffenbacher et al. (2000) found that driving anger was related to greater self-reported lifetime collisions and minor collisions during the past year, as well as greater moving violations. In contrast, according to Deffenbacher, Lynch, Oetting, and Yingling (2001), although trait anger was related to low levels of concentration, poor vehicle control, and close calls, it did not predict collisions or violations. Similarly, Sullman (2006) found that the DAS was unrelated to self-reported collisions during the past 5 years. This incongruity may be due to the relative rarity of collisions as an outcome measure, difficulties in collecting collision data accurately (particularly self-reports of active or at-fault collisions), as well as numerous indirect and contextual influences on collisions. Deffenbacher, Deffenbacher, Lynch, and Richards (2003) also identified aggression as a consequence of elevated driving anger independent of demographic variables. However, others would contend that the link between anger and state aggression is not always strong or direct. Considering the total volume of drivers and distance driven per day, and numerous potential moderating and mediating factors in the traffic environment, aggression may not be an overly common event (Parker, Lajunen, & Summala, 2002). In this respect, Neighbors, Vietor, and Knee (2002) advocated only a modest link between anger and aggression. It could be maintained that the strength and nature of such a relationship ultimately depend on a number of interacting personal and situational factors, such as the type and nature of interactions, the degree of frustration, motivation, and public self-consciousness (Lajunen & Parker, 2001; Millar, 2007). One consideration concerning driving anger research is the fact that some researchers neglect to differentiate state from trait anger, to the extent that some treat their trait
153
measurement as a state outcome. Although trait anger may predict greater state anger in driving situations, it is not deterministic. Another concern is the fact that a metaanalytic review of the angereaggression link by Bettencourt, Talley, Benjamin, and Valentine (2006) showed that trait anger predicts aggression only in hostile situations. Similarly, Wilkowski and Robinson (2010) argued that anger predicts reactive forms of aggression. This might account for the contention from Van Rooy, Rotton, and Burns (2006) that the DAS and Driving Vengeance Questionnaire (DVQ) are indistinguishable from one another based on their extremely high correlation. The DVQ was designed as a measure of attitudes toward vengeance in the driving environment, where those who have been wronged or provoked by other drivers would seek revenge through reactive types of aggression (Wiesenthal, Hennessy, & Gibson, 2000). Hence, the high correlation found by Van Rooy et al. (2006) may be due to confounding by reactive aggression. However, it could just as easily be argued that trait vengeful drivers become more easily and intensely angered and as a result are more inclined to respond to provocation from others through aggression.
2.5. Trait Driver Stress Susceptibility Most traffic research has treated driver stress as the outcome of a negative cognitive appraisal of driving situations (Glendon et al., 1993; Hennessy & Wiesenthal, 1997). It is only when driving is interpreted as demanding or overly taxing that stress manifests itself in psychological (e.g., anxiety and negative mood; Gulian, Matthews, et al., 1989; Wiesenthal, Hennessy, & Totten, 2000), cognitive (e.g., task-relevant cognitive interference and loss of attention; Desmond & Matthews, 2009; Matthews, 2002), or physical symptoms (e.g., increased heart rate and blood pressure; Stokols, Novaco, Stokols, & Campbell, 1978). In this respect, trait driver stress has immediate and long-term consequences to drivers. According to the transactional model of driver stress (Matthews, 2002), trait stress susceptibility is an important personal factor that can potentially interact with the situation to impact driving outcomes. This model holds that driver stress is the product of the dynamic interaction between personal and environmental factors that are mediated by cognitive processes (e.g., interpretation of events and selection of coping resources). Repeated stressful experiences or ineffective coping strategies can lead to feedback that dynamically alters the transactional process and eventual behavior of that driver. For example, repeated stressful outcomes may lead to a generalized trait driver stress susceptibility, which may then impact the driving activities and resulting situational factors that might be encountered, subsequently altering perceptions and
154
interpretations of those events and ultimately impacting the state emotional experience in actual conditions. Based on this model, Gulian, Matthews, et al. (1989) developed the Driving Behaviour Inventory (DBI) as a tool to measure the factors, or characteristics, associated with appraisal styles and coping behaviors reflective of driver stress. It consists of both a five-factor solution and a single “general” stress factor solution. The five-factor solution includes three main factors labeled as Dislike of Driving, Driving Aggression, and Alertness, as well as two minor factors identified as Irritation When Overtaken and Frustration When Overtaking. The General factor is composed of a much smaller set of items from the five-factor solution. Although the General factor has been found to predict state driver stress and negative emotion in actual driving conditions (Hennessy & Wiesenthal, 1997), and has been linked to speeding convictions and minor accidents (Matthews, Dorn, & Glendon, 1991), the three main factors from the five-factor solution are most commonly used in research. Dislike of Driving (DIS) represents items related to anxiety, unhappiness, and lack of confidence while driving, particularly when driving conditions are difficult. DIS has been found to predict negative mood in a simulated driving task (Dorn & Matthews, 1995) as well as postcommute tension and depression (Matthews et al., 1991). High DIS drivers show greater impaired lateral control (Matthews, Sparkes, & Bygrave, 1996) and mistakes and lapses (Kontogiannis, 2006), but they also perceive themselves as less skilled, showing reduced self-driving confidence leading to greater cautiousness, slower speeds, and fewer speeding tickets (Matthews et al., 1991, 1998). Driving Aggression (AG) items focus on reactions to irritation in the driving context and impatience shown as a result of impedance from others. Those high on the AG factor demonstrate more dangerous driving patterns, including control errors when overtaking (Matthews et al., 1998); mistakes, lapses, and speeding (Kontogiannis, 2006); and tailgating, confrontation, negative evaluation of other drivers, and minor collisions (Matthews et al., 1991). The Alertness (AL) factor predominantly measures a tendency to monitor the driving situation for hazards. Previous research has found that AL is considerably lower in reliability than the DIS or AG factors (Glendon et al., 1993; Lajunen & Summala, 1995) and has weak predictive relationships with behavioral and emotional outcomes (Matthews et al., 1998). This may partially account for conflicting results in research attempting to link AL with negative driving outcomes. For example, although AL has been associated with reduced collisions (Matthews, Desmond, Joyner, Carcardy, & Gilliland, 1997) and fewer speeding convictions (Matthews et al., 1991), Kontogiannis (2006) failed to find a direct relationship with either speeding convictions or collisions.
PART | II
Key Variables to Understand in Traffic Psychology
As an alternative, Matthews et al. (1997) revised and extended the item pool of the DBI to develop the Driver Stress Inventory (DSI), which consists of five factors or “dimensions” of stress vulnerability. According to Matthews (2002), Dislike of Driving (DIS), Driver Aggression (AG), and Fatigue Proneness (FAT) are reflective of subjective states of disturbance while driving. Thrill Seeking (TS) represents an enjoyment of danger, and Hazard Monitoring (HM) is consistent with a vigilance to danger while driving. The DSI is believed to be more predictive of negative driving outcomes consistent with ¨ z, O ¨ zkan, and stress than general personality constructs. O Lajunen (2010) confirmed the five-factor solution of the DSI among Turkish drivers, and regression analysis showed that DIS and TS predicted speeding, whereas AG, DIS, and low HM were related to accident involvement. Matthews (2001) also reported that AG, TS, and low DIS were related to speeding and self-reported violations, and that AG, TS, DIS, FAT, and low HM were linked to higher rates of unintentional errors. However, some inconsistencies have ¨ z et al. did find negative outcomes from DIS, been noted. O AG, and FAT that would be expected in comparison to the work of Matthews and colleagues, but they found that TS was related to less “dangerous” activity in the sense of slower driving and that HM was related to greater collisions. They argued that other factors, such as unique sample makeup, risk perception, and hazard detection, might account for these differences, which would be consistent with the transactional model on which the DSI is based. One issue with using the DBQ and DSI is the fact that the factors overlap with other driving personality constructs, most notably aggression and thrill seeking of sensation seeking scales. Although these legitimately appear to represent components of driver stress within a transactional model, caution needs to be exercised when examining trait driver stress in combination with tools more narrowly designed to measure related constructs to avoid inflated relationships due to item overlap. Another issue is that age differences have been found in both total and subfactor scores of driver stress, with older drivers typically showing lower levels of stress (Gulian, Matthews, et al., 1989; Langford & Glendon, 2002). Given that age has also been negatively linked to several driving outcomes, including aggression, speeding, riskiness, and collisions (Hennessy & Wiesenthal, 2002; Waylen & McKenna, 2008), oversampling of young participants or failure to control for age may exaggerate the negative impact of stress on driving. A similar case could be made for gender due to the fact that men are also highly represented in dangerous driving activities and have been concurrently found to show higher levels of some components of driver stress, such as driving aggression and irritation when overtaking (Matthews et al., 1999). Hence, the
Chapter | 12
Social, Personality, and Affective Constructs in Driving
disproportionate inclusion of male drivers in research could overstate the link between components of drivers stress, such as aggression, and negative outcomes, such as collisions.
2.6. Locus of Control Locus of control (LOC) represents an enduring belief about the source of cause for, or control over, personal behavior and is typically polarized as internal (I) and external (E) (Rotter, 1966). Externals are more likely to find responsibility for their actions in other individuals, luck, chance, or situational factors beyond their control, which may contribute to lack of caution or fewer precautions to prevent negative outcomes in life. According to Montag and Comrey (1987), externals represent a danger in the traffic environment due to their passive tendencies leading to fewer personal precautions. In contrast, internals tend to attribute their activities more often to stable, internal attributes (skill or effort) and as such also take more active responsibility for their actions and steps to alter negative future outcomes. However, Arthur and Doverspike (1992) argued that internals may represent a greater danger in the driving environment due to an overconfidence and overestimation of their skills or abilities while driving. Overall, LOC has been linked to various driving attitudes and behaviors, although there have been mixed findings. An external locus has been linked with elevated collisions (Lajunen & Summala, 1995), drinking and driving offenses (Cavaiola & DeSordi, 2000), errors (Breckenridge & Dodd, 1991), and intention to commit violations (Yagil, 2001), whereas an internal locus has been associated with increased seat belt use (Hoyt, 1973) and cautious behavior (Montag & Comrey, 1987). However, others have failed to find a link between LOC and collisions (Guastello & Guastello, 1986; Iversen & Rundmo, 2002) or seat belt use (Riccio-Howe, 1991) and determined that different measurement approaches can lead to contrasting results (Cavaiola & DeSordi, 2000). One reason for such inconsistency could be that Rotter’s original I/E scale may be too general to predict situationspecific LOC. In response, Montag and Comrey (1987) developed two separate scales to measure driving-specific internality (DI) and externality (DE) (MDIE), with which they identified a positive relationship between DE, as well as a negative relationship between DI, and collisions. However, using the MDIE, Arthur and Doverspike (1992) and Iversen and Rundmo (2002) failed to find any significant association with collisions. Holland, Geraghty, and Shah (2010) noted that such variation could be due to an overuse of male drivers, who are also more likely to possess an elevated self-bias and, as such, may give responses that make themselves appear less internally responsible for
155
collisions, hence masking possible relationships between DI and collisions. Another reason may be that the link between LOC and driving behavior is more indirect than direct. For example, although Guastello and Guastello (1986) did not find a link between Rotter’s I/E and collisions, they did note that a belief about the causes of accidents, or attributional style, may translate to specific activities such as collisions. Furthermore, Holland et al. (2010) found that externals are more likely to possess a negative driving style (e.g., either a “dissociative style” that includes distractibility or an “anxious style” that involves lack of confidence and distress) that may then indirectly impact collisions. Yagil (2001) similarly argued that externality can indirectly influence intentions to commit violations through positive attitudes toward violations. Others have noted that LOC acts as the moderator. According to Gidron, Gal, and Desevilya (2003), the impact of hostility on collisions is moderated by increased DI. In another study, Miller and Mulligan (2002) found that driving LOC altered the outcome of mortality salience on subsequent risky driving behavior. After receiving a mortality salience treatment in which participants were given 15 true/false items designed to stimulate thoughts about their own death, internals indicated decreased, whereas externals reported increased, future risky driving behavior. This suggests that the LOC orientation was accountable for changes in the perception and personalization of mortality-related information in the driving context, in advance of future risk-taking driving behavior. A third reason for the discrepancies could be that the concept of LOC is more varied and multidimensional than originally proposed. In this respect, a single bipolar distinction of control may be insufficient to examine the complexities of driving behavior. Levenson (1981) extended the original dimension of I/E to represent independent orientations of internal, chance, and powerful others. In a similar respect, Lajunen developed a multidi¨ zkan & mensional traffic-specific LOC scale (T-LOC; O Lajunen, 2005a), which asks questions about the source of control for driving outcomes. The four subscales of the TLOC include the Self (similar to Internal), the Vehicle/ Environment, Other Drivers, and Fate (similar to External). ¨ zkan and Lajunen (2005a) Using a Turkish translation, O found that those with an elevated “self” LOC reported elevated ordinary and aggressive violations, errors, and active accidents, which they interpreted within the framework of an elevated self-bias where self-confident drivers downplay the likelihood of negative outcomes for more dangerous driving behavior due to their own ability or skill. ¨ zkan, Using a Swedish translation of the T-LOC, Warner, O and Lajunen (2010) found a five-factor structure with similar external factors (Vehicle/Environment, Other Drivers, and Fate) and two internal factors related to the
156
Self (Own Skill and Own Behavior). Own Behavior predicted self-reported speed on 90 km/h roads, supporting the notion that internal aspects of the T-LOC may be a useful alternative in understanding riskier driving activities.
3. THE CONTEXT IN DRIVING OUTCOMES The context has a complex and ongoing impact on the thoughts, feelings, and actions of drivers. For the purpose of this chapter, the driving context includes a combination of physical (e.g., temperature), social (e.g., culture), and temporal (e.g., time urgency) factors on both distal/macro (laws and norms) and proximal/micro (weather and traffic congestion) levels. Part of the challenge in considering context factors in traffic research, particularly those on the proximal level, is the fact that they are fluid and part of an ongoing, and often changing, process. As mentioned previously, context factors are important to understand on their own, but they are best appreciated in interaction with personal factors. It is also important to note that context factors interact with one another, and that issues, events, and experiences from one context can carry over from one environment to impact subsequent environments, often outside of conscious awareness (Hennessy, 2008). Events from the traffic environment can influence drivers in subsequent nondriving settings (Hennessy & Jakubowski, 2007), and drivers can be changed by factors from outside the driving environment (Wickens & Wiesenthal, 2005). In this respect, combinations of contextual variables may be conceptually limitless, although the following represent a small subset of context factors and their recognized impact on driving.
3.1. Traffic Congestion: Impedance, Time Urgency, and “Other” Drivers Perhaps the most recognizable contextual factor in the driving environment is traffic congestion, which extends beyond the number of vehicles to represent the perception and interpretation of that volume. In line with the distinction between density and crowding in environmental psychology, the volume of traffic embodies the physical number of vehicles (or drivers) per unit of space, whereas congestion occurs when drivers believe that there are too many vehicles or too little space at that given time and location. Also, although a greater number of vehicles are likely to lead to perceptions of congestion, it is not necessary nor sufficient. For example, regular commuters, who have adjusted their expectations to fit the predictably slower moving areas during “rush hour,” may not necessarily view a specific commute as congested, especially if traffic flows faster than usual. In contrast, during periods that are typically low in volume, drivers who are forced to
PART | II
Key Variables to Understand in Traffic Psychology
slow their pace or drive closer to other vehicles than anticipated for some unexpected reason may experience a sense of congestion at that particular time and location. In this respect, negative outcomes are not dependent solely on how many vehicles are present on the roadways, which would explain why congestion, and related outcomes such as stress, can be experienced in areas that have relatively small driving populations rather than solely in large city centers. According to Stokols et al. (1978), one factor that may facilitate the perception of congestion is impedance, which represents the behavioral constraints that occur in traffic due to the distance a driver must travel and the time spent in transit during a given trip. They found that high impedance (longer distances at a slower pace) leads to greater perceptions of congestion, greater physiological arousal, and more negative evaluations of other drivers. Schaeffer, Street, Singer, and Baum (1988) subsequently argued that distance and time are so highly correlated that the average speed of a commute was a better predictor of impedance, which they also found to be related to physiological arousal and decreased performance on postcommute proofreading tasks (identifying errors in spelling, grammar, and punctuation in a written passage). Thus, travel impedance appears to create an aversive condition resulting from blocked goals (i.e., expectations of traveling at a certain speed over a certain distance in a specified time frame) due to constrained movements from other drivers (and their vehicles), which can lead to negative physical, cognitive, and behavioral outcomes. A major consequence of traffic congestion is heightened stress and anxiety. Several studies have found that drivers report greater stress on days when levels of congestion are highest (Gulian, Matthews, et al., 1989; Hennessy & Wiesenthal, 1999). Using predetermined routes that were naturally either high or low in congestion, Van Rooy (2006) found that participants randomly assigned to the highly congested route reported greater stress and negative mood. Similarly, Hennessy and Wiesenthal (1997) had commuters drive their regular daily routes and administered state measures of driver stress while they drove in areas of both high and low congestion during the same trip. They found that state stress was significantly greater in high-congestion than in low-congestion conditions, but it was moderated by trait driver stress susceptibility. In studies examining bus drivers who routinely experience a wide spectrum of traffic conditions, Evans and colleagues determined that higher congestion was associated with elevated stress but was exaggerated by personal factors of lost control (Evans & Carre`re, 1991) and learned helplessness (Evans & Stecker, 2004) in such conditions. Traffic congestion can also increase anger and aggression in drivers. Deffenbacher et al. (2003) found that anger predicted aggression in simulated driving situations in which participants were impeded by other traffic. Using
Chapter | 12
Social, Personality, and Affective Constructs in Driving
audio recorders, Underwood et al. (1999) had participants document various experiences from their commute and found that congested encounters were linked to elevated anger during that commute. Gulian, Glendon, Davies, Matthews, and Debney (1989) determined that a large proportion of highway drivers in the United Kingdom frequently experienced irritation in traffic congestion, and that the source of frustration and aggression they reported was typically “other drivers.” Hennessy and Wiesenthal (1999) confirmed this in Canadian drivers interviewed via cellular telephones while engaged in actual low- and highcongestion conditions, where state aggression toward other drivers occurred more frequently in high congestion. This is in line with the notion by Shinar (1998) that roadway aggression may be explained by the frustration aggression hypothesis, in which frustration is more typical when the actions of other drivers block or impede an intended goal. This frustration in turn increases the likelihood of aggression (depending on personal factors, past experience, attributions of others’ actions, and anticipation of outcomes). Several studies have supported the proposition that frustration and irritation from other drivers can lead to driver aggression (Bjo¨rklund, 2008; Hennessy & Wiesenthal, 2001b; Lajunen & Parker, 2001; McGarva, Ramsey, & Shear, 2006). Interestingly, traffic congestion has also been shown to have an enduring impact outside the actual driving environment. Spillover effects have been detected, where congestion-induced elevations in driver stress subsequently impact postdriving outcomes. For example, traffic congestion has been linked to performance decrements in workplace tasks, such as greater errors in proofreading (Schaeffer et al., 1988) and greater time needed to complete simple visual spatial assignments (Hennessy & Jakubowski, 2007), as well as disturbances in interpersonal factors, including increased levels of workplace aggression during that workday (specifically, obstructionism and expressed hostility among men; Hennessy, 2008) and negative evaluations of unqualified job candidates (Van Rooy, 2006). Hennessy (2008) proposed that unresolved stress from the traffic environment likely continues to unconsciously influence and intensify subsequent reactions to stressors experienced after a commute ends, even though the immediate emotional effects from driving dissipate quickly postcommute. An important consideration in interpreting the impact of congested conditions is the confounding issue of time urgency. For many, the delays caused by other drivers can alter the economic or social experience of time, particularly during “rush hour.” Time pressure or time urgency can modify the perception of traffic events, flow, and actions of other drivers, subsequently increasing stress, irritation, frustration, negative affect, and aggression (Evans & Carre`re, 1991; Hennessy, Wiesenthal, & Kohn, 2000;
157
Koslowsky, 1997; Lucas & Heady, 2002; Neighbors et al., 2002). Everyday hassles, such as time pressure and adverse driving conditions, typically have an additive effect, in which the influence of one event can add to the severity of another. According to Shinar and Compton (2004), driver aggression increases in frequency during periods of high time value (rush hour) even after controlling for traffic volume. In addition, Lucas and Heady (2002) discovered that regular commuters with greater flextime in their work schedule showed less perceived time urgency and resulting driver stress. Clearly, a hurried pace and lifestyle, a focus on time issues, and the escalated potential for blocked goals have the capacity to exaggerate the impact of congestion on negative driving outcomes. Interpretations of research on traffic congestion, however, must also be qualified by methodological considerations. Like many contextual factors, it is difficult to accurately measure effects of traffic congestion outside of the actual driving experience. The use of hypothetical situations or participant recall of previous trips (even those recently completed) may be problematic because of memory issues, social desirability, experimenter bias, and expectancy effects. However, there are still concerns of control and predictability of real-time traffic events that can alter in situ evaluations of traffic congestion, such as police presence, weather effects, or construction. Also, simulated driving events are limited by the degree they are considered or experienced as “real” by participants, particularly in light of the perceptual nature of congestion in comparison to volume. Nonetheless, there is a degree of consistency across various methods that support the relevance of traffic congestion in negative driving outcomes.
3.2. Physical Environment Previous research has established a greater risk of collisions due to adverse weather, including rain, snow, and fog (Andrey, Mills, Leahy, & Suggett, 2003), but many drivers attempt to compensate for such conditions by driving more cautiously, such as reducing speed and increasing distance between vehicles (de Waard, Kruizinga, & Brookhuis, 2008; Harris & Houston, 2010). This is in line with the zero risk theory (Na¨a¨ta¨nen & Summala, 1976), which holds that drivers adapt to risk (i.e., driving more slowly) to the point that their subjective risk approaches zero. However, Kilpela¨inen and Summala (2007) argued that many are not accurate in predicting risks during poor weather and subsequently do not adjust their driving behavior properly, particularly on slippery roads. One possible problem is that although individual motivation might be geared toward increasing personal safety, perception may be impacted by the conditions (e.g., inaccurate interpretations of distances, speed of traffic, and severity of conditions), leading to unintended risky
158
behavior, such as speeds that are too fast for the conditions and reduced headway between vehicles (Caro, Cavallo, Marendaz, Boer, & Vienne, 2009). Cognitive workload may also be impacted during adverse weather conditions because drivers must focus on slower moving traffic and deal with reduced visibility, thus decreasing their capacity to estimate safety margins and increasing strain (Cavallo, Mestre, & Berthelon, 1997). Hill and Boyle (2007) also found that adverse weather and limited visibility can lead to elevated stress levels, which have been independently connected to increased riskiness and collisions. Heat is another environmental factor that can impact drivers. Heat may be a source of psychological stress and physiological arousal that alters cognitions, emotions, and behavior. Using a traffic simulator, Wyon, Wyon, and Norin (1996) discovered that heat impacted vigilance, where drivers in hot conditions missed a greater number of signals to press a pedal. Heat has also been linked to increased aggressive behavior in drivers. Kenrick and MacFarlane (1986) had a confederate vehicle cause delays at a green light and found increased horn honking among those with open driver-side windows (implying the lack of airconditioning and “hot” drivers) during periods of higher temperatures. However, it should be noted that the impact of high temperatures may be qualified by confounding factors of humidity, wind, and air pressure, which can alter the experience of “heat.” Similarly, duration of temperature needs to be considered given that prolonged heat will have different outcomes than those resulting from temporary or fluctuating heat. This is highlighted in research on the heateaggression hypothesis that suggests a curvilinear relationship, where escape may be a more preferable behavior selection than aggression at prolonged extreme temperatures (Bell & Fusco, 1989). It is also important to note that negative outcomes from heat are typically instigated through some social contact or interaction. The target of heat-related aggression is typically others perceived as a source of irritation, frustration, or wrongdoing. The physical environment may also have beneficial effects on driving outcomes in that pleasant sceneries may favorably impact cognitions, attitudes, affect, and behavior. Although evaluations and perceptions of scenery are subjective, consistent patterns have been found, including a greater general preference for natural as opposed to built urban scenery (Kaplan & Kaplan, 1982). This has been confirmed in the traffic environment, in which drivers report that undeveloped landscapes are less cluttered, as well as more pleasant, useful, and attractive (Evans & Wood, 1980). Antonson, Ma˚rdh, Wiklund, and Blomqvist (2009) reviewed existing literature related to the effect of roadside scenery on driving behavior and noted that varied and vegetated landscapes can have a beneficial impact. Such landscapes may increase curiosity, decrease boredom, and
PART | II
Key Variables to Understand in Traffic Psychology
help maintain focus through anticipation of upcoming threats or challenges. Alternatively, common, predictable, or homogeneous landscapes may lead to monotony, fatigue, and risk. They found that those who traveled a vegetated route in a simulator drove slower and closer to the center of the road than those who traveled an open landscape. However, they also found greater stress, which may be explained by a greater need for vigilance due to a more rapidly changing scenery on an unfamiliar roadway. Alternatively, Parsons, Tassinary, Ulrich, Hebl, and Grossman-Alexander (1998) measured autonomic stress in participants who experienced mild stressors before and after viewing driving scenarios that varied in degree of roadside vegetation. They found that viewing the more vegetated environments did lead to decreased stress responses and increased recovery from that stress. Similarly, Cackowski and Nasar (2003) found that those who viewed routes with more roadside vegetation showed greater subsequent frustration tolerance by spending greater time attempting to solve an insolvable puzzle task. These findings highlight the relevance of the physical environment in understanding the personesituation interaction of drivers.
3.3. Driving Culture and Norms Culture represents shared norms, values, traditions, and customs of a group that typically define and guide appropriate and inappropriate attitudes and behaviors. These can occur on a macro level (e.g., national customs and religious holidays) or a more micro level (e.g., family traditions and peer activities). Driving behavior and driving style should be influenced by these cultural processes, given that the driving environment is a social context with very distinct rules and norms that are transmitted between road users across time and generations. Also, although driving laws, licensing procedures, road types, driving styles, and actual driving behaviors will vary regionally and internationally, dangerous and risky driving practices occur universally. Attitudes about driving and personal driving styles are largely learned, which includes influence from parents, peers, media, and other drivers regarding the overall riskiness of driving, as well as the probability of experiencing negative outcomes (Hennessy, Hemingway, & Howard, 2007; Shope & Bingham, 2008). Using Ajzen’s (1985) theory of planned behavior as a framework, Elliott, Armitage, and Baughan (2007) found that subjective norms (the perceived pressure or acceptance of others toward a behavior) were associated with elevated speeding intentions, which subsequently predicted both self-reported and observed speeding behavior in a simulator. In essence, the normative belief that there is a consensus or commonality to unsafe driving behavior may serve as a justification for its personal adoption (Forward, 2009).
Chapter | 12
Social, Personality, and Affective Constructs in Driving
Parents represent one potential source of cultural and normative influence on driving behavior, particularly for young drivers. Parental influence on driving begins early in life, well before formal “training years,” through modeling of driving styles, attitudes about safety, reactions to other drivers, and respect for traffic laws (Summala, 1987). Familial models that validate recklessness often encourage riskiness of young drivers (Taubman-Ben-Ari, 2008). Bianchi and Summala (2004) further proposed that parental influence may impart from genetic dispositions that guide personal tendencies of parent drivers, such as sensation seeking or attention, which are then demonstrated to their children. Research has consistently shown that parental attitudes and activities outside the driving environment, particularly lenience of restrictions and control, can impact driving behavior and style of young drivers (Hartos, Eitel, Haynie, & Simons-Morton, 2000). Shope, Waller, Raghunathan, and Patil (2001) found that parental monitoring, nurturing, and connectedness in 10th grade of high school were subsequently linked to lower rates of serious offenses (alcohol related, speeding, and reckless driving) and crashes (single vehicle, at fault, and alcohol related), whereas lower monitoring, nurturance, connectedness, and a greater lenience toward drinking had the opposite effect. Similarly, Prato, Toledo, Lotan, and Taubman-Ben-Ari (2010) determined that risk indexes for young drivers were lower for those whose parents were actively involved in monitoring their child’s driving behavior, whereas lack of supervision exaggerated existing dangerous driving tendencies in their child, such as sensation seeking, increasing overall risk. Another important source of cultural and normative driving influence is the media, which is prone to promote danger and risk over safety as a “normal” part of driving. Although the media has the potential to help promote a culture of driving safety, in many cultures, television and movies glorify and promote speeding, risk taking, and dangerous driving practices as acceptable or even admirable (Hennessy et al., 2007), particularly for young male drivers. Consistent with social and cognitive learning theories, one primary mechanism by which behavior is acquired is through observing and imitating others. By placing emphasis on the consequences of others’ actions, observation serves as a vicarious learning experience (Bandura, 2001). Hennessy et al. found that speeding, lane violations, and near collisions were elevated in a simulator among drivers previously exposed to a short movie scene of dangerous driving. It is possible that watching media portrayals that endorse driving that is competitive, performed at excessive speeds, or otherwise unsafe increases viewers’ perceived acceptability of such actions and reduces the expectancy of a tragic outcome, which then increases the likelihood they will engage in dangerous driving themselves.
159
4. CONCLUSIONS AND FUTURE DIRECTIONS This chapter highlighted the importance of personal and social influences on driving outcomes. Each driver brings a unique and rich experience base, skill set, expectation, interpretation, and reaction to the distinctive state driving context, yet research has been successful in identifying consistent patterns that can heighten the likelihood of negative driving outcomes, only a fraction of which were presented here. The purpose of such research is to concurrently understand problematic aspects within the person and the context to ultimately improve traffic safety. However, there are still many more factors in this process to be discovered. One future focus could be on understanding aspects that influence more positive or safe driving emotions and actions. For example, Peˆcher et al. (2009) found that listening to happy music decreased mean speed and hard shoulder deviations, suggesting that positive emotions could ultimately distract drivers from dangerous activi¨ zkan and Lajunen (2005b) argued ties. Furthermore, O that promotion of polite driving through media sources could be a means of reducing aggressive driving. Likewise, psychology often emphasizes negative outcomes, sometimes at the expense of characteristics that can improve or increase more appropriate consequences. Anger, aggression, and revenge have been examined more frequently than their constructive counterpart, forgiveness. It is possible that understanding the process by which drivers forgive transgressions of others in the traffic context (rather than become agitated, ruminate, and harm others) may ultimately hold greater long-term benefit in helping to improve driver interactions (Moore & Dahlen, 2008). Another direction could be increased expansion of the personesituation interaction to include the vehicle. Whereas safety issues surrounding the vehicle have been the focus of engineering and human factors, traffic psychology as a whole has been less inclined to incorporate such aspects into research. The vehicle is more than just the means of transporting individuals; rather, it is part of a driver’s life space. Issues such as comfort, layout of instruments, safety features, overall condition of the vehicle, or technology can alter the driving experience variably across situations. In many instances, the vehicle becomes an extension of the driver and his or her space, with personal meaning that can lead to protection and defense (Marsh & Collett, 1987). It can also provide a degree of anonymity from others that allows expression of emotions and actions not typical in other situations (Li et al., 2004). In this respect, a full understanding of the transactional aspects of driving may not be possible without inclusion of the vehicle.
160
Finally, there is a need for continued globalization of traffic research. Although current research has begun to identify those aspects that are unique and those that are consistent across cultures, this is only a start. This is particularly pressing in light of the recent emphasis on the role of traffic psychology in helping create a driving safety culture, with a concentration on establishing values, attitudes, and behaviors that permeate across a society that favors safe over unsafe driving practices as the expectation and norm for all. For example, as laws have changed and seat belt use has become an expectation, seat belt use has increased dramatically during the past few decades in many countries (European Traffic Safety Council, 2006). In such instances, successive generations come to view seat belt use as normative, and non-users are explicitly and implicitly pressured to comply more frequently. As has been highlighted in this chapter, within the personesituation interaction, an enduring alteration in cultural or normative expectations for safety should have widespread national and international impact on individual driving outcomes; hence, pervasive changes hold great potential for improved global driver safety.
REFERENCES Ajzen, I. (1985). From intentions to actions: A theory of planned behavior. In J. Kuhl, & J. Beckman (Eds.), Action control: From cognition to behavior (pp. 11e39). Heidelberg, Germany: Springer. Andrew, M., & Cronin, C. (1997). Two measures of sensation seeking as predictors of alcohol use among high school males. Personality and Individual Differences, 22, 393e401. Andrey, J., Mills, B., Leahy, M., & Suggett, J. (2003). Weather as a chronic hazard for road transportation in Canadian cities. Natural Hazards, 28, 319e343. Antonson, H., Ma˚rdh, S., Wiklund, M., & Blomqvist, G. (2009). Effect of surrounding landscape on driving behavior: A driving simulator study. Journal of Environmental Psychology, 29, 493e502. Arnett, J. (1994). Sensation seeking: A new conceptualisation and a new scale. Personality and Individual Differences, 16, 289e296. Arnett, J., & Balle-Jensen, L. (1993). Cultural bases of risk behavior: Danish adolescents. Development, 64, 1842e1855. Arthur, W., & Doverspike, D. (1992). Locus of control and auditory selective attention as predictors of driving accident involvement: A comparative longitudinal investigation. Journal of Safety Research, 23, 73e80. Bandura, A. (2001). Social cognitive theory: An agentic perspective. Annual Review of Psychology, 52, 1e26. Beirness, D. J. (1993). Do we really drive as we live? The role of personality factors in road crashes. Alcohol, Drugs and Driving, 9, 129e143. Bell, P. A., & Fusco, M. E. (1989). Heat and violence in the Dallas field data: Linearity, curvilinearity, and heteroscedasticity. Journal of Applied Social Psychology, 19, 1479e1482. ¨ zkan, T., & Lajunen, T. (2008). The Driver Behaviour Bener, A., O Questionnaire in Arab Gulf countries: Qatar and United Arab Emirates. Accident Analysis and Prevention, 40, 1411e1417.
PART | II
Key Variables to Understand in Traffic Psychology
Berkowitz, L. (1993). Aggression: Its causes, consequences, and control. Boston: McGraw-Hill. Bettencourt, B. A., Talley, A., Benjamin, A. J., & Valentine, J. (2006). Personality and aggressive behavior under provoking and neutral conditions: A meta-analytic review. Psychological Bulletin, 132, 751e777. Bianchi, A., & Summala, H. (2004). The "genetics" of driving behavior: Parents’ driving style predicts their children’s driving style. Accident Analysis and Prevention, 36, 655e659. Bjo¨rklund, G. M. (2008). Driver irritation and aggressive behaviour. Accident Analysis and Prevention, 40, 1069e1077. Breckenridge, R. L., & Dodd, M. O. (1991). Locus of control and alcohol placebo-effects on performance in a driving simulator. Perceptual and Motor Skills, 72, 751e756. Britt, T. W., & Garrity, M. J. (2006). Attributions and personality as predictors of the road rage response. British Journal of Social Psychology, 45, 127e147. Bulmash, E. L., Moller, H. J., Kayumov, L., Shen, J. H., Wang, X. H., & Shapiro, C. M. (2006). Psychomotor disturbance in depression: Assessment using a driving simulator paradigm. Journal of Affective Disorders, 93, 213e218. Burns, P. C., & Wilde, G. J. S. (1995). Risk-taking in male taxi drivers: Relationships among personality, observational data and driver records. Personality and Individual Differences, 18, 267e278. Cackowski, J. M., & Nasar, J. L. (2003). The restorative effects of roadside vegetation: Implications for automobile driver anger and frustration. Environment and Behavior, 35, 736e751. Caro, S., Cavallo, V., Marendaz, C., Boer, E. R., & Vienne, F. (2009). Can headway reduction in fog be explained by impaired perception of relative motion? Human Factors, 51, 378e392. Cavaiola, A. A., & DeSordi, E. G. (2000). Locus of control in drinking driving offenders and non-offenders. Alcoholism Treatment Quarterly, 18, 63e73. Cavallo, V., Mestre, D., & Berthelon, C. (1997). Visual and spatiotemporal factors. In J. A. Rothengatter, & E. J. Carbonell Vaya (Eds.), Time-to-collision judgments: Traffic and transport psychology: Theory and applications (pp. 97e111). Amsterdam: Pergamon. Dahlen, E. R., Martin, R. C., Ragan, K., & Kuhlman, M. M. (2005). Driving anger, sensation seeking, impulsiveness, and boredom proneness in the prediction of unsafe driving. Accident Analysis and Prevention, 37, 341e348. Dahlen, E. R., & White, R. P. (2006). The Big Five factors, sensation seeking, and driving anger in the prediction of unsafe driving. Personality and Individual Differences, 41, 903e915. de Waard, D., Kruizinga, A., & Brookhuis, K. A. (2008). The consequences of an increase in heavy goods vehicles for passenger car drivers’ mental workload and behaviour: A simulator study. Accident Analysis and Prevention, 40, 818e828. Deffenbacher, J. L., Deffenbacher, D. M., Lynch, R. S., & Richards, T. L. (2003). Anger, aggression, and risky behavior: A comparison of high and low anger drivers. Behaviour Research and Therapy, 41, 701e718. Deffenbacher, J. L., Huff, M. E., Lynch, R. S., Oetting, E. R., & Salvatore, N. F. (2000). Characteristics and treatment of high-anger drivers. Journal of Counseling Psychology, 47, 5e17. Deffenbacher, J. L., Lynch, R. S., Oetting, E. R., & Yingling, D. A. (2001). Driving anger: Correlates and a test of state-trait theory. Personality and Individual Differences, 31, 1321e1331.
Chapter | 12
Social, Personality, and Affective Constructs in Driving
Deffenbacher, J. L., Oetting, E. R., & Lynch, R. S. (1994). Development of a driving anger scale. Psychological Reports, 74, 83e91. Desmond, P. A., & Matthews, G. (2009). Individual differences in stress and fatigue in two field studies of driving. Transportation Research Part F: Traffic Psychology and Behaviour, 12, 265e276. Doob, A. N., & Gross, A. E. (1968). Status of frustrator as an inhibitor of horn honking responses. Journal of Social Psychology, 76, 213e218. Dorn, L., & Matthews, G. (1995). Prediction of mood and risk appraisals from trait measures: Two studies of simulated driving. European Journal of Personality, 9, 25e42. Elliott, M. A., Armitage, C. J., & Baughan, C. J. (2007). Using the theory of planned behaviour to predict observed driving behaviour. British Journal of Social Psychology, 46, 69e90. European Transport Safety Council. (2006). Seat belt reminders increasingly standard in Europe but not in all countries. http://www.etsc.eu/ documents/Press%20release%20PIN%20Flash%203.pdf. Accessed July 5, 2009. Evans, G. W., & Carre`re, S. (1991). Traffic congestion, perceived control, and psychophysiological stress among urban bus drivers. Journal of Applied Psychology, 76, 658e663. Evans, G. W., & Stecker, R. (2004). Motivational consequences of environmental stress. Journal of Environmental Psychology, 24, 143e165. Evans, G. W., & Wood, K. W. (1980). Assessment of environmental aesthetics in scenic highway corridors. Environment and Behavior, 12, 255e273. Forward, S. E. (2009). The theory of planned behaviour: The role of descriptive norms and past behaviour in the prediction of drivers’ intentions to violate. Transportation Research Part F: Traffic Psychology and Behaviour, 12, 198e207. Gidron, Y., Gal, R., & Desevilya., H. S. (2003). Internal locus of control moderates the effects of road hostility on recalled driving behavior. Transportation Research Part F: Traffic Psychology and Behaviour, 6, 109e116. Glendon, A. I., Dorn, L., Matthews, G., Gulian, E., Davies, D. R., & Debney, L. M. (1993). Reliability of the Driving Behavior Inventory. Ergonomics, 36, 719e726. Greene, K., Krcmar, M., Walters, L. H., Rubin, D. L., & Hale, J. L. (2000). Targeting adolescent risk-taking behaviors: The contributions of egocentrism and sensation seeking. Journal of Adolescence, 23, 439e461. Guastello, S. J., & Guastello, D. D. (1986). The relation between the locus of control construct and involvement in traffic accidents. Journal of Psychology, 120, 293e297. Gulian, E., Glendon, A. I., Davies, D. R., Matthews, G., & Debney, L. M. (1989). Coping with driver stress. In F. McGuigan, W. E. Sime, & J. M. Wallace (Eds.), Stress and tension control, Vol. 3 (pp. 173e186). New York: Plenum. Gulian, E., Matthews, G., Glendon, A. I., & Davies, D. R. (1989). Dimensions of driver stress. Ergonomics, 32, 585e602. Gulliver, P., & Begg, D. (2007). Personality factors as predictors of persistent risky driving behavior and crash involvement among young adults. Injury Prevention, 13, 376e381. Harris, P. B., & Houston, J. M. (2010). Recklessness in context: Individual and situational correlates to aggressive driving. Environment and Behavior, 42, 44e60. Hartos, J. L., Eitel, P., Haynie, D. L., & Simons-Morton, B. G. (2000). Can I take the car? Relations among parenting practices and
161
adolescent problem-driving practices. Journal of Adolescent Research, 15, 352e367. Hennessy, D. A. (2008). The impact of commuter stress on workplace aggression. Journal of Applied Social Psychology, 38, 2315e2335. Hennessy, D. A., Hemingway, J., & Howard, S. (2007). The effects of media portrayals of dangerous driving on young drivers’ performance. In F. N. Gustavvson (Ed.), New transportation research progress (pp. 143e156). New York: Nova Science. Hennessy, D. A., & Jakubowski, R. (2007). The impact of visual perspective and anger on the actoreobserver bias among automobile drivers. Traffic Injury Prevention, 8, 115e122. Hennessy, D. A., & Wiesenthal, D. L. (1997). The relationship between traffic congestion, driver stress, and direct versus indirect coping behaviour. Ergonomics, 40, 348e361. Hennessy, D. A., & Wiesenthal, D. L. (1999). Traffic congestion, driver stress, and driver aggression. Aggressive Behavior, 25, 409e423. Hennessy, D. A., & Wiesenthal, D. L. (2001a). Further validation of the driving vengeance questionnaire. Violence and Victims, 16, 565e573. Hennessy, D. A., & Wiesenthal, D. L. (2001b). Gender, driver aggression, and driver violence: An applied evaluation. Sex Roles, 44, 661e676. Hennessy, D. A., & Wiesenthal, D. L. (2002). The relationship between driver aggression, vengeance, and violence. Violence and Victims, 17, 707e718. Hennessy, D. A., & Wiesenthal, D. L. (2005). Driving vengeance and willful violations: Clustering of problem driving attitudes. Journal of Applied Social Psychology, 35, 61e79. Hennessy, D. A., Wiesenthal, D. L., & Kohn, P. M. (2000). The influence of traffic congestion, daily hassles, and trait stress susceptibility on state driver stress: An interactive perspective. Journal of Applied Biobehavioral Research, 5, 162e179. Hill, J. D., & Boyle, L. N. (2007). Driver stress as influenced by driving maneuvers and roadway conditions. Transportation Research Part F: Traffic Psychology and Behaviour, 10, 177e186. Holland, C., Geraghty, J., & Shah, K. (2010). Differential moderating effect of locus of control on effect of driving experience in young male and female drivers. Personality and Individual Differences, 48, 821e826. Hoyt, M. F. (1973). Internaleexternal control and beliefs about automobile travel. Journal of Research in Personality, 7, 288e293. Iversen, H., & Rundmo, T. (2002). Personality, risky driving and accident involvement among Norwegian drivers. Personality and Individual Differences, 33, 1251e1263. Jonah, B. A. (1997). Sensation seeking and risky driving: A review and synthesis of the literature. Accident Analysis and Prevention, 29, 651e665. Jonah, B. A., Thiessen, R., & Au-Yeung, E. (2001). Sensation seeking, risky driving and behavioral adaptation. Accident Analysis and Prevention, 33, 679e684. Kaplan, S., & Kaplan, R. (1982). Cognition and environment: Functioning in an uncertain world. New York: Praeger. Kenrick, D. T., & MacFarlane, S. W. (1986). Ambient temperature and horn honking: A field study of the heat aggression relationship. Environment and Behavior, 18, 179e191. Kilpela¨inen, M., & Summala, H. (2007). Effects of weather and weather forecasts on driver behaviour. Transportation Research Part F: Traffic Psychology and Behaviour, 10, 288e299. King, Y., & Parker, D. (2008). Driving violations, aggression and perceived consensus. European Review of Applied Psychology, 58, 43e49.
162
Kontogiannis, T. (2006). Patterns of driver stress and coping strategies in a Greek sample and their relationship to aberrant behaviors and traffic accidents. Accident Analysis and Prevention, 38, 913e924. Kontogiannis, T., Kossiavelou, Z., & Marmaras, N. (2002). Self-reports of aberrant behaviour on the roads: Errors and violations in a sample of Greek drivers. Accident Analysis and Prevention, 34, 381e399. Koslowsky, M. (1997). Commuting stress: Problems of definition and variable identification. Applied Psychology: An International Review, 46, 153e173. Lajunen, T., Corry, A., Summala, H., & Hartley, L. (1998). Cross-cultural differences in drivers’ self-assessments of their perceptual-motor and safety skills: Australians and Finns. Personality and Individual Differences, 24, 539e550. Lajunen, T., & Parker, D. (2001). Are aggressive people aggressive drivers? A study of the relationship between self-reported general aggressiveness, driver anger and aggressive driving. Accident Analysis and Prevention, 33, 243e255. Lajunen, T., Parker, D., & Summala, H. (2004). The Manchester Driver Behaviour Questionnaire: A cross-cultural study. Accident Analysis and Prevention, 36, 231e238. Lajunen, T., & Summala, H. (1995). Driving experience, personality, and skill and safety motive dimensions in drivers self-assessments. Personality and Individual Differences, 19, 307e318. Langford, C., & Glendon, A. I. (2002). Effects of neuroticism, extraversion, circadian type and age on reported driver stress. Work and Stress, 16, 316e334. Levenson, H. (1981). Differentiating among internality, powerful others, and chance. In H. Lefcourt (Ed.), Research with the locus of control construct, Vol. 1 (pp. 15e63). New York: Academic Press. Lawton, R., Parker, D., Manstead, A. S. R., & Stradling, S. G. (1997). The role of affect in predicting social behaviors: The case of road traffic violations. Journal of Applied Social Psychology, 27, 1258e1276. Li, F., Li, C., Long, Y., Zhan, C., & Hennessy, D. A. (2004). Reliability and validity of aggressive driving measures in China. Traffic Injury Prevention, 5, 349e355. Lucas, J. L., & Heady, R. B. (2002). Flextime commuters and their driver stress, feelings of time urgency, and commute satisfaction. Journal of Business and Psychology, 16, 565e571. Marsh, P., & Collett, P. (1987). The car as a weapon. Et Cetera, 44, 146e151. Matthews, G. (2001). A transactional model of driver stress. In P. Hancock, & P. Desmond (Eds.), Stress, workload, and fatigue (pp. 133e166). Mahwah, NJ: Erlbaum. Matthews, G. (2002). Towards a transactional ergonomics for driver stress and fatigue. Theoretical Issues in Ergonomics Science, 3, 195e211. Matthews, G., Desmond, P. A., Joyner, L. A., Carcardy, B., & Gilliland, K. (1997). A comprehensive questionnaire measure of driver stress and affect. In J. A. Rothengatter, & E. J. Carbonell Vay‘a (Eds.), Traffic and transport psychology: Theory and applications (pp. 317e324). Amsterdam: Pergamon. Matthews, G., Dorn, L., & Glendon, A. I. (1991). Personality-correlates of driver stress. Personality and Individual Differences, 12, 535e549. Matthews, G., Dorn, L., Hoyes, T. W., Davies, D. R., Glendon, A. I., & Taylor, R. G. (1998). Driver stress and performance on a driving simulator. Human Factors, 40, 136e149. Matthews, G., Sparkes, T. J., & Bygrave, H. M. (1996). Attentional overload, stress, and simulated driving performance. Human Performance, 9, 77e101.
PART | II
Key Variables to Understand in Traffic Psychology
Matthews, G., Tsuda, A., Xin, G., & Ozeki, Y. (1999). Individual differences in driver stress vulnerability in a Japanese sample. Ergonomics, 42, 401e415. Maxwell, J. P., Grant, S., & Lipkin, S. (2005). Further validation of the Propensity for Angry Driving Scale in British drivers. Personality and Individual Differences, 38, 213e224. McGarva, A. R., Ramsey, M., & Shear, S. A. (2006). Effects of driver cell-phone use on driver aggression. Journal of Social Psychology, 146, 133e146. Mesken, J., Hagenzieker, M. P., Rothengatter, T., & de Waard, D. (2007). Frequency, determinants, and consequences of different drivers’ emotions: An on-the-road study using self reports, (observed) behaviour, and physiology. Transportation Research Part F: Traffic Psychology and Behaviour, 10, 458e475. Millar, M. (2007). The influence of public self-consciousness and anger on aggressive driving. Personality and Individual Differences, 43, 2116e2126. Miller, R. L., & Mulligan, R. D. (2002). Terror management: The effects of mortality salience and locus of control on risk-taking behaviors. Personality and Individual Differences, 33, 1203e1214. Montag, I., & Comrey, A. L. (1987). Internality and externality as correlates of involvement in fatal driving accidents. Journal of Applied Psychology, 72, 339e343. Moore, M., & Dahlen, E. R. (2008). Forgiveness and consideration of future consequences in aggressive driving. Accident Analysis and Prevention, 40, 1661e1666. Na¨a¨ta¨nen, R., & Summala, H. (1976). Road user behavior and traffic accidents. Amsterdam: North Holland. Neighbors, C., Vietor, N. A., & Knee, C. R. (2002). A motivational model of driving anger and aggression. Personality and Social Psychology Bulletin, 28, 324e335. Oltedal, S., & Rundmo, T. (2006). The effects of personality and gender on risky driving behaviour and accident involvement. Safety Science, 44, 621e628. ¨ zkan, T., & Lajunen, T. (2010). Professional and non-profes¨ z, B., O O sional drivers’ stress reactions and risky driving. Transportation Research Part F: Traffic Psychology and Behaviour, 13, 32e40. ¨ zkan, T., & Lajunen, T. (2005a). Multidimensional Traffic Locus of O Control Scale (T-LOC): Factor structure and relationship to risky driving. Personality and Individual Differences, 38, 533e545. ¨ zkan, T., & Lajunen, T. (2005b). A new addition to DBQ: Positive O Driver Behaviours Scale. Transportation Research Part F: Traffic Psychology and Behaviour, 8, 355e368. Parker, D., Lajunen, T., & Summala, H. (2002). Anger and aggression among drivers in three European countries. Accident Analysis and Prevention, 34, 229e235. Parsons, R., Tassinary, L. G., Ulrich, R. S., Hebl, M. R., & GrossmanAlexander, M. (1998). The view from the road: Implications for stress recovery and immunization. Journal of Environmental Psychology, 18, 113e140. Peˆcher, C., Lemercier, C., & Cellier, J. M. (2009). Emotions drive attention: Effects on driver’s behaviour. Safety Science, 47, 1254e1259. Prato, C. G., Toledo, T., Lotan, T., & Taubman-Ben-Ari, O. (2010). Modeling the behavior of novice young drivers during the first year after licensure. Accident Analysis and Prevention, 42, 480e486. Riccio-Howe, L. A. (1991). Health values, locus of control, and cues to action as predictors of adolescent safety belt use. Journal of Adolescent Health, 12, 256e262.
Chapter | 12
Social, Personality, and Affective Constructs in Driving
Richer, I., & Bergeron, J. (2009). Driving under the influence of cannabis: Links with dangerous driving, psychological predictors, and accident involvement. Accident Analysis and Prevention, 41, 299e307. ˚ berg, L. (1999). On the distinction between violations Rimmo¨, A., & A and errors: Sensation seeking associations. Transportation Research Part F: Traffic Psychology and Behaviour, 2, 151e166. Rotter, J. B. (1966). Generalized expectancies for internal versus external control of reinforcement. Psychological Monographs, 80, 1e28. Rotton, J., Gregory, P. J., & Van Rooy, D. L. (2005). Behind the wheel: Construct validity of aggressive driving scales. In D. A. Hennessy, & D. L. Wiesenthal (Eds.), Contemporary issues in road user behavior and traffic safety (pp. 37e46). New York: Nova Science. Schaeffer, M. H., Street, S. W., Singer, J. E., & Baum, A. (1988). Effects of control on the stress reactions of commuters. Journal of Applied Social Psychology, 18, 944e957. Schwerdtfeger, A., Heims, R., & Heer, J. (2010). Digit ratio (2D:4D) is associated with traffic violations for male frequent car drivers. Accident Analysis and Prevention, 42, 269e274. Shahar, A. (2009). Self-reported driving behaviors as a function of trait anxiety. Accident Analysis and Prevention, 41, 241e245. Shinar, D. (1998). Aggressive driving: The contribution of the drivers and the situation. Transportation Research Part F: Traffic Psychology and Behaviour, 1, 137e159. Shinar, D., & Compton, R. (2004). Aggressive driving: An observational study of driver, vehicle, and situational variables. Accident Analysis and Prevention, 36, 429e437. Shope, J. T., & Bingham, C. R. (2008). Teen driving: Motor-vehicle crashes and factors that contribute. American Journal of Preventive Medicine, 35, S261eS271. Shope, J. T., Waller, P. F., Raghunathan, T. E., & Patil, S. M. (2001). Adolescent antecedents of high-risk driving behavior into young adulthood: Substance use and parental influences. Accident Analysis and Prevention, 33, 649e658. Speilberger, C. D. (1988). State-Trait Anger Expression Inventory. Odessa, FL: Psychological Assessment Resources. Stephens, A. N., & Groeger, J. A. (2009). Situational specificity of trait influences on drivers’ evaluations and driving behaviour. Transportation Research Part F: Traffic Psychology and Behaviour, 12, 29e39. Stokols, D., Novaco, R. W., Stokols, J., & Campbell, J. (1978). Traffic congestion, type-A behavior, and stress. Journal of Applied Psychology, 63, 467e480. Sullman, M. J. M. (2006). Anger amongst New Zealand drivers. Transportation Research Part F: Traffic Psychology and Behaviour, 9, 173e184. Su¨mer, N. (2003). Personality and behavioral predictors of traffic accidents: Testing a contextual mediated model. Accident Analysis and Prevention, 35, 949e964.
163
Summala, H. (1987). Young driver accidents: Risk taking or failure of skills? Alcohol, Drugs and Driving, 3, 79e91. Taubman-Ben-Ari, O. (2008). Motivational sources of driving and their associations with reckless driving cognitions and behavior. European Review of Applied Psychology, 58, 51e64. Underwood, G., Chapman, P., Wright, S., & Crundall, D. (1999). Anger while driving. Transportation Research Part F: Traffic Psychology and Behaviour, 2, 55e68. Van Rooy, D. L. (2006). Effects of automobile commute characteristics on affect and job candidate evaluations: A field experiment. Environment and Behavior, 38, 626e655. Van Rooy, D. L., Rotton, J., & Burns, T. M. (2006). Convergent, discriminant, and predictive validity of aggressive driving inventories: They drive as they live. Aggressive Behavior, 32, 89e98. ¨ zkan, T., & Lajunen, T. (2010). Can the traffic locus of Warner, H. W., O control (T-LOC) scale be successfully used to predict Swedish drivers’ speeding behaviour? Accident Analysis & Prevention, 42, 1113e1117. Waylen, A. E., & McKenna, F. P. (2008). Risky attitudes towards road use in pre-drivers. Accident Analysis and Prevention, 40, 905e911. Wickens, C. M., & Wiesenthal, D. L. (2005). State driver stress as a function of occupational stress, traffic congestion, and trait stress susceptibility. Journal of Applied Biobehavioral Research, 10, 83e97. Wiesenthal, D. L., Hennessy, D. A., & Gibson, P. M. (2000). The Driving Vengeance Questionnaire (DVQ): The development of a scale to measure deviant drivers’ attitudes. Violence and Victims, 15, 115e136. Wiesenthal, D. L., Hennessy, D. A., & Totten, B. (2000). The influence of music on driver stress. Journal of Applied Social Psychology, 30, 1709e1719. Wilkowski, B. M., & Robinson, M. D. (2010). The anatomy of anger: An integrative cognitive model of trait anger and reactive aggression. Journal of Personality, 78, 9e38. Wong, J. T., Chung, Y. S., & Huang, S. H. (2010). Determinants behind young motorcyclists’ risky riding behavior. Accident Analysis and Prevention, 42, 275e281. Wyon, D. P., Wyon, I., & Norin, F. (1996). Effects of moderate heat stress on driver vigilance in a moving vehicle. Ergonomics, 39, 61e75. Yagil, D. (2001). Reasoned action and irrational motives: A prediction of drivers’ intention to violate traffic laws. Journal of Applied Social Psychology, 31, 720e740. Yasak, Y., & Esiyok, B. (2009). Anger amongst Turkish drivers: Driving Anger Scale and its adapted, long and short version. Safety Science, 47, 138e144. Zuckerman, M. (1994). Behavioral expressions and biosocial bases of sensation seeking. New York: Cambridge University Press. Zuckerman, M. (2007). The Sensation Seeking Scale V (SSS-V): Still reliable and valid. Personality and Individual Differences, 43, 1303e1305.
This page intentionally left blank
Chapter 13
Mental Health and Driving Joanne E. Taylor Massey University, Palmerston North, New Zealand
1. MENTAL HEALTH IMPACTS On November 6, 2009, a high court jury in New Zealand found a 50-year-old man guilty of murdering a young newlywed on June 3, 2008, after he allegedly crashed into her car in an attempt to end his life. He had tried to kill himself several times before, including one occasion when he drank aftershave and detergent. On the night of the crash, he left his house after an argument with his wife. He had visited an alcohol store three times that afternoon and was purportedly swaying on his feet before he got into his car. He called his wife at approximately 6 pm, saying he was looking for a large semitrailer truck. Later in the evening, he drove directly at four separate cars, crossing the center line and making no attempt to swerve. He subsequently collided with a car carrying a man and his two young children on their way home from soccer practice, and then he hit the newlywed’s car head-on. She did not regain consciousness and died of significant internal injuries. Emergency service personnel discovered empty beer bottles in his car and a full can of beer propped on his lap after the crash. Suicide by intentional car crash is an extreme example of the effects of mental health on driving. There are many other ways in which mental health can impact driving and, conversely, driving can affect mental health, especially following a motor vehicle crash (MVC). The dynamic nature of the driving environment makes this a challenging area for researchers. Currently, there are two very separate bodies of literature that examine this area: traffic psychology research, which examines the effects of mental health on driving, mostly in an attempt to identify factors that might increase the likelihood of MVCs, and mental health research, which investigates the various psychological consequences of MVCs.
2. THE EFFECTS OF MENTAL HEALTH ON DRIVING Driving is a highly complex process. As information processors in the driving system, drivers must constantly Handbook of Traffic Psychology. DOI: 10.1016/B978-0-12-381984-0.10013-X Copyright Ó 2011 Elsevier Inc. All rights reserved.
receive, process, and respond to information derived from a constantly changing environment as well as modulate their internal states. Therefore, they require efficient cognitive and intrapersonal function. Several factors can influence the efficiency of cognitive function and in different ways, which may place drivers and other road users at increased risk of involvement in an MVC. In an attempt to understand the human error causes of MVCs, researchers have studied an exhaustive array of human factors, including mood, aggression, risk-taking behavior, fatigue, stress, age, gender, brain injury, drug-taking behavior, and psychiatric symptoms (Little, 1970; McDonald & Davey, 1996; Shinar, 1978; Taylor & Dorn, 2006). Some of these factors are considered in detail elsewhere in this volume (e.g., see Chapter 12 on social, personality, and affective constructs in driving; Chapter 17 on impaired driving; and Chapter 21 on fatigued driving), so the present material focuses on the relationships between mental health and driving. Much of the research in this area has retrospectively examined the prevalence of psychopathology in MVC victims or conducted laboratory-based investigations of driving with groups identified on the basis of particular mental health characteristics, examining factors that may impair and affect driving ability, such as attention, concentration, memory, vigilance, impulse control, judgment, problem solving, reaction time, and psychomotor control. It has long been suggested that mental health might contribute to road safety, and specifically that those who experience mental health problems, such as psychotic, mood, anxiety, or substance use difficulties, are more likely to be involved in MVCs. Several reviews of this area have been published, although they are somewhat dated (McDonald & Davey, 1996; Metzner et al., 1993; Noyes, 1985; Silverstone, 1988; Tsuang, Boor, & Fleming, 1985). Early studies examined what were then called “accidentprone” drivers and reported on their socially deviant characteristics (Tillman & Hobbs, 1949), as well as noting high levels of alcoholism in this group (Selling, 1940). Crancer and Quiring (1969) found that people with personality 165
166
disorders had an accident rate 144% higher than that of a matched control group, whereas it was 49% higher in a group with psychoneurotic disorders, and there was no increased rate compared with controls for a group with schizophrenia. Since these early studies, other researchers have reported higher traffic accident rates among people with alcohol use and personality disorders, particularly antisocial personality disorder (Armstrong & Whitlock, 1980; Dumais et al., 2005; Elkema, Brosseau, Koshick, & McGee, 1970; Selzer, Payne, Westervelt, & Quinn, 1967; Waller & Turkel, 1966; see also the review by Tsuang et al., 1985), as well as higher mortality rates in MVCs for these groups (Rorsman, Hagnell, & Lanke, 1982; Schuckit & Gunderson, 1977). Others have found no such increased accident likelihood for those with psychiatric histories, although these studies are limited by methodological issues such as excluding people with substance-related problems (Cushman, Good, & States, 1990; Kastrup, Dunpont, Bille, & Lund, 1977). Laboratory- or field-based studies have reported mixed findings, such as slowed driving speed and more errors and collisions on a simulator for a sample with schizophrenia compared with a matched control group (St. Germain, Kurtz, Pearlson, & Astur, 2004) and marked problems with psychomotor performance in a sample of psychiatric outpatients (de la Cuevas Castresana & Alvarez, 2009). However, the relationship between mental health and driving is complex, and simply examining the differential rates of MVC in those with or without various forms of psychopathology does not provide evidence that psychopathology plays a causal role in accidents. Accidents are often preceded by stressful life events, such as problems in interpersonal relationships, as well as driving under the influence of alcohol and other substances (Noyes, 1985). Various mechanisms have been proposed to explain the relationship between various types of mental health difficulty and MVCs. McDonald and Davey (1996) provide a detailed review of these factors, which are briefly outlined here.
2.1. Alcohol and Substance Use Problems The physiological, cognitive, and behavioral effects of alcohol, such as slowed reaction time, problems with coordination and attention, and lowered behavioral inhibition, clearly increase MVC risk, and this risk is greater for those who have a pathological problem with alcohol, such as alcohol abuse or dependence (which may also be comorbid with other psychopathology, such as antisocial personality and conduct disorder traits, depression, and post-traumatic stress disorder; del Rio & Alvarez, 2001; del Rio, GonzalezLuque, & Alvarez, 2001; McDonald & Davey, 1996; McMillan et al., 2008; Stoduto et al., 2008). Several epidemiological studies have examined psychopathology
PART | II
Key Variables to Understand in Traffic Psychology
in those convicted of driving under the influence of alcohol and have reported lifetime rates of alcohol use disorder of 41e91% and other drug use disorder of 26e40% (de Baca, Lapham, Skipper, & Hunt, 2004; Lapham, de Baca, McMillan, & Lapidus, 2006; Lapham et al., 2001; McCutcheon et al., 2009; Palmer, Ball, Rounsaville, & O’Malley, 2007; Shaffer et al., 2007). Psychotropic medication use is also relevant for drivers with mental health problems, although most of these medications are generally considered not to interfere with driving performance unless alcohol is also taken or the drugs are being abused (Hole, 2007, 2008).
2.2. Personality Traits and Disorders In some cases, such as alcohol abuse, the symptoms of the problem can directly influence vehicle control and therefore contribute to risky or unsafe driving. In others, however, such as personality disorders, the link between the mental health problem and driving behavior is less clear, and is probably influenced by a combination of factors, including personality traits such as aggression, hostility, impulsivity, and sensation-seeking. These latter traits might have indirect effects on MVCs through driving behaviors such as errors and violations (Donovan & Marlatt, 1982; Shaw & Sichel, 1971; Wilson & Jonah, 1988; Zuckerman & Neeb, 1980). Several studies have attempted to predict accident involvement from a variety of personality factors and driving behaviors (Kim, Nitz, Richardson, & Li, 1995; ˚ berg, 1999; Norris, Matthews, & Riad, 2000; Rimmo¨ & A Ulleberg & Rundmo, 2003; West, Elander, & French, 1993). Su¨mer (2003) developed a fairly sophisticated contextual mediated model of personality and behavioral predictors of MVCs that distinguished distal and proximal factors in 295 professional drivers in Turkey. Psychological symptoms, including anxiety, depression, hostility, and psychoticism, had direct effects on aberrant driving behaviors, and dysfunctional drinking had an indirect effect (via aberrant driving behaviors) on the number of accidents (Su¨mer, 2003). Other researchers have also supported the use of multiple personality predictors of unsafe driving that are related in different ways to different aspects of driving behavior (Dahlen & White, 2006).
2.3. Anger In addition to being investigated in research on personality traits and driving, anger in the driving situation has increasingly been studied during approximately the past decade, fuelled by interest in the phenomenon of road rage (Galovski, Malta, & Blanchard, 2006; Hole, 2007). Research has focused on developing instruments to measure driving anger that can distinguish aggressive driving behavior and
Chapter | 13
Mental Health and Driving
thoughts as well as risky, nonaggressive behavior (Deffenbacher, Lynch, Oetting, & Swaim, 2002; Deffenbacher, Oetting, & Lynch, 1994; Deffenbacher, White, & Lynch, 2004; DePasquale, Geller, Clarke, & Littleton, 2001; Lajunen, Parker, & Stradling, 1998). Others have explored associations of driving anger and aggression with psychopathology and have found that aggressive drivers are more likely to meet criteria for various mental health problems, particularly intermittent explosive disorder, current or past alcohol or substance abuse or dependence, antisocial and borderline personality disorders, conduct disorder, and attention-deficit/hyperactivity disorder (Fong, Frost, & Stansfeld, 2001; Galovski, Blanchard, & Veazey, 2002; Malta, Blanchard, & Freidenberg, 2005). Victims of road rage are also at risk of developing mental health problems (Smart, Ashbridge, Mann, & Adlaf, 2003). Despite the tendency for angry and aggressive drivers to report more risky driving behaviors, most research has found no significant correlations between measures of driving anger and accident involvement (Sullman, 2006; Van Rooy, Rotton, & Burns, 2006), although a few have found overall anger, in addition to other variables, to significantly predict crash involvement (Sullman, Gras, Cunill, Planes, & Font-Mayolas, 2007). Driving anger and aggression is generally considered to be a complex problem that depends on the characteristics of the driver and the situation (Lajunen & Parker, 2001; Shinar, 1998).
2.4. Depression Rates of depression in accident samples have not been clearly determined given the overlap in some cases of low mood and depression with self-harm and suicide (McDonald & Davey, 1996) and the use of self-report questionnaires instead of diagnostic methods to determine depression (Hilton, Staddon, Sheridan, & Whiteford, 2009). Other research on depression and driving has taken place in the broader context of evaluating the influence of emotions on driving, which may be related to various factors, including prejourney emotions and circumstances, the traffic situation while driving, and thoughts that arise during travel (Ban˜uls Egeda, Carbonell Vaya, Casanoves, & Chisvert, 1997; Groeger, 1997; Levelt, 2003). For example, Mesken, Hagenzieker, Rothengatter, and de Waard (2007) found that emotions while driving were related to emotional traits as well as traffic events such that anger was associated with other-blame and events affecting progress, and anxiety was associated with situation-blame and events affecting safety.
2.5. Anxiety Discussions in the general driving literature that have related anxiety to driving have come from broader studies
167
of personality typologies and disorders (Evans, 1991; Foot & Chapman, 1982; Heimstra, Ellingstad, & DeKock, 1967; Little, 1970; Loo, 1979; Shinar, 1978; Shoham, Rahav, Markovski, Chard, & Baruch, 1984; Silverstone, 1988; Wilson & Greensmith, 1983), and stress (Gulian, Glendon, Matthews, Davies, & Debney, 1988, 1990; Heimstra, 1970; Hentschel, Bijleveld, Kiessling, & Hosemann, 1993). Some research suggests that anxiety necessarily impairs driving. Shoham et al. (1984) used a combination of personality variables to predict the likelihood of recidivist traffic accident involvement, and they reported that drivers characterized as anxious manifested high internalization of traffic norms and high levels of anxiety and “were found to have lowered bio-psychogenic [sic] control over the basic mechanisms required for driving” (p. 184). Other authors have considered that anxiety affects driving in a more complex manner and may also have some facilitative or positive effects that are specific to driver behavior and driving skills (Kottenhoff, 1961; O’Hanlon, Vermeeren, Uiterwijk, van Veggel, & Swijgman, 1995; Payne & Corley, 1994; Silverstone, 1988). For example, a moderate amount of anxiety may enable the driver to carry out all of the basic skills required for driving, as well as to pay sufficient attention to potential hazards so that the appropriate action can be taken if required (Walklin, 1993), whereas high levels of anxiety could interfere with driving performance and increase the risk of an MVC through errors, indecision, and hesitation (Carbonell, Banuls, Chisvert, Monteagudo, & Pastor, 1997; Silverstone, 1988; Walklin, 1993). Yinon and Levian (1988) found that anxiety about being in the presence of other drivers leads to the division of attention between self- and task-relevant stimuli, although the focus on threat appraisals in anxiety disorders may actually explain why the MVC rate is no higher than population norms for this group (McDonald & Davey, 1996). Indeed, Taylor, Deane, and Podd (2007) found that anxious drivers made more errors than controls in an on-road driving test, but there were no differences in MVC history.
2.6. Attention-Deficit/Hyperactivity Disorder During approximately the past decade, researchers have started to investigate the influence of adult attention-deficit/ hyperactivity disorder (ADHD) on driving behavior, particularly because of the features of difficulties with attention and impulsivity that characterize this condition. ADHD has been found to present risks to safe driving in terms of traffic violations, license suspensions, less safe driving practices, more driving errors, and crashes in simulator performance as well as on the road (Fischer, Barkley, Smallish, & Fletcher, 2007; Nada-Raja et al., 1997). These risks may be partly accounted for by the fact that ADHD can also be associated with other risk factors, such as tendencies to frustration and aggression (Richards, Deffenbacher,
168
Rose´n, Barkley, & Rodricks, 2006), as well as substance use and conduct problems (Jerome, Segal, & Habinski, 2006). Treatment with long-acting methylphenidate improves the driving performance of adolescents and adults with ADHD (Barkley, Murphy, O’Connell, & Connor, 2005), although it is unknown whether this also reduces their risk of MVCs or traffic violations (Barkley, 2010).
2.7. Stress Although much research has examined the relationship between various types of mental health difficulty and driving performance as well as MVCs, another strand of research has investigated the role of stress as a more general problem linked to mental health that might make people with (or without) psychopathology more vulnerable to MVCs. One difficulty in this research is that stress can act both as a cause and as a consequence of mental health problems, so studies that do not control for psychopathology are limited (McDonald & Davey, 1996). Matthews (2001) attempted to identify the informationprocessing functions that mediate the effects of stress on driving performance impairment in developing a transactional model of driver stress (Matthews et al., 1998). Stress variables used have been based on factor analyses of the Driving Behaviour Inventory (Glendon et al., 1993; Gulian, Matthews, Glendon, Davies, & Debney, 1989) and its revision, the Driver Stress Inventory (Matthews, Desmond, Joyner, Carcary, & Gilliland, 1997), which are considered to represent vulnerabilities to different types of stress outcome, including aggression and anxiety. Effects of driver stress on driving performance can depend on the nature of the driver’s stress reactions (e.g., appraisal of the demands of the traffic environment, including other drivers; appraisal of personal competence; and coping strategy), the traffic environment, and the demands of the driving task (Matthews, 2001; Matthews, Emo, & Funke, 2005; Matthews et al., 1997, 1998, 1999). The approach used to cope with stress has been identified as an important factor influencing the perception of stress and whether it affects driving behavior, and maladaptive coping strategies such as alcohol abuse may increase accident risk (McDonald & Davey, 1996).
3. EFFECTS OF DRIVING ON MENTAL HEALTH Various aspects of driving can also impact mental health. Most notably, involvement in MVCs can have varied mental health effects, ranging from little or no impact to significant and marked difficulties. Several books have been dedicated to this topic (Blanchard & Hickling, 2004; Duckworth, Iezzi, & O’Donohue, 2008; Hickling & Blanchard, 1999; Mitchell, 1997). It is also important to
PART | II
Key Variables to Understand in Traffic Psychology
note that, in addition to those directly involved and injured in the crash as primary victims, MVCs can also affect (1) those who are uninjured but are still involved in the crash as either participants or witnesses; (2) friends and family who hear descriptions of the event from someone involved; and (3) those involved in dealing with the after effects of MVCs, including police, fire, ambulance, and hospital personnel, as well as those who are charged with preparing medical or legal reports (Mayou, 1997; Mitchell, 1999; Taylor & Koch, 1995).
3.1. Mental Health Consequences of Motor Vehicle Crashes The nature and sheer frequency of MVCs suggests that at least some people are likely to suffer psychological repercussions. Research has demonstrated that people who have been involved in MVCs and other common accidents may manifest chronic psychological dysfunction, even in the context of minimal physical injury and good recovery (Horne, 1993; Pilowsky, 1985). Longitudinal research has demonstrated a variety of psychosocial sequelae in injured MVC survivors, including worsened family or spouse relationships as well as reduced social contact, pleasure from leisure activities, and work capacity (Malt, Hfivik, & Blikra, 1993). Research on the psychiatric consequences of MVCs has documented widespread implications for psychological functioning, including depressive, anxious, and phobic symptoms, as well as multiple disturbances. However, some of these studies have combined MVC victims with victims from a range of industrial and workrelated accidents, making it impossible to clearly determine psychiatric morbidity after MVCs (Culpan & Taylor, 1973; Jones & Riley, 1987; Shalev et al., 1998). Post-MVC research has ranged from examining specific effects such as depersonalization responses (Noyes, Hoenk, Kuperman, & Slymen, 1977) to comprehensive studies of MVC-induced psychopathology. The most common psychological sequelae of MVCs include driving-related fears and avoidance, post-traumatic stress disorder (PTSD), depression, and pain-related syndromes (Blanchard, Hickling, Taylor, Loos, & Gerardi, 1994; Goldberg & Gara, 1990; Koch & Taylor, 1995; Kuch, Cox, Evans, & Shulman, 1994; Malt, 1988; Mayou, 1992; Mayou, Bryant, & Duthie, 1993). For victims with multiple injuries, depression and anxiety are particularly common (Mayou et al., 1993). Approximately 40% of MVC victims suffer comorbid conditions, such as major depression, panic disorder, specific phobia, eating disorder, substance abuse, and personality disorder (Blanchard et al., 1994). Furthermore, blaming others for the crash has been found to be associated with higher levels of psychological distress and lower psychological well-being for both passengers and
Chapter | 13
Mental Health and Driving
drivers (Ho, Davidson, Van Dyke, & Agar-Wilson, 2000). However, consistent prevalence data for the various psychological outcomes have been elusive, likely due to the impact of several methodological issues, such as varying definitions of terms, the use of samples that are not generalizable to MVC victims (e.g., medicolegal samples, victims referred for treatment, and hospital or primary care attendees), and lack of consideration of injury severity (Blaszczynski et al., 1998). Significant psychological problems have been identified even among victims of minor traffic crashes. For example, one longitudinal study reported that 4 months after a minor MVC (i.e., outpatient treatment only, with no hospitalization or head injury), 13% of 39 participants scored above the cutoff for a positive PTSD diagnosis on a self-report measure, 36% reported symptoms of anxiety, and 16% described avoiding using their car, motorcycle, or bicycle (Smith, MacKenzie-Ross, & Scragg, 2007).
3.1.1. Post-Traumatic Stress Reactions The most commonly studied psychological outcomes of MVCs are those involving some kind of trauma response, either in the immediate weeks following the crash or in the longer term, and that may range from subthreshold-level symptoms to full-blown clinical syndromes, such as acute and post-traumatic stress disorder. Both are characterized by problematic experiences of anxiety that can occur following a traumatic event involving “actual or threatened death or serious injury, or a threat to the physical integrity of self or others” and in which the person responds to the event with horror, fear, or helplessness (American Psychiatric Association, 2000, pp. 467 and 471). The symptoms of these trauma responses can include psychologically re-experiencing the trauma (e.g., intrusive thoughts and nightmares), increased physical arousal (e.g., exaggerated startle response and irritability), and persistent avoidance related to the crash (e.g., avoidance of or reluctance to drive and avoiding thoughts or conversations about the crash). A body of research exists that indicates the frequent occurrence of PTSD in MVC victims and its impact on quality of life (Gudmundsdottir, Beck, Coffey, Miller, & Palyo, 2004), and it has been suggested that PTSD thoroughly captures the psychological consequences of MVCs (Burstein, 1989b; Davis & Breslau, 1994; Hickling, Blanchard, Silverman, & Schwarz, 1992; Kuch, Swinson, & Kirby, 1985; Platt & Husband, 1987). Studies have investigated the treatment of post-traumatic responses after MVCs (Blanchard, Hickling, Taylor, et al., 1996; Brom, Kleber, & Hofmann, 1993; Fairbank, DeGood, & Jenkins, 1981; Green, McFarlane, Hunter, & Griggs, 1993; McCaffrey & Fairbank, 1985; Walker, 1981) as well as the complicating nature of PTSD in post-traumatic headache (Davis & Breslau, 1994; Hickling, Blanchard, Schwarz, &
169
Silverman, 1992; Hickling, Blanchard, Silverman, et al., 1992) and the nature of psychophysiological responding in MVC-related PTSD (Blanchard, Hickling, & Taylor, 1991). However, it has been noted that avoidant symptoms may obscure the identification of PTSD reactions in MVC victims, and PTSD may go unrecognized for some time after the accident (Burstein, 1989a, 1989b; Epstein, 1993). Reported incidence rates of MVC-related PTSD vary considerably across studies, largely due to methodological differences, especially in terms of the sample included and approach used to ascertain PTSD. Table 13.1 provides an overview of studies that have examined adult MVC samples using well-validated structured diagnostic interviews or self-report measures that represent diagnostic criteria. Investigations have differentiated between PTSD among respondents involved in “serious” MVCs, in which there was some degree of physical injury requiring hospitalization, and “non-serious” MVCs, or those not resulting in bodily injury or involving the subjective experience of psychological injury. However, because victims of both serious and non-serious MVCs have been found to experience PTSD, these distinctions may have little utility. Table 13.1 demonstrates this phenomenon, whereby research using seriously injured victims has found a range of rates of PTSD (from 1 to 100%), as have studies that have utilized victims sustaining relatively minor injuries (15e50%). Although there appears to be a larger range for serious MVCs, the overlap in incidence rates is substantial and may be due to definitional (e.g., different criteria for severity of injury, diagnosis, and time since the MVC) and methodological differences (e.g., whether the sample was seeking treatment or not). The various studies shown in Table 13.1 indicate that the rate of PTSD for seriously injured victims 1 month following an MVC is approximately 25e56% (Blanchard et al., 1994; Blanchard, Hickling, & Barton, 1996; Blanchard, Hickling, & Taylor, et al., 1996; Chubb & Bisson, 1999; Feinstein & Dolan, 1991; Ursano et al., 1999). Rates at 3e6 months decrease to 7e30% (Bryant, Harvey, Guthrie, & Moulds, 2000; Ehlers, Mayou, & Bryant, 1998; Hamanaka et al., 2006; Harvey & Bryant, 1998; Mayou et al., 1993; Ursano et al., 1999; Yasan, Gu¨zel, Tamam, & Ozkan, 2009), and they decrease to 5e32% at 12 months (Blanchard, Hickling, Barton, et al., 1996; Blanchard, Hickling, Taylor, et al., 1996; Ehlers et al., 1998; Green et al., 1993; Koren, Arnon, & Klein, 1999; Mayou et al., 1993). Increasingly, studies have documented MVC-related PTSD in children, some of which have reported cases with onset as young as 2 years (Jaworowski, 1992; Jones & Peterson, 1993; McDermott & Cvitanovich, 2000; Thompson & McArdle, 1993). Symptomatology in such cases has included reliving the MVC through nightmares, conduct difficulties, separation anxiety, enuresis, fear of the dark, trauma-specific fears, sleep disturbance, violent play,
170
PART | II
reluctance to cross roads or travel by car, and a preoccupation with road safety (Canterbury & Yule, 1997; Jones & Peterson, 1993; Taylor & Koch, 1995; Thompson & McArdle, 1993). Studies using approaches that permit diagnostic assessment have typically found higher rates of PTSD in
Key Variables to Understand in Traffic Psychology
children than have studies using other criteria to determine the presence of PTSD, such as cutoff scores on a measure of PTSD-type symptoms (e.g., 12e18% from 3 to 12 months post-MVC; Landolt, Vollrath, Gnehm, & Sennhauser, 2009; Landolt, Vollrath, Timm, Gnehm, & Sennhauser, 2005; Sturms et al., 2005).
TABLE 13.1 Prevalence of PTSD in Adults Following MVCs Using DSM-Based Assessment References
N
Injury Severity Criteria
% PTSD
Time since MVC
Kuch et al. (1985)
30
Medical attention sought
100a
NR
Malt (1988)
107
Hospitalized
1
6 months
Feinstein and Dolan (1991)
48
Accidentally injured
25 15
6 weeks 6 months
Hickling, Blanchard, Silverman, et al. (1992)
20
Post-traumatic headache
75b
NR
b
NR
b
Serious Injury
Epstein (1993)
15
Serious injury
40
Green et al. (1993)
24
Severe injury
25
18 months
Mayou et al. (1993)
188
Multiple injury or whiplash neck injury
7e9 5e11
3 months 12 months
Blanchard et al. (1994)
50
Medical attention sought
46
1e4 months
Blanchard, Hickling, Barton, et al. (1996); Blanchard, Hickling, Taylor, et al. (1996)
158
Medical attention sought
39b 12b
1e4 months 12 months
Ehlers et al. (1998)
967
Emergency department attendees
23.1c c
Mayou et al. (2002) Harvey and Bryant (1998)
71
Hospitalized
3 months
16.5 11c
12 months 3 years
25.4c
6 months
c
Chubb and Bisson (1999)
24
Many physically injured
56.3 37.5c
5 weeks 9 months
Koren et al. (1999)
74
Hospitalized
32b
12 months b
Ursano et al. (1999)
122
Most hospitalized
34.4 25.3b 18.2b 17.4b 14b
1 month 3 months 6 months 9 months 12 months
Bryant et al. (2000)
113
Hospitalized
21c
6 months
Hamanaka et al. (2006) Matsuoka et al. (2008) Yasan et al. (2009)
100 100
Severe injuries In intensive care
c
6 months
8.5 c
1 month
8
c
95
Emergency department attendees
29.8 23.1c 17.9c
3 months 6 months 12 months
Goldberg and Gara (1990)
55
Not resulting in death or major bodily injury
15
M ¼ 15 months
Kuch et al. (1994)
21
Minimal injury and chronic pain
38b
NR
Non-Serious Injury
Kupchik et al. (2007)
60
General health outpatient
c
50
M ¼ 44 months
Chapter | 13
Mental Health and Driving
171
TABLE 13.1 Prevalence of PTSD in Adults Following MVCs Using DSM-Based Assessmentdcont’d References
N
Injury Severity Criteria
% PTSD
Time since MVC
20
NR
50b
NR
a
Injury Criteria Not Given Hickling and Blanchard (1992) Horne (1993)
7
NR
43
Dalal and Harrison (1993)
86
NR
32
Kuch, Cox, and Direnfeld (1995)
54
NR
22
Chan, Air, and McFarlane (2003)
391
NR
29
M ¼ 2 years M ¼ 2.7 years M ¼ 3.6 years
c
9 months
NR, not reported. a DSM-III. b DSM-III-R. c DSM-IV.
Diagnostic studies have reported higher child PTSD rates both in the short term and in the longer term following an MVC, including up to 4e6 weeks (23e35%; MeiserStedman, Smith, Glucksman, Yule, & Dalgleish, 2008; Stallard, Velleman, & Baldwin, 1998), 3 months (22e25%; McDermott & Cvitanovich, 2000; Scha¨fer, Berkmann, Riedasser, & Schulte-Markwort, 2006), and 6 months (13e19%; Meiser-Stedman et al., 2008; Mirza, Bhadrinath, Goodyer, & Gilmour, 1998). Various predictors of PTSD have been identified in the literature, related to both the early development of the trauma response and whether that response becomes chronic. Predictors of early stage development of PTSD include the presence of acute stress disorder, persistent physical disability, severity of physical injury, a sense of threat to life, dissociation during the crash, low perceptions of coping self-efficacy, and lower perceived social support (Benight, Cieslak, Molton, & Johnson, 2008; Hamanaka et al., 2006; Koren et al., 1999; Matsuoka et al., 2008; Yasan et al., 2009). Factors that have been shown to predict chronic PTSD (symptoms experienced for 1 year or more) include some of the previously mentioned factors, such as early trauma symptoms (including sleep problems), perceived threat, dissociation during the crash, and persistent health problems, as well as other variables, including hospitalization for injuries, persistent financial problems, litigation, female gender, unemployment, emotional problems prior to the crash (including distress from and amount of prior trauma), alcohol abuse, and poor social support before and after the crash (Ameratunga, Tin Tin, Coverdale, Connor, & Norton, 2009; Beck, Palyo, Canna, Blanchard, & Gudmundsdottir, 2006; Blanchard, Hickling, Barton, et al., 1996; Buckley, Blanchard, & Hickling, 1996; Do¨rfel, Rabe, & Karl, 2008; Ehlers et al., 1998; Fujita & Nishida, 2008; Irish et al., 2008; Koren, Arnon, Lavie, & Klein, 2002; Mayou & Bryant, 2002; Mayou, Ehlers, & Bryant,
2002; Mayou, Tyndel, & Bryant, 1997; Murray, Ehlers, & Mayou, 2002). Several cognitive factors, especially cognitive processes, have also been found to maintain and predict PTSD (in some cases to a greater degree than the established predictors noted previously). These cognitive factors include negative interpretations of intrusive memories, rumination, memory disorganization, thought suppression, anger cognitions, and general negative post-traumatic thoughts (Ehring, Ehlers, & Glucksman, 2006, 2008; Ehring, Frank, & Ehlers, 2008; Holeva, Tarrier, & Wells, 2001; Karl, Rabe, Zo¨llner, Maercker, & Stopa, 2009; Murray et al., 2002). A few studies have examined predictors of PTSD development in children who have experienced MVCs, and early trauma symptom severity has been identified as the strongest predictor (Landolt et al., 2005; Scha¨fer et al., 2006), along with severity of MVC-related PTSD in the father. However, variables such as age, gender, injury severity, threat appraisal, prior trauma exposure, prior mental health problems, and maternal MVC-related PTSD have not been found to be significant predictors (Landolt et al., 2005; Meiser-Stedman, Dalgleish, Glucksman, Yule, & Smith, 2009). The presence of nightmares with content that exactly matched the trauma has been found to strongly predict PTSD scores at 2 and 6 months post-MVC, although this finding needs to be replicated with a larger sample (Wittman, Zehnder, Schredl, Jenni, & Landolt, 2010). In line with adult research, studies are beginning to document evidence for the role of various cognitive factors in predicting the development and maintenance of PTSD following MVCs in children. For example, research has examined maladaptive cognitive appraisals such as the meaning of the trauma and trauma symptoms, future vulnerability, rumination, anxiety sensitivity, and the quality of trauma memories (Meiser-Stedman et al., 2009). Information about approaches to treatment of PTSD
172
following an MVC is widely available, and readers are referred to the existing literature for more information (Blanchard & Hickling, 2004; Duckworth et al., 2008; Hickling & Blanchard, 1999; Hickling, Kuhn, & Beck, 2008).
3.1.2. Driving-Related Fear, Phobia, and Travel Anxiety In addition to post-traumatic reactions, research has also demonstrated that other types of fear reactions to MVCs are common and can be extremely debilitating (Herda, Ehlers, & Roth, 1993; Kuch, 1997; Kuch, Cox, & Evans, 1996; Kuch, Evans, & Mueller-Busch, 1993). Fear of driving is a diverse experience, ranging from mild driving reluctance to driving phobia as a variant of specific phobia. Research on these phobic fear reactions has focused on avoidance of or reduction in driving, endurance of driving with marked discomfort, and the effect of fear on a person’s lifestyle and everyday functioning. However, variations in terminology and definitions used as well as sampling issues have led to vast inconsistencies in prevalence estimates for problems such as driving phobia, accident phobia, travel phobia, driving reluctance, and phobic travel anxiety. In particular, studies that have used broader criteria in which complete avoidance is unnecessary have reported higher rates (e.g., 60e77%; Hickling & Blanchard, 1992; Kuch et al., 1985) than those that have specified total avoidance for diagnosis (e.g., 2e6%; Blanchard et al., 1994; Hickling & Blanchard, 1999). Driving fear and phobia can also occur in the absence of MVCs but still be severe and have a marked effect on functioning (Ehlers, Hofmann, Herda, & Roth, 1994; Taylor & Deane, 1999, 2000; Taylor, Deane, & Podd, 1999, 2000). Driving phobia is most appropriately considered to be a situational type of specific phobia that is characterized by marked and persistent fear that is excessive or unreasonable, is cued by anticipation of or exposure to driving stimuli, is associated with avoidance of driving stimuli or endurance of such stimuli with considerable anxiety or distress, and has a marked impact on the person’s functioning (American Psychiatric Association, 2000). The content of fear can be much broader than the fear of driving and can relate to various aspects of travel and accident-related stimuli, such as fear of riding in a vehicle as a passenger while having no fear of driving (Koch & Taylor, 1995). Blanchard and Hickling (2004) refer to less phobic forms of driving-related fear as driving reluctance, where the person is able to make essential journeys but avoids nonessential travel or tolerates it with some degree of anxiety. Several reviews of driving-related fear and phobia provide more comprehensive information on this topic, along with information on appropriate interventions (Taylor, 2008; Taylor, Deane, & Podd, 2002).
PART | II
Key Variables to Understand in Traffic Psychology
3.1.3. Other Problems Several studies have documented increased rates of depression and other mood disorders following MVCs that may or may not be comorbid with PTSD, with rates of major depression ranging from 6 to 53% 1 year post-MVC (Blanchard et al., 2004; Blanchard, Hickling, Taylor, & Loos, 1995; Dickov et al., 2009; Ehlers et al., 1998). Blanchard, Hickling, Taylor, et al. (1996) also reported that major depression prior to the MVC was a significant predictor of post-MVC PTSD. Depression has been identified as a common consequence of MVCs and one that can overlap with physical effects such as chronic pain and head injury and contribute markedly to functional limitations following a crash (Duckworth, 2008). Substance use disorders have been examined in relation to MVCs, although the mixed findings reported in the literature suggest that longitudinal research is needed to more clearly identify the relationship between substance use and MVC (O’Donnell, Creamer, & Ludwig, 2008). Psychological factors can also contribute in many ways to pain-related syndromes that occur following MVC-related injury (Duckworth et al., 2008).
4. SUMMARY The relationship between mental health and driving is complex. Mental health can have an impact on driving behavior and performance, although the relationships are multifaceted and depend on factors related to the specific nature of the problem, other characteristics of the individual, and specific aspects of the traffic environment and driving situation. However, higher accident rates for those with alcohol-related disorders as well as antisocial personality disorder are relatively consistent findings, although these conditions are also frequently comorbid. The absence of clear information from methodologically sound studies about how mental health affects safe driving can present difficulties for health professionals who are required to make decisions regarding fitness to drive in cases in which mental health is an issue (Knapp & VandeCreek, 2009; Me´nard et al., 2006). The complexity of the relationships involved necessitates that current guidelines focus on the importance of individualized assessment and consideration of factors such as acute illness symptoms as well as side effects of, interactions among, and compliance with medication (Carr, Schwartzberg, Manning, & Sempek, 2010; Land Transport Safety Authority, 2002). Several quite different types of mental health problems have tended to emerge from research on the psychological consequences of MVCs and highlight the diverse and complex types of experiences that might also be influenced by pre-accident mental health as well as the specific nature of the incident and response characteristics.
Chapter | 13
Mental Health and Driving
Although the two fields of traffic psychology and mental health and driving have historically been considered separately, considering them both provides a more comprehensive overview of the role of mental health in driving, particularly in highlighting the ways that mental health might influence driving and be affected by the driving environment.
REFERENCES American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (4th ed.). text revision. Washington, DC. Ameratunga, S., Tin Tin, S., Coverdale, J., Connor, J., & Norton, R. (2009). Posttraumatic stress among hospitalized and nonhospitalized survivors of serious car crashes: A population-based study. Psychiatric Services, 60, 402e404. Armstrong, J. L., & Whitlock, F. A. (1980). Mental illness and road traffic accidents. Australian and New Zealand Journal of Psychiatry, 14, 53e60. Ban˜uls Egeda, F., Carbonell Vaya, E., Casanoves, M., & Chisvert, M. (1997). Different emotional responses in novice and professional drivers. In T. Rothegatter, & E. Carbonell Vaya (Eds.), Traffic and transport psychology: Theory and application (pp. 343e352). Oxford: Pergamon. Barkley, R. A. (2010). Attention deficit hyperactivity disorder in adults: The latest assessment and treatment strategies. Sudbury, MA: Jones & Bartlett. Barkley, R. A., Murphy, K. R., O’Connell, T., & Connor, D. F. (2005). Effects of two doses of methylphenidate on simulator driving performance in adults with attention deficit hyperactivity disorder. Journal of Safety Research, 36, 121e131. Beck, J. G., Palyo, S. A., Canna, M. A., Blanchard, E. B., & Gudmundsdottir, B. (2006). What factors are associated with the maintenance of PTSD after a motor vehicle accident? The role of sex differences in a help-seeking population. Journal of Behavior Therapy and Experimental Psychiatry, 37, 256e266. Benight, C. C., Cieslak, R., Molton, I. R., & Johnson, L. E. (2008). Selfevaluative appraisals of coping capability and posttraumatic distress following motor vehicle accidents. Journal of Consulting and Clinical Psychology, 76, 677e685. Blanchard, E. B., & Hickling, E. J. (2004). After the crash: Psychological assessment and treatment of survivors of motor vehicle accidents (2nd ed.). Washington, DC: American Psychological Association. Blanchard, E. B., Hickling, E. J., Barton, K. A., Taylor, A. E., Loos, W. R., & Jones-Alexander, J. (1996). One-year prospective follow-up of motor vehicle accident victims. Behaviour Research and Therapy, 34, 775e786. Blanchard, E. B., Hickling, E. J., Freidenberg, B. M., Malta, L. S., Kuhn, E., & Sykes, M. A. (2004). Two studies of psychiatric morbidity among motor vehicle accident survivors 1 year after the crash. Behaviour Research and Therapy, 42, 569e583. Blanchard, E. B., Hickling, E. J., & Taylor, A. E. (1991). The psychophysiology of motor vehicle accident related post-traumatic stress disorder. Biofeedback and Self-Regulation, 16, 449e458. Blanchard, E. B., Hickling, E. J., Taylor, A. E., & Loos, W. R. (1995). Psychiatric morbidity associated with motor vehicle accidents. Journal of Nervous and Mental Disease, 183, 495e504.
173
Blanchard, E. B., Hickling, E. J., Taylor, A. E., Loos, W. R., Forneris, C. A., & Jaccard, J. (1996). Who develops PTSD from motor vehicle accidents? Behaviour Research and Therapy, 34, 1e10. Blanchard, E. B., Hickling, E. J., Taylor, A. E., Loos, W. R., & Gerardi, R. J. (1994). Psychological morbidity associated with motor vehicle accidents. Behaviour Research and Therapy, 32, 283e290. Blaszczynski, A., Gordon, K., Silove, D., Sloane, D., Hillman, K., & Panasetis, P. (1998). Psychiatric morbidity following motor vehicle accidents: A review of methodological issues. Comprehensive Psychiatry, 39, 111e121. Brom, D., Kleber, R. J., & Hofmann, M. C. (1993). Victims of traffic accidents: Incidence and prevention of post-traumatic stress disorder. Journal of Clinical Psychology, 49, 131e140. Bryant, R. A., Harvey, A. G., Guthrie, R. M., & Moulds, M. L. (2000). A prospective study of physiological arousal, acute stress disorder, and posttraumatic stress disorder. Journal of Abnormal Psychology, 109, 341e344. Buckley, T. C., Blanchard, E. B., & Hickling, E. J. (1996). A prospective examination of delayed onset PTSD secondary to motor vehicle accidents. Journal of Abnormal Psychology, 105, 617e625. Burstein, A. (1989a). Intrusion and avoidance symptoms in PTSD. American Journal of Psychiatry, 146, 1518. Burstein, A. (1989b). Posttraumatic stress disorder in victims of motor vehicle accidents. Hospital and Community Psychiatry, 40, 295e297. Canterbury, R., & Yule, W. (1997). The effects on children of road accidents. In M. Mitchell (Ed.), The aftermath of road accidents: Psychological, social and legal consequences of an everyday trauma (pp. 59e69). London: Routledge. Carbonell, E. J., Banuls, R., Chisvert, M., Monteagudo, M. J., & Pastor, G. (1997). A comparative study of anxiety responses in traffic situations as predictors of accident rates in professional drivers. In Human factors in road traffic II: Traffic psychology and engineering. Proceedings of the second seminar on human factors in road traffic (pp. 186e192). Braga, Portugal: University of Minho. Carr, D. B., Schwartzberg, J. G., Manning, L., & Sempek, J. (2010). Physician’s guide to assessing and counseling older drivers (2nd ed.). Washington, DC: National Highway Traffic Safety Administration. Chan, A. O. M., Air, T. M., & McFarlane, A. C. (2003). Posttraumatic stress disorder and its impact on the economic and health costs of motor vehicle accidents in South Australia. Journal of Clinical Psychiatry, 64, 175e181. Chubb, H. L., & Bisson, J. I. (1999). The development of psychological reactions in a group of individuals with pre-existing and enduring psychiatric disorder following a fatal coach crash. Journal of Mental Health, 8, 203e212. Crancer, A., Jr., & Quiring, D. L. (1969). The mentally ill as motor vehicle operators. American Journal of Psychiatry, 126, 807e813. Culpan, R., & Taylor, C. (1973). Psychiatric disorders following road traffic and industrial injuries. Australian and New Zealand Journal of Psychiatry, 7, 32e39. Cushman, L. A., Good, R. G., & States, J. D. (1990). Psychiatric disorders and motor vehicle accidents. Psychological Reports, 67, 483e489. Dahlen, E. R., & White, R. P. (2006). The Big Five factors, sensation seeking, and driving anger in the prediction of unsafe driving. Personality and Individual Differences, 41, 903e915. Dalal, B., & Harrison, G. (1993). Psychiatric consequences of road traffic accidents: Consider somatoform pain disorder. British Medical Journal, 307, 1282.
174
Davis, G. C., & Breslau, N. (1994). Post-traumatic stress disorder in victims of civilian trauma and criminal violence. Psychiatric Clinics of North America, 17, 289e299. de Baca, J. C., Lapham, S. C., Skipper, B. J., & Hunt, W. C. (2004). Psychiatric disorders of convicted DWI offenders: A comparison among Hispanics, American Indians and non-Hispanic whites. Journal of Studies on Alcohol, 65, 419e427. de la Cuevas Castresana, C., & Alvarez, E. J. S. (2009). Mental illness and road safety. Actas Espanolas de Psiquitria, 37, 75e81. Deffenbacher, J. L., Lynch, R. S., Oetting, E. R., & Swaim, R. C. (2002). The Driving Anger Expression Inventory: A measure of how people express their anger on the road. Behaviour Research and Therapy, 40, 717e737. Deffenbacher, J. L., Oetting, E. R., & Lynch, R. S. (1994). Development of a driving anger scale. Psychological Reports, 74, 83e91. Deffenbacher, J. L., White, G. S., & Lynch, R. S. (2004). Evaluation of two new scales assessing driving anger: The Driving Anger Expression Inventory and the Driver’s Angry Thoughts Questionnaire. Journal of Psychopathology and Behavioral Assessment, 26, 87e99. del Rio, M. C., & Alvarez, F. J. (2001). Illicit drugs and fitness to drive: Assessment in Spanish medical driving test centres. Drug and Alcohol Dependence, 64, 19e25. del Rio, M. C., Gonzalez-Luque, J. C., & Alvarez, F. J. (2001). Alcoholrelated problems and fitness to drive. Alcohol and Alcoholism, 36, 256e261. DePasquale, J. P., Geller, E. S., Clarke, S. W., & Littleton, L. C. (2001). Measuring road rage: Development of the Propensity for Angry Driving Scale. Journal of Safety Research, 32, 1e16. Dickov, A., Martinovic-Mitrovic, S., Vuckovic, N., SiladjiMladenovic, D., Mitrovic, D., Jovicevic, M., et al. (2009). Psychiatric consequences of stress after a vehicle accident. Psychiatria Danubina, 21, 483e489. Donovan, D. M., & Marlatt, G. A. (1982). Personality subtypes among driving-while-intoxicated offenders: Relationship to drinking behaviour and driving risk. Journal of Consulting and Clinical Psychology, 50, 241e249. Do¨rfel, D., Rabe, S., & Karl, A. (2008). Coping strategies in daily life as protective and risk factors for post traumatic stress in motor vehicle accident survivors. Journal of Loss and Trauma, 13, 422e440. Duckworth, M. P. (2008). Psychological conditions associated with motor vehicle collisions. In M. P. Duckworth, T. Iezzi, & W. T. O’Donohue (Eds.), Motor vehicle collisions: Medical, psychosocial, and legal consequences (pp. 311e344). San Diego: Academic Press. Duckworth, M. P., Iezzi, T., & O’Donohue, W. T. (Eds.). (2008). Motor vehicle collisions: Medical, psychosocial, and legal consequences. San Diego: Academic Press. Dumais, A., Lesage, A. D., Boyer, R., Lalovic, A., Chawky, N., Me´nardButeau, et al.. (2005). Psychiatric risk factors for motor vehicle fatalities in young men. Canadian Journal of Psychiatry, 50, 838e844. Ehlers, A., Hofmann, S. G., Herda, C. A., & Roth, W. T. (1994). Clinical characteristics of driving phobia. Journal of Anxiety Disorders, 8, 323e339. Ehlers, A., Mayou, R. A., & Bryant, B. (1998). Psychological predictors of chronic posttraumatic stress disorder after motor vehicle accidents. Journal of Abnormal Psychology, 107, 508e519. Ehring, T., Ehlers, A., & Glucksman, E. (2006). Contribution of cognitive factors to the prediction of post-traumatic stress disorder, phobia and
PART | II
Key Variables to Understand in Traffic Psychology
depression after motor vehicle accidents. Behaviour Research and Therapy, 44, 1699e1716. Ehring, T., Ehlers, A., & Glucksman, E. (2008). Do cognitive models help in predicting the severity of posttraumatic stress disorder, phobia, and depression after motor vehicle accidents? A prospective longitudinal study. Journal of Consulting and Clinical Psychology, 76, 219e230. Ehring, T., Frank, S., & Ehlers, A. (2008). The role of rumination and reduced concreteness in the maintenance of posttraumatic stress disorder and depression following trauma. Cognitive Therapy and Research, 32, 488e506. Elkema, R. C., Brosseau, J., Koshick, R., & McGee, C. (1970). A statistical study on the relationship between mental illness and traffic accidents: A pilot study. American Journal of Public Health, 60, 459e469. Epstein, R. S. (1993). Avoidant symptoms cloaking the diagnosis of PTSD in patients with severe accidental injury. Journal of Traumatic Stress, 6, 451e458. Evans, L. (1991). Traffic safety and the driver. New York: Van Nostrand Reinhold. Fairbank, J. A., DeGood, D. E., & Jenkins, C. W. (1981). Behavioral treatment of a persistent post-traumatic startle response. Journal of Behavior Therapy and Experimental Psychiatry, 12, 321e324. Feinstein, A., & Dolan, R. (1991). Predictors of post-traumatic stress disorder following physical trauma: An examination of the stressor criterion. Psychological Medicine, 21, 85e91. Fischer, M., Barkley, R. A., Smallish, L., & Fletcher, K. (2007). Hyperactive children as young adults: Driving abilities, safe driving behavior, and adverse driving outcomes. Accident Analysis and Prevention, 39, 94e105. Fong, G., Frost, D., & Stansfeld, S. (2001). Road rage: A psychiatric phenomenon? Social Psychiatry and Psychiatric Epidemiology, 36, 277e286. Foot, H. C., & Chapman, A. J. (1982). Road safety and driver behaviour. Ergonomics, 25, 863e865. Fujita, G., & Nishida, Y. (2008). Association of objective measures of trauma exposure from motor vehicle accidents and posttraumatic stress symptoms. Journal of Traumatic Stress, 21, 425e429. Galovski, T., Blanchard, E. B., & Veazey, C. (2002). Intermittent explosive disorder and other psychiatric comorbidity among court-referred and self-referred aggressive drivers. Behaviour Research and Therapy, 40, 641e651. Galovski, T. E., Malta, L. S., & Blanchard, E. B. (2006). Road rage: Assessment and treatment of the angry, aggressive driver. Washington, DC: American Psychological Association. Glendon, A. I., Dorn, L., Matthews, G., Gulian, E., Davies, D. R., & Debney, L. M. (1993). Reliability of the Driving Behaviour Inventory. Ergonomics, 36, 719e726. Goldberg, L., & Gara, M. A. (1990). A typology of psychiatric reactions to motor vehicle accidents. Psychopathology, 23, 15e20. Green, M. M., McFarlane, A. C., Hunter, C. E., & Griggs, W. M. (1993). Undiagnosed post-traumatic stress disorder following motor vehicle accidents. Medical Journal of Australia, 159, 529e534. Groeger, J. A. (1997). Mood and driving: Is there an effect of affect? In T. Rothegatter, & E. Carbonell Vaya (Eds.), Traffic and transport psychology: Theory and application (pp. 335e342) Oxford: Pergamon. Gudmundsdottir, B., Beck, J. G., Coffey, S. F., Miller, L., & Palyo, S. A. (2004). Quality of life and post trauma symptomatology in motor
Chapter | 13
Mental Health and Driving
vehicle accident survivors: The mediating effects of depression and anxiety. Depression and Anxiety, 20, 187e189. Gulian, E., Glendon, A. I., Matthews, G., Davies, D. R., & Debney, L. M. (1988). Exploration of driver stress using self-reported data. In T. Rothengatter, & R. de Bruin (Eds.), Road user behaviour: Theory and research (pp. 342e347). Assen, The Netherlands: van Gorcum. Gulian, E., Glendon, A. I., Matthews, G., Davies, D. R., & Debney, L. M. (1990). The stress of driving: A diary study. Work & Stress, 4, 7e16. Gulian, E., Matthews, G., Glendon, A. I., Davies, D. R., & Debney, L. M. (1989). Dimensions of driver stress. Ergonomics, 32, 585e602. Hamanaka, S., Asukai, N., Kamijo, Y., Hatta, K., Kishimoto, J., & Miyaoka, H. (2006). Acute stress disorder and posttraumatic stress disorder symptoms among patients severely injured in motor vehicle accidents in Japan. General Hospital Psychiatry, 28, 234e241. Harvey, A. G., & Bryant, R. A. (1998). The relationship between acute stress disorder and posttraumatic stress disorder: A prospective evaluation of motor vehicle accident survivors. Journal of Consulting and Clinical Psychology, 66, 507e512. Heimstra, N. W. (1970). The effects of “stress fatigue” on performance in a simulated driving situation. Ergonomics, 13, 209e218. Heimstra, N. W., Ellingstad, V. S., & DeKock, A. R. (1967). Effects of operator mood on performance in a simulated driving task. Perceptual and Motor Skills, 25, 729e735. Hentschel, U., Bijleveld, C. C., Kiessling, M., & Hosemann, A. (1993). Stress-related psychophysiological reactions of truck drivers in relation to anxiety, defense, and situational factors. Accident Analysis and Prevention, 25, 115e121. Herda, C. A., Ehlers, A., & Roth, W. T. (1993). Diagnostic classification of driving phobia. Anxiety Disorders Practice Journal, 1, 9e16. Hickling, E. J., & Blanchard, E. B. (1992). Post-traumatic stress disorder and motor vehicle accidents. Journal of Anxiety Disorders, 6, 285e291. Hickling, E. J., & Blanchard, E. B. (Eds.). (1999). The international handbook of road traffic accidents and psychological trauma: Current understanding, treatment and law. New York: Pergamon. Hickling, E. J., Blanchard, E. B., Schwarz, S. P., & Silverman, D. J. (1992). Headaches and motor vehicle accidents: Results of the psychological treatment of post-traumatic headache. Headache Quarterly, 3, 285e289. Hickling, E. J., Blanchard, E. B., Silverman, D. J., & Schwarz, S. P. (1992). Motor vehicle accidents, headaches and post-traumatic stress disorder: Assessment findings in a consecutive series. Headache, 32, 147e151. Hickling, E. J., Kuhn, E. R., & Beck, J. G. (2008). Treatment of posttraumatic stress disorder consequent to motor vehicle collisions: Contributions from clinical science. In M. P. Duckworth, T. Iezzi, & W. T. O’Donohue (Eds.), Motor vehicle collisions: Medical, psychosocial, and legal consequences (pp. 365e387). San Diego: Academic Press. Hilton, M. F., Staddon, Z., Sheridan, J., & Whiteford, H. A. (2009). The impact of mental health symptoms on heavy goods vehicle drivers’ performance. Accident Analysis and Prevention, 41, 453e461. Ho, R., Davidson, G., Van Dyke, M., & Agar-Wilson, M. (2000). The impact of motor vehicle accidents on the psychological well-being of at-fault drivers and related passengers. Journal of Health Psychology, 5, 33e51. Hole, G. (2007). The psychology of driving. Totowa, NJ: Erlbaum. Hole, G. (2008). Predictors of motor vehicle collisions. In M. P. Duckworth, T. Iezzi, & W. T. O’Donohue (Eds.), Motor
175
vehicle collisions: Medical, psychosocial, and legal consequences (pp. 13e43). San Diego: Academic Press. Holeva, V., Tarrier, N., & Wells, A. (2001). Prevalence and predictors of acute stress disorder and PTSD following road traffic accidents: Thought control strategies and social support. Behavior Therapy, 32, 65e83. Horne,de L., D. J. (1993). Traumatic stress reactions to motor vehicle accidents. In J. P. Wilson, & B. Raphael (Eds.), International handbook of traumatic stress syndromes (pp. 499e506). New York: Plenum. Irish, L., Ostrowski, S. A., Fallon, W., Spoonster, E., van Dulmen, M., Sledjeski, E. M., & Delahunty, D. L. (2008). Trauma history characteristics and subsequent PTSD symptoms in motor vehicle accident victims. Journal of Traumatic Stress, 21, 377e384. Jaworowski, S. (1992). Traffic accident injuries of children: The need for prospective studies of psychiatric sequelae. Israeli Journal of Psychiatry and Related Sciences, 29, 174e184. Jerome, L., Segal, A., & Habinski, L. (2006). What we know about ADHD and driving risk: A literature review, meta-analysis and critique. Journal of the American Academy of Child and Adolescent Psychiatry, 15, 105e125. Jones, I. H., & Riley, W. T. (1987). The post-accident syndrome: Variations in the clinical picture. Australian and New Zealand Journal of Psychiatry, 21, 560e567. Jones, R. W., & Peterson, L. W. (1993). Post-traumatic stress disorder in a child following an automobile accident. Journal of Family Practice, 36, 223e225. Karl, A., Rabe, S., Zo¨llner, T., Maercker, A., & Stopa, L. (2009). Negative self-appraisals in treatment-seeking survivors of motor vehicle accidents. Journal of Anxiety Disorders, 23, 775e781. Kastrup, M., Dunpont, A., Bille, M., & Lund, H. (1977). Traffic accidents involving psychiatric patients: Description of the material and general results. Acta Psychologica Scandinavica, 55, 355e368. Kim, K., Nitz, L., Richardson, J., & Li, L. (1995). Personal and behavioural predictors of automobile crash and injury severity. Accident Analysis and Prevention, 27, 469e481. Knapp, S., & VandeCreek, L. (2009). Driving and operating other equipment: Legal and ethical issues. In J. L. Werth, Jr., E. R. Welfel, & G. A. H. Benjamin (Eds.), The duty to protect: Ethical, legal, and professional considerations for mental health professionals (pp. 127e140). Washington, DC: American Psychological Association. Koch, W. J., & Taylor, S. (1995). Assessment and treatment of motor vehicle accident victims. Cognitive and Behavioral Practice, 2, 327e342. Koren, D., Arnon, I., & Klein, E. (1999). Acute stress response and posttraumatic stress disorder in traffic accident victims: A one-year prospective, follow-up study. American Journal of Psychiatry, 156, 367e373. Koren, D., Arnon, I., Lavie, P., & Klein, E. (2002). Sleep complaints as early predictors of posttraumatic stress disorder: A 1-year prospective study of injured survivors of motor vehicle accidents. American Journal of Psychiatry, 159, 855e857. Kottenhoff, H. (1961). Interrelations of steering skill measures in neurotic and other patients. Perceptual and Motor Skills, 12, 289e290. Kuch, K. (1997). Accident phobia. In G. C. L. Davey (Ed.), Phobias: A handbook of theory, research and treatment (pp. 153e162). Chichester, UK: Wiley.
176
Kuch, K., Cox, B. J., & Direnfeld, D. M. (1995). A brief self-rating scale for PTSD after road vehicle accident. Journal of Anxiety Disorders, 9, 503e514. Kuch, K., Cox, B. J., & Evans, R. J. (1996). Posttraumatic stress disorder and motor vehicle accidents: A multidisciplinary overview. Canadian Journal of Psychiatry, 41, 429e434. Kuch, K., Cox, B. J., Evans, R., & Shulman, I. (1994). Phobias, panic, and pain in 55 survivors of road vehicle accidents. Journal of Anxiety Disorders, 8, 181e187. Kuch, K., Evans, R. J., & Mueller-Busch, C. (1993). Accidents, anxiety and chronic pain. The Pain Clinic, 6, 3e7. Kuch, K., Swinson, R. P., & Kirby, M. (1985). Post-traumatic stress disorder after car accidents. Canadian Journal of Psychiatry, 30, 426e427. Kupchik, M., Strous, R. D., Erez, R., Gonen, N., Weizman, A., & Spivak, B. (2007). Demographic and clinical characteristics of motor vehicle accident victims in the community general health outpatient clinic: A comparison of PTSD and non-PTSD subjects. Depression and Anxiety, 24, 244e250. Lajunen, T., & Parker, D. (2001). Are aggressive people aggressive drivers? A study of the relationship between self-reported general aggressiveness, driver anger and aggressive driving. Accident Analysis and Prevention, 33, 243e255. Lajunen, T., Parker, D., & Stradling, S. G. (1998). Dimensions of driver anger, aggressive and highway code violations and their mediation by safety orientation in UK drivers. Transportation Research Part F, 1, 107e121. Land Transport Safety Authority. (2002). Medical aspects of fitness to drive. Wellington, New Zealand. Landolt, M. A., Vollrath, M. E., Gnehm, H. E., & Sennhauser, F. H. (2009). Post-traumatic stress impacts on quality of life in children after road traffic accidents: Prospective study. Australian and New Zealand Journal of Psychiatry, 43, 746e753. Landolt, M. A., Vollrath, M., Timm, K., Gnehm, H. E., & Sennhauser, F. H. (2005). Predicting posttraumatic stress symptoms in children after road traffic accidents. Journal of the American Academy of Child and Adolescent Psychiatry, 44, 1276e1283. Lapham, S. C., de Baca, J. C., McMillan, G. P., & Lapidus, J. (2006). Psychiatric disorders in a sample of repeat impaired-driving offenders. Journal of Studies on Alcohol, 67, 707e713. Lapham, S. C., Smith, E., de Baca, J. C., Chang, I., Skipper, B. J., Baum, G., et al. (2001). Prevalence of psychiatric disorders among persons convicted of driving while impaired. Archives of General Psychiatry, 58, 943e949. Levelt, P. B. M. (2003). Field study of emotions in traffic (SWOV Report No. R-2003-08). Leidschendam, The Netherlands: SWOV. Little, A. D. (1970). The state of the art of traffic safety: A comprehensive review of existing information. New York: Praeger. Loo, R. (1979). Role of primary personality factors in the perception of traffic signs and driver violations and accidents. Accident Analysis and Prevention, 11, 125e127. Malt, U. (1988). The long-term psychiatric consequences of accidental injury: A longitudinal study of 107 adults. British Journal of Psychiatry, 153, 810e818. Malt, U. F., Hfivik, B., & Blikra, G. (1993). Psychosocial consequences of road accidents. European Psychiatry, 8, 227e228. Malta, L. S., Blanchard, E. B., & Freidenberg, B. M. (2005). Psychiatric and behavioral problems in aggressive drivers. Behaviour Research and Therapy, 43, 1467e1484.
PART | II
Key Variables to Understand in Traffic Psychology
Matsuoka, Y., Nishi, D., Nakajima, S., Kim, Y., Homma, M., & Otomo, Y. (2008). Incidence and prediction of psychiatric morbidity after a motor vehicle accident in Japan: The Tachikawa cohort of motor vehicle accident study. Critical Care Medicine, 36, 74e80. Matthews, G. (2001). A transactional model of driver stress. In P. A. Hancock, & P. A. Desmond (Eds.), Stress, workload, and fatigue (pp. 133e163). Mahwah, NJ: Erlbaum. Matthews, G., Desmond, P. A., Joyner, L., Carcary, B., & Gilliland, K. (1997). A comprehensive questionnaire measure of driver stress and affect. In T. Rothengatter, & E. C. Vaya (Eds.), Traffic and transport psychology: Theory and application (pp. 317e324). Amsterdam: Pergamon. Matthews, G., Dorn, L., Hoyes, T. W., Davies, D. R., Glendon, A. I., & Taylor, R. G. (1998). Driver stress and performance on a driving simulator. Human Factors, 40, 136e149. Matthews, G., Emo, A. K., & Funke, G. J. (2005). The transactional model of driver stress and fatigue and its implications for driver training. In L. Dorn (Ed.), Driving behaviour and training, Vol. 2 (pp. 273e285). Hampshire, UK: Ashgate. Matthews, G., Tsuda, A., Xin, G., & Ozeki, Y. (1999). Individual differences in driver stress vulnerability in a Japanese sample. Ergonomics, 42, 401e415. Mayou, R. (1997). The psychiatry of road traffic accidents. In M. Mitchell (Ed.), The aftermath of road accidents: Psychological, social and legal consequences of an everyday trauma (pp. 33e48). London: Routledge. Mayou, R. (1992). Psychiatric aspects of road traffic accidents. International Review of Psychiatry, 4, 45e54. Mayou, R., & Bryant, B. (2002). Outcome 3 years after a road traffic accident. Psychological Medicine, 32, 671e675. Mayou, R., Bryant, B., & Duthie, R. (1993). Psychiatric consequences of road traffic accidents. British Medical Journal, 307, 647e651. Mayou, R., Ehlers, A., & Bryant, B. (2002). Posttraumatic stress disorder after motor vehicle accidents: 3-year follow-up of a prospective longitudinal study. Behaviour Research and Therapy, 40, 665e675. Mayou, R., Tyndel, S., & Bryant, B. (1997). Long-term outcome of motor vehicle accident injury. Psychosomatic Medicine, 59, 578e584. McCaffrey, R. J., & Fairbank, J. A. (1985). Behavioral assessment and treatment of accident-related posttraumatic stress disorder: Two case studies. Behavior Therapy, 16, 406e416. McCutcheon, V. V., Heath, A. C., Edenberg, H. J., Grucza, R. A., Hesselbrock, V. M., Kramer, J. R., et al. (2009). Alcohol criteria endorsement and psychiatric and drug use disorders among DUI offenders: Greater severity among women and multiple offenders. Addictive Behaviors, 34, 432e439. McDermott, B. M., & Cvitanovich, A. (2000). Posttraumatic stress disorder and emotional problems in children following motor vehicle accidents: An extended case series. Australian and New Zealand Journal of Psychiatry, 34, 446e452. McDonald, A. S., & Davey, G. C. L. (1996). Psychiatric disorder and accidental injury. Clinical Psychology Review, 16, 105e127. McMillan, G. P., Timken, D. S., Lapidus, J., de Baca, J. C., Lapham, S. C., & McNeal, M. (2008). Underdiagnosis of comorbid mental illness in repeat DUI offenders mandated to treatment. Journal of Substance Abuse Treatment, 34, 320e325. Meiser-Stedman, R., Dalgleish, T., Glucksman, E., Yule, W., & Smith, P. (2009). Maladaptive cognitive appraisals mediate the evolution of posttraumatic stress reactions: A 6-month follow-up of child and
Chapter | 13
Mental Health and Driving
adolescent assault and motor vehicle accident survivors. Journal of Abnormal Psychology, 118, 778e787. Meiser-Stedman, R., Smith, P., Glucksman, E., Yule, W., & Dalgleish, T. (2008). The posttraumatic stress disorder diagnosis in preschool- and elementary school-age children exposed to motor vehicle accidents. American Journal of Psychiatry, 165, 1326e1337. Me´nard, I., Korner-Bitensky, N., Dobbs, B., Cascalenda, N., Beck, P. R., Ge´linas, I., et al. (2006). Canadian psychiatrists’ current attitudes, practices, and knowledge regarding fitness to drive in individuals with mental illness: A cross-Canada survey. Canadian Journal of Psychiatry, 51, 836e846. Mesken, J., Hagenzieker, M. P., Rothengatter, T., & de Waard, D. (2007). Frequency, determinants, and consequences of different drivers’ emotions: An on-the-road study using self-reports, (observed) behaviour, and physiology. Transportation Research Part F, 10, 458e475. Metzner, J. L., Dentino, A. N., Godard, S. L., Hay, D. P., Hay, L., & Linnoila, M. (1993). Impairment in driving and psychiatric illness. Neuropsychiatric Practice and Opinion, 5, 211e220. Mirza, K. A. H., Bhadrinath, B. R., Goodyer, I. M., & Gilmour, C. (1998). Post-traumatic stress disorder in children and adolescents following road traffic accidents. British Journal of Psychiatry, 172, 443e447. Mitchell, M. (Ed.). (1997). The aftermath of road accidents: Psychological, social and legal consequences of an everyday trauma. London: Routledge. Mitchell, M. (1999). Psychological distress in police officers attending serious and fatal road traffic accidents. In E. J. Hickling, & E. B. Blanchard (Eds.), The international handbook of road traffic accidents and psychological trauma: Current understanding, treatment and law (pp. 129e139). New York: Pergamon. Murray, J., Ehlers, A., & Mayou, R. A. (2002). Dissociation and post-traumatic stress disorder: Two prospective studies of road traffic accident survivors. British Journal of Psychiatry, 180, 363e368. Nada-Raja, S., Langley, J. D., McGee, R., Williams, S. M., Begg, D. J., & Reeder, A. I. (1997). Inattentive and hyperactive behaviors and driving offenses in adolescence. Journal of the American Academy of Child and Adolescent Psychiatry, 36, 515e522. Norris, F. H., Matthews, B. A., & Riad, J. K. (2000). Characterological, situational, and behavioral risk factors for motor vehicle accidents: A prospective examination. Accident Analysis and Prevention, 32, 505e515. Noyes, R. (1985). Motor vehicle accidents related to psychiatric impairment. Psychosomatics, 26, 569e580. Noyes, R., Hoenk, P. R., Kuperman, S., & Slymen, D. J. (1977). Depersonalisation in accident victims and psychiatric patients. Journal of Nervous and Mental Disease, 164, 401e407. O’Donnell, M. L., Creamer, M., & Ludwig, G. (2008). PTSD and associated mental health consequences of motor vehicle collisions. In M. P. Duckworth, T. Iezzi, & W. T. O’Donohue (Eds.), Motor vehicle collisions: Medical, psychosocial, and legal consequences (pp. 345e363). San Diego: Academic Press. O’Hanlon, J. F., Vermeeren, A., Uiterwijk, M. M., van Veggel, L. M., & Swijgman, H. F. (1995). Anxiolytics’ effects on the actual driving performance of patients and healthy volunteers in a standardized test. Neuropsychobiology, 31, 81e88. Palmer, R. S., Ball, S. A., Rounsaville, B. J., & O’Malley, S. S. (2007). Concurrent and predictive validity of drug use and psychiatric
177
diagnosis among first-time DWI offenders. Alcoholism: Clinical and Experimental Research, 31, 619e624. Payne, R. B., & Corley, T. J. (1994). Motivational effects of anxiety on psychomotor performance. Perceptual and Motor Skills, 79, 1507e1521. Pilowsky, I. (1985). Cryptotrauma and "accident neurosis.". British Journal of Psychiatry, 147, 310e311. Platt, J. J., & Husband, S. D. (1987). Posttraumatic stress disorder and the motor vehicle accident victim. American Journal of Forensic Psychology, 5, 39e42. Richards, T. L., Deffenbacher, J. L., Rose´n, L. A., Barkley, R. A., & Rodricks, T. (2006). Driving anger and driving behavior in adults with ADHD. Journal of Attention Disorders, 10, 54e64. ˚ berg, L. (1999). On the distinction between violations and Rimmo¨, P., & A errors: Sensation seeking associations. Transportation Research Part F, 2, 151e166. Rorsman, B., Hagnell, O., & Lanke, J. (1982). Violent death and mental disorders in the Landby study. Neuropsychobiology, 8, 233e240. Scha¨fer, I., Berkmann, C., Riedasser, P., & Schulte-Markwort, M. (2006). Posttraumatic syndromes in children and adolescents after road traffic accidents: A prospective cohort study. Psychopathology, 39, 159e164. Schuckit, M. A., & Gunderson, E. K. E. (1977). Accident and assault deaths in the United States Navy: Demography and preliminary interpretations. Military Medicine, 142, 607e610. Selling, L. S. (1940). The psychiatric findings in the cases of 500 traffic offenders and accident prone drivers. American Journal of Psychiatry, 97, 68e79. Selzer, M. L., Payne, C. E., Westervelt, F. H., & Quinn, J. (1967). Automobile accidents as an expression of psychopathology in an alcoholic population. Quarterly Journal of Studies on Alcohol, 28, 505e516. Shaffer, H. J., Nelson, S. E., LaPlante, D. A., LaBrie, R. A., Albanese, M., & Caro, G. (2007). The epidemiology of psychiatric disorders among repeat DUI offenders accepting a treatmentsentencing option. Journal of Consulting and Clinical Psychology, 75, 795e804. Shalev, A. Y., Freedman, S., Peri, T., Brandes, D., Sahar, T., Orr, S. P., et al. (1998). Prospective study of posttraumatic stress disorder and depression following trauma. American Journal of Psychiatry, 155, 630e637. Shaw, L., & Sichel, H. (1971). Accident proneness. Oxford: Pergamon. Shinar, D. (1978). Psychology on the road: The human factor in traffic safety. New York: Wiley. Shinar, D. (1998). Aggressive driving: The contribution of the drivers and the situation. Transportation Research Part F, 1, 137e160. Shoham, S. G., Rahav, G., Markovski, R., Chard, F., & Baruch, I. (1984). "Anxious" and "reckless" drivers. Deviant Behaviour, 5, 181e191. Silverstone, T. (1988). The influence of psychiatric disease and its treatment on driving performance. International Clinical Psychopharmacology, 3(Suppl. 1), 59e66. Smart, R. G., Ashbridge, M., Mann, R. E., & Adlaf, E. M. (2003). Psychiatric distress among road rage victims and perpetrators. Canadian Journal of Psychiatry, 48, 681e688. Smith, B., MacKenzie-Ross, S., & Scragg, P. (2007). Prevalence of poor psychological morbidity following a minor road traffic accident (RTA): The clinical implications of a prospective longitudinal study. Counselling Psychology Quarterly, 20, 149e155.
178
St. Germain, S. A., Kurtz, M. M., Pearlson, G. D., & Astur, R. S. (2004). Virtual driving in individuals with schizophrenia. Annual Review of CyberTherapy and Telemedicine, 2, 153e159. Stallard, P., Velleman, R., & Baldwin, S. (1998). Prospective study of post-traumatic stress disorder in children involved in road traffic accidents. British Medical Journal, 317, 1619e1623. Stoduto, G., Dill, P., Mann, R. E., Wells-Parker, E., Toneatto, T., & Shuggi, R. (2008). Examining the link between drinking-driving and depressed mood. Journal of Studies on Alcohol and Drugs, 69, 777e780. Sturms, L. M., van der Sluis, C., Stewart, R. E., Groothoff, J. W., ten Duis, H. J., & Eisma, W. H. (2005). A prospective study on paediatric traffic injuries: Health-related quality of life and post-traumatic stress. Clinical Rehabilitation, 19, 312e322. Sullman, M. (2006). Anger amongst New Zealand drivers. Transportation Research Part F, 9, 173e184. Sullman, M., Gras, M. E., Cunill, M., Planes, M., & Font-Mayolas, S. (2007). Driving anger in Spain. Personality and Individual Differences, 42, 701e713. Su¨mer, N. (2003). Personality and behavioral predictors of traffic accidents: Testing a contextual mediated model. Accident Analysis and Prevention, 35, 949e964. Taylor, A. H., & Dorn, L. (2006). Stress, fatigue, health, and risk of road traffic accidents among professional drivers: The contribution of physical inactivity. Annual Review of Public Health, 27, 371e391. Taylor, J. E. (2008). Driving phobia consequent to motor vehicle collisions. In M. P. Duckworth, T. Iezzi, & W. O’Donohue (Eds.), Motor vehicle collisions: Medical, psychosocial, and legal consequences (pp. 389e416). New York: Elsevier. Taylor, J. E., & Deane, F. P. (1999). Acquisition and severity of drivingrelated fears. Behaviour Research and Therapy, 37, 435e449. Taylor, J. E., & Deane, F. P. (2000). Comparison and characteristics of motor vehicle accident (MVA) and non-MVA driving fears. Journal of Anxiety Disorders, 14, 281e298. Taylor, J. E., Deane, F. P., & Podd, J. V. (1999). Stability of driving fear acquisition pathways over one year. Behaviour Research and Therapy, 37, 927e939. Taylor, J. E., Deane, F. P., & Podd, J. V. (2000). Determining the focus of driving fears. Journal of Anxiety Disorders, 14, 453e470. Taylor, J. E., Deane, F. P., & Podd, J. V. (2002). Driving-related fear: A review. Clinical Psychology Review, 22, 631e645. Taylor, J. E., Deane, F. P., & Podd, J. V. (2007). Driving fear and driving skills: Comparison between fearful and control samples using standardised on-road assessment. Behaviour Research and Therapy, 45, 805e818. Taylor, S., & Koch, W. J. (1995). Anxiety disorders due to motor vehicle accidents: Nature and treatment. Clinical Psychology Review, 15, 721e738.
PART | II
Key Variables to Understand in Traffic Psychology
Thompson, A., & McArdle, P. (1993). Psychiatric consequences of road traffic accidents: Children may be seriously affected. British Medical Journal, 307, 1282e1283. Tillman, W. A., & Hobbs, G. E. (1949). The accident-prone automobile driver: A study of the psychiatric and social background. American Journal of Psychiatry, 106, 321e331. Tsuang, M. T., Boor, M., & Fleming, J. A. (1985). Psychiatric aspects of traffic accidents. American Journal of Psychiatry, 142, 538e546. Ulleberg, P., & Rundmo, T. (2003). Personality, attitudes and risk perception as predictors of risky driving behavior among young drivers. Safety Science, 41, 427e443. Ursano, R. J., Fullerton, C. S., Epstein, R. S., Crowley, B., Kao, T., Vance, K., et al. (1999). Acute and chronic posttraumatic stress disorder in motor vehicle accident victims. American Journal of Psychiatry, 156, 589e595. Van Rooy, D. L., Rotton, J., & Burns, T. M. (2006). Convergent, discriminant, and predictive validity of aggressive driving inventories: They drive as they live. Aggressive Behavior, 32, 89e98. Walker, J. I. (1981). Posttraumatic stress disorder after a car accident. Postgraduate Medicine, 69, 82e86. Walklin, L. (1993). Instructional techniques and practice for driving instructors (2nd ed.). Cheltenham, UK: Stanley Thornes. Waller, J. A., & Turkel, H. W. (1966). Alcoholism and traffic deaths. New England Journal of Medicine, 275, 532e536. West, R., Elander, J., & French, D. (1993). Mild social deviance, Type-A behavior pattern and decision-making style as predictors of selfreported driving style and traffic risk. British Journal of Psychology, 84, 207e219. Wilson, R. J., & Jonah, B. A. (1988). The application of problem behaviour theory to the understanding of risky driving. Alcohol, Drugs and Driving, 4, 173e191. Wilson, T., & Greensmith, J. (1983). Multivariate analysis of the relationship between drivometer variables and drivers’ accident, sex, and exposure status. Human Factors, 25, 303e312. Wittman, L., Zehnder, D., Schredl, M., Jenni, O. G., & Landolt, M. A. (2010). Posttraumatic nightmares and psychopathology in children after road traffic accidents. Journal of Traumatic Stress, 23, 232e239. Yasan, A., Gu¨zel, A., Tamam, Y., & Ozkan, M. (2009). Predictive factors for acute stress disorder and posttraumatic stress disorder after motor vehicle accidents. Psychopathology, 42, 236e241. Yinon, Y., & Levian, E. (1988). Presence of other drivers as a determinant of traffic violations. In T. Rothengatter, & R. de Bruin (Eds.), Road user behaviour: Theory and research (pp. 274e278). Assen: The Netherlands: van Gorcum. Zuckerman, M., & Neeb, M. (1980). Demographic influences in sensation seeking and expressions of sensation seeking in religion, smoking and driving habits. Personality and Individual Differences, 1, 197e206.
Chapter 14
Person and Environment: Traffic Culture ¨ zkan and Timo Lajunen Tu¨rker O Middle East Technical University, Ankara, Turkey
1. PERSON AND ENVIRONMENT: BEHAVIOR AND ACCIDENTS Behavior is a result of a contribution of the person, the situation or environment, and some probabilistic interactive function of person and environment (Lewin, 1952, p. 25). The person is labeled as a human factor component, whereas the situation and/or environment are labeled as vehicle-related factors and road environment in traffic. A human (i.e., road user) is also embedded in a complex multilevel sociocultural and technical environment of traffic. Any outcome, such as an accident, is therefore a result of the contribution of human factor (i.e., road user), environment, and the probabilistic interaction of human ¨ zkan, 2006). factor and the environment (O
1.1. Accident Causation: Perspectives, Theories, and Periods The perspectives, theories, and periods of human error and/ or accident causation have actually evolved systematically throughout the years. Salmon, Lenne, Stanton, Jenkins, and Walker (2010) stated that human factor (error) models can basically be categorized as either person models (e.g., the generic error modeling system by Reason (1990)), focusing on the errors made at an individual operator (e.g., driver) level, or system models (e.g., the Swiss cheese model by Reason (1990)), focusing on the interaction between wider systematic failures and errors made by an individual operator. Elvik (1996) described accident theories that have been proposed to explain road accidents and presented them chronologically as random events (1900e1920), accident proneness (1918e1955), causal theory (1940e1960), systems theory (1955e1980), and behavioral theory (1978e2000). According to Elvik, the theory of accidents as random events and accident proneness theory were designed to explain why some people have more accidents than othersdthat is, their objective was to explain variation in the number of accidents within a certain group (or even “innate characteristics”). Causal accident theory was Handbook of Traffic Psychology. DOI: 10.1016/B978-0-12-381984-0.10014-1 Copyright Ó 2011 Elsevier Inc. All rights reserved.
developed to identify the real causes of accidents by probing the events leading to each accident in detail (e.g., in-depth accident analysis). Systems theory, on the other hand, takes the total number of accidents in a system as the starting point of its explanatory efforts. Systems theory proposed that accidents are the result of maladjustments in the interaction between the components of complex systems. Behaviorally oriented accident theories have once more focused on individual road user behavior as a critical determinant of accident occurrence. The basic idea of these theories is that human risk assessment and human risk acceptance are very important determinants of the actual number of accidents that occur during an activity. Similarly, it has been proposed (i.e., by risk homeostasis theory) that every society has the number of accidents it wants to have, and the only way to permanently lower this number is to change the target level of risk (or the desired level of safety; e.g., the number of accidents, injuries, fatalities can be tolerated by the society including decision makers and public). In summary, Elvik stated that (1) all accident theories that have been proposed contain an element of truth, (2) none of the theories tell the whole truth, and (3) almost all theories have been proposed as means of reducing accidents rather than out of intellectual curiosity. Hale and Hovden (1998) described the three ages of safety management as an expansion of perspectives on accident phenomena by emphasizing their supplementary characteristics. The first period was mainly associated with technical measures, whereas the second one focused on behavioral factors and individual behavior. The third period was influenced by ergonomics and later merged with sociotechnical approaches (Hovden, Albrechtsen, & Herrera, 2010). Wiegmann, von Thaden, and Gibbons (2007) claimed that recent years have witnessed the development of a fourth stage, the “safety culture period.” Operators are performing their duties or interacting with technology as coordinated teams embedded within a particular culture (e.g., organizational culture). In this chapter, we propose a framework as a product of intellectual curiosity to “fight” accidents. It is hoped that this framework will also contain an element of truth of 179
180
accident occurrence in the “whole truth” of accident causation. In addition, we aim to merge person (i.e., the role of behavioral factors in traffic accidents) and environment perspectives (i.e., the structure of the complex multilevel sociocultural and technical environment of traffic and its goals and mechanisms) in “the fourth age of safety” (i.e., “traffic safety culture”).
1.2. Behavioral Factors in Accidents: Driver Behaviors and Performance Most road traffic accidents can be directly attributed to behavioral factors as a sole or a contributory factor (Lewin, 1982). Behavioral factors in driving can be investigated under two separate components: driver behavior/style and performance/skills (Elander, West, & French, 1993). Driver behavior refers to the ways drivers choose to drive or habitually drive, including the choice of driving speed, habitual level of general attentiveness, and gap acceptance (Elander et al., 1993). In other words, it explains what drivers usually “do.” Driver performance includes information processing and motor and safety skills, which improve with practice and training (i.e., with driving experience). In other words, it explains what drivers “can” do. Reason, Manstead, Stradling, Baxter, and Campbell (1990) classified driver behaviors into errors, violations, slips, and lapses. They defined errors as “the failure of planned actions to achieve their intended consequences” and violations as “deliberate deviations from those practices believed necessary to maintain the safe operation of a potentially hazardous system” (p. 1316). Reason et al. (1990) also identified a third Driver Behavior Questionnaire (DBQ) factor, which they named “slips and lapses.” This factor includes attention and memory failures, which can cause embarrassment but are unlikely to have an impact on driving safety (Parker, West, Stradling, & Manstead, 1995). Lawton, Parker, Manstead, and Stradling (1997) extended the violations scale by adding more items and split it into two distinctive scales, ordinary violations and aggressive violations, according to the reason why drivers violate. However, the distinction between violations and errors is also supported by the fact that this two-factor solution was the most stable solution (among possible solutions with two to six factors) in a 3-year follow-up ¨ zkan, Lajunen, & Summala, 2006). study in Finland (O Finally, to extend the DBQ to an omnibus measure of driver ¨ zkan and Lajunen (2005) added to the DBQ behavior, O a scale for measuring positive driver behavior and obtained a clear three-factor solution: violations, errors, and positive behaviors. Spolander (1983) differentiated driver performance as technical (i.e., quick and fluent car control and traffic situation management) and defensive driving skills
PART | II
Key Variables to Understand in Traffic Psychology
(i.e., anticipatory accident avoidance skills). Hatakka, Keskinen, Laapotti, Katila, and Kiiski (1992) used an internal reference (i.e., the drivers were asked to assess their own abilities in different aspects of driving skills) based on a well-known finding that the majority of drivers assess themselves as better than average drivers in their technical and defensive skills (Na¨a¨ta¨nen & Summala, 1976). Later, Lajunen and Summala (1995) developed the Driver Skill Inventory (DSI) to further assess both general perceptual-motor performance and safety concerns and verified the two-factor structure of the DSI as perceptualmotor skills (i.e., perception, decision making, and motor control-related skills) and safety skills (i.e., anticipatory accident avoidance skills). A consistent factor structure and high reliability of the DSI were obtained in different studies across countries (Lajunen, Corry, Summala, & Hartley, ¨ zkan, Lajunen, Chliaoutakis, Parker, & Summala, 1998; O 2006a). It is well-known that both the components of the human factor in driving (i.e., driver behavior and skills) are associated with different traffic outcomes (e.g., offenses, speeding, and accidents) (Lajunen & Summala, 1995; Reason et al., 1990). Thus, the following section describes a complex multilevel sociocultural and technical environment of traffic in which behavioral factors are embedded.
1.3. Structure of the Multilevel Sociocultural and Technical Environment of Traffic 1.3.1. Level 1: Micro LeveldIndividual Level Characteristics of Behavioral Factors in Driving Driver behaviors and performance can be assumed to reflect many drivers’ individual characteristics, such as personality, attitudes, motives or “extramotives,” and perceptualmotor and information-processing capacities (Elander et al., 1993; Groeger, 2000; Na¨a¨ta¨nen & Summala, 1976). Here, some individual-level characteristics (i.e., age, sex, and cognitive process and/or biases) are presented as examples of critical behavioral factors in driving. Sex and age are directly linked to driver behaviors, performance, and accident liability. Young people are more involved in accidents in virtually every country, and the majority of these drivers are young males (Blockey & Hartley, 1995; Doherty, Andrey, & MacGregor, 1998; Evans, 1991). In addition, men and young drivers tend to commit violations more frequently than women and older drivers. In contrast, women and older drivers committed more errors than males and young drivers (Reason et al., 1990). Road users also have to interact with each other and to take into account each other’s intentions and behaviors in order to drive safely. Thus, drivers’ cognitive processes (i.e., causal attributions) might be a source of their own and others’ risky driving behaviors, performance, and accident
Chapter | 14
Person and Environment: Traffic Culture
liability. Attribution refers to the process by which individuals arrive at casual explanations for their own and others’ behavior (Ross, 1977). Most studies that have examined attribution biases in traffic have investigated the false consensus bias and actoreobserver effect (Baxter, Macrae, Manstead, Stradling, & Parker, 1990; Bjo¨rklund, 2005; Manstead, Parker, Stradling, Reason, & Baxter, 1992). False consensus refers to the tendency of persons “to see their own behavioral choices and judgments as relatively common and appropriate to existing circumstances while viewing alternative responses as uncommon, deviant, or inappropriate” (Ross, Greene, & House, 1977, p. 280). Manstead et al. (1992), for example, found that compared to drivers who did not commit specific violations and errors, drivers who committed these driver behaviors perceived these behaviors as being committed by a higher proportion of drivers than they were in reality. The actoreobserver effect refers to a “pervasive tendency for actors to attribute their actions to situational requirements, whereas observers tend to attribute the same actions to stable personal dispositions” (Jones & Nisbett, 1972, p. 80). When reporting causes for close following and running traffic lights, for instance, drivers attribute their own violations to situational factors and others’ violations to their personal dispositions (Baxter et al., 1990). Based on the literature, it can be assumed that age, sex, and cognitive process and/or biases are “universal” individual-level factors influencing driving behavior and performance and accident involvement.
1.3.2. Level 2: Meso LeveldOrganizational/ Company and Group/Community Level Factors 1.3.2.1. Organizational/Company Level Factors Compared to nonprofessional drivers, professional drivers’ driving is a less self-paced task. Nonprofessional drivers can principally determine the difficulties and risk level of their driving (Caird & Kline, 2004). They can choose the mode of transportation, time of travel, and target speed. On the other hand, many factors (e.g., time schedule and working shifts and hours) can increase professional drivers’ task demands. In addition, other factors, such as a company’s culture and/or climate including safety policy and practices (Caird & Kline, 2004), seem to largely determine how, why, when, and where they drive. This might clearly indicate the importance of the role of organizational culture and/or climate in professional drivers’ driving. Organizations are complex systems with different values, principles, attitudes, and viewpoints (Arnold, 2005). As a component of this complex system, organizational climate can be defined as “a summary of molar perceptions that employees share about their work environments”
181
(Zohar, 1980, p. 96). Those perceptions are thought to have a psychological utility in serving as a frame of reference for ¨ z, O ¨ zkan, guiding appropriate and adaptive task behaviors. O and Lajunen (2010) used a general organizational culture scale (i.e., Hofstede’s organizational culture scale) to investigate the relationship between organizational culture and/or climate and driver behaviors (i.e., errors, violations, and positive behaviors) among professional drivers. It was found that the highest number of violations was reported when both a low score for work orientation (i.e., low organizational importance for the work being done, rules and regulations, etc.) and a low score for employee consideration (i.e., the employees are given less consideration for their presence in and adaptation to the organization, etc.) were reported. In contrast, the lowest number of violations was reported when both work orientation (i.e., high organizational importance on the work being done, rules and regulations, etc.) and employee consideration scores were high (i.e., the employees are given more consideration for their presence in and adaptation to the organization, etc.). ¨ z, O ¨ zkan, and Lajunen (under review) developed the O Transportation Companies’ Climate Scale, which yielded three factorsdgeneral safety management, specific practices and precautions, and work and time pressure. It was found that drivers with high scores (a low level of pressure) of work and time pressure reported significantly lower frequencies of errors and violations than did drivers with low scores of work orientation. Drivers with high scores of general safety management reported significantly higher safety skills compared to drivers with low scores of general safety management. There was no main effect of any organizational climate dimensions on either the positive driver behaviors scale of the DBQ or perceptual-motor skills dimension of the DSI. Logistic regression analysis revealed a significant relationship between work and time pressure and accident involvement. Therefore, it can be assumed that organizational factors influence especially professionals’ driving behavior and performance and accident involvement, which in turn influence other road users’ driving. 1.3.2.2. Group/Community Level Factors The boundaries of the system become more open and relatively ill-defined as it is considered at higher levels (i.e., community/group level). Also, the effect of the group/ community level on an individual driver’s behaviors and performance might be getting narrower in scope and magnitude. In other words, road users are continuously interacting with each other, and they may not be under supervision as in closed systems (i.e., transportation companies). On the other hand, different cities do appear to have distinct driving cultures, such as differences in overall accident rates (Allstate Insurance Company, 2006) and road rage behaviors (Prince Market Research, 2006). In
182
addition, in the same country and even in the same city, drivers from different driver groups (e.g., a truck driver versus a private car user or a young versus an old driver) might follow informal rules of their own group rather than formal rules in driving and therefore develop a different general driving style and pose different levels of accident ¨ zkan, 2002). risks (Su¨mer & O Bener et al. (2008), for example, found that four-wheel drivers committed more violations, errors, and lapses than small car users. Lapses were associated with accident involvement among four-wheel drivers, whereas both errors and aggression speeding were related to accident involvement among small car users. Four-wheel drivers also reported lower seat belt usage and higher speeding compared to small car users. Bener et al. found that four-wheel drivers were involved in nearly 40.3% of road traffic accidents. ¨ z, O ¨ zkan, and Lajunen (2009) investigated stress O reactions, speeding, number of penalties, and accident involvement among different driver groups (taxi drivers, minibus drivers, heavy vehicle drivers, and nonprofessional drivers). The results revealed differences between different driver groups in terms of both risky driving behaviors and stress reactions (aggression, dislike of driving, hazard monitoring, fatigue proneness, and thrill-seeking) in traffic. The nonprofessional drivers drove faster than the taxi, minibus, and heavy vehicle drivers on highways and faster than the heavy vehicle and minibus drivers on city roads. In addition, the minibus drivers reported more penalties than the heavy vehicle drivers. Moreover, the minibus drivers were more aggressive compared to the nonprofessional drivers. The nonprofessional drivers were better with hazard monitoring in traffic compared to the minibus and heavy vehicle drivers. Finally, the heavy vehicle drivers reported more fatigue proneness compared to the nonprofessional drivers. Aggression, dislike of driving, and hazard monitoring dimensions were also related to accident involvement, whereas dislike of driving and thrill-seeking dimensions were related to speeding on city roads.
1.3.3. Level 3: Macro Level: National Level Factors The same drivers can engage in different driver behaviors and display different performance and pose different accident risks in two different countries (Finland and Russia) with roughly the same climate but different traffic safety regulations and practices (Levia¨kangas, 1998) and public awareness and government policies (Svedung & Rasmussen, 1998). For example, Gaygisiz (2010) investigated the relationship between governance quality and road fatality rates in a sample of 46 countries. The Worldwide Governance Indicators (WGI) was used to measure six dimensions of governance. Voice and Accountability measures the extent to which a country’s citizens are able to
PART | II
Key Variables to Understand in Traffic Psychology
participate in selecting their government as well as freedom of expression, freedom of association, and free media. Political Stability and Absence of Violence refers to the likelihood that the government will be destabilized or overthrown by unconstitutional or violent means, including politically motivated violence and terrorism. Government Effectiveness is a measure of the quality of public services, the quality of the civil service and the degree of its independence from political pressures, the quality of policy formulation and implementation, and the credibility of the government’s commitment to such policies. Regulatory Quality measures the ability of the government to formulate and implement sound policies and regulations that permit and promote private sector development. Rule of Law is the extent to which agents have confidence in and abide by the rules of society, particularly the quality of contract enforcement, property rights, the police, and the courts, as well as the likelihood of crime and violence. Control of Corruption measures the extent to which public power is exercised for private gain, including both petty and grand forms of corruption, as well as “capture” of the state by elites and private interests. The correlations between these indexes and traffic fatality rates were -0.41, -0.48, -0.42, -0.51, -0.41, and -0.43, respectively. In other words, WGI and traffic fatalities are significantly associated with each other, and the better governance a country has, the lower traffic fatality rate tends to be. On the other hand, governance including laws and policies might force individuals to behave appropriately and safely in traffic, but it does not necessarily affect the way in which people think overall. Indeed, policies may often fail when they are not supported by the “upper” and “lower” levels of the system and the culture.
1.3.4. Level 4: Magna Level: Ecocultural Sociopolitical Level Factors The factors on the level of ecocultural sociopolitics, called exogenous variables in the traffic literature (Page, 2001), include the usual ecological components of a traffic culture, such as economic, demographic (e.g., population), ecologic (e.g., latitude) (Hofstede, 2001), and broader cultural factors (Levia¨kangas, 1998). These factors are highly correlated with each other (Hofstede, 2001) and cannot be modified by safety policies in the short term. They mostly have indirect, and rarely direct, effects on the level of mobility and safety by interacting with engineering and road user factors of everyday traffic in a country. Economic, societal, and cultural factors appear to be the most important variables in traffic safety (Gaudry & Lassarre, 2000). 1.3.4.1. Economy A high-income country can invest in its road infrastructure, maintenance of infrastructure, traffic safety
Chapter | 14
Person and Environment: Traffic Culture
work, vehicles, and driver education, whereas a lowincome country or country in an economic depression will invest less in traffic safety. The composition of the driver population may also change from a dominant majority of professional drivers to private car drivers in periods of economic boom (e.g., China; Zhang, Huang, Roetting, Wang, & Wei, 2006). Along with economic development, the male-dominant traffic society may change to one that is more balanced between females and males; that is, the proportion of male and female driver’s license holders changes, resulting in a higher number of female drivers (United Nations, 1997). The number of young, inexperienced drivers, however, is relatively high in high-income countries. Page (2001) indicated that an increase of 10% in the young population, with all other factors held constant, leads to an 8.3% increase in fatalities. During an economic boom, young adults have more money to spend on leisure activities such as driving. This increases the exposure and probability of accident involvement. New (and safer) car sales (Pelzman, 1975) and car ownership rates are also relatively high in high-income countries. According to the well-known Smeed’s law (Smeed, 1949), traffic casualties are related to the cube root of car ownership. It was evidenced (with data) in 20 different countries that the death rate per vehicle declined when ownership increased. In addition, Smeed’s law was valid for a variety of countries (e.g., in Great Britain from 1909 to 1973) over time and for the data from 62 countries (Adams, 1987, 1995). The economic system, on the other hand, influences the price mechanism (e.g., price of fuel), household consumption and vacation practices (e.g., holiday travel), modes of personal travel (e.g., home-to-work trips), and industrial activity for the transport of goods (Jaeger & Lassarre, 2000). It was found that factors such as the high occurrence of home-to-work trips and holiday travels, greater number of commercial vehicles per unit of work, wine consumption, and low price of fuel explain increases in both total mileage and accident risk. In the SARTRE 1 study conducted in October 1991 and June 1992 targeting major road safety concerns, it was found that the differentiation among drivers of the 15 European countries with regard to their attitudes and behaviors toward major road safety concerns (i.e., alcohol, speed, and seat belt use) was also partly associated with the economic prosperity of the individual countries (i.e., “safe” or “high-income” west/north vs. “dangerous” and “low-income” south) (SARTRE, 1998). ¨ zkan and Lajunen (2007) found that gross national O product (GNP) was the most important predictor for traffic safety in countries and the main reason for regional differences among countries in traffic safety. GNP per capita was negatively related to traffic
183
fatalities. GNP also correlated with both culture dimensions and values. 1.3.4.2. Culture Hofstede’s (2001) culture dimensions include inequality between people (“power distance”), the level of stress in a society related to an unknown future (“uncertainty avoidance”), the integration of individuals into primary groups (“individualism versus collectivism”), the division of emotional roles between males and females (“masculinity versus femininity”), and the time perspective of individuals (“long-term versus short-term orientation”). Schwartz values are based on three main concerns that all societies have to confront and solve. According to Schwartz (1999), the first concern, a society’s answer to the question of to what extent persons are either autonomous or embedded in their group, can be summarized by using three value types: “conservatism” (or embeddedness in Schwartz (2004) i.e., social order, “intellectual autonomy” (i.e., curiosity), and “affective autonomy” (i.e., pleasure). The second concern is to guarantee responsible behavior that will preserve the social fabric. Value types “hierarchy” (i.e., authority) and “egalitarianism” (i.e., equality) are the main solutions for preserving the social structure of the society. The third concern is the relationship between an individual and the natural and social environment. The relationship between human and environment can be based on two value types, “mastery” and “harmony.” In this dichotomy, mastery emphasizes a human’s wish to shape his or her environment according to his or her needs, whereas harmony refers to values in which protection of the environment is emphasized. Southern European countries generally score higher on uncertainty avoidance, power distance, collectivism, egalitarianism, and masculinity and are less conservative than northern European countries (Hofstede, 2001; Schwartz, 1992, 1999). Specifically, “dangerous” Greece scores the highest on uncertainty avoidance and mastery scores. “Dangerous” Turkey also has very high scores on uncertainty avoidance, power distance, conservatism, and hierarchy. “Safe” Great Britain and The Netherlands have very high scores on individualism, and Great Britain has a very low score on uncertainty avoidance. “Safe” Finland also has low scores on masculinity, power distance, and uncertainty avoidance. It was found that masculinity dimensions of a culture were positively related to high speed limits in 14 European countries (Hofstede, 2001). In addition, Hofstede reported that uncertainty avoidance and masculinity were positively related to traffic death rates in 1971 in 14 European countries, whereas individualism was negatively related to the accident rate. Drivers in individualistic cultures show a more calculative involvement in traffic (Hofstede, 2001),
184
PART | II
which leads to safer driving. After controlling the effect of GNP per capita, as in earlier studies (Hofstede, 2001), uncertainty avoidance was positively related to traffic fatalities. Partial correlation coefficients showed that conservatism correlated negatively and egalitarianism ¨ zkan & correlated positively with traffic fatalities (O Lajunen, 2007). These findings indicate that the role of economic, societal, and cultural factors should be taken into account to explain the regional differences in traffic safety among countries.
1.4. Summary of the Structure of the Multilevel Sociocultural and Technical Environment of Traffic As presented in Figure 14.1, a road user’s and/or country’s level of safety in traffic is mostly determined by how and to what extent external factors (ecocultural sociopolitical, national, group, organizational, and individual factors) influence either directly or indirectly internal factors (road users/behavioral factors, roads, and environment/engineering), which in turn affect exposure and accident risk. It is highly likely that factors such as geography or climate, which remain relatively constant over time and resist change (Evans, 2004), would have a more direct effect on engineering (e.g., roads and vehicles) than would road users. On the other hand, it is likely that climate (e.g., snow) A country’s environment (external factors)
Key Variables to Understand in Traffic Psychology
could reduce drivers’ exposure and behavior, particularly speed, which in turn might increase the number of accidents but lower the risk of severe injuries (Evans, 2004). However, external factors cannot be restricted to only environment-related factors (i.e., climate); in other words, other variables can be present at the ecocultural sociopolitical level (i.e., economy and culture) as well. For example, the same drivers can engage in different driver behaviors and display different performance and pose different accident risk in two different countries (Finland and Russia) with roughly the same climate but different traffic safety regulations and practices (Levia¨kangas, 1998) and public awareness and government policies (Svedung & Rasmussen, 1998). In the same country, and even in the same city, drivers from different driver groups (e.g., a truck driver versus a private car user or a young versus an old driver) might follow informal rules of their own group rather than formal rules in driving and, therefore, develop a different general driving style and pose ¨ zkan, 2002). different levels of accident risks (Su¨mer & O Organizational culture factorsdthat is, management or company policy (Svedung & Rasmussen, 1998)dmight be more important than formal traffic code and informal group code for professional drivers. In other words, drivers from the same driver group but from different companies, who even drive the same route and vehicles, might have different driver behaviors and performance and accident ¨ z et al., 2010). Driving is therefore to some extent risk (O
Traffic components (internal factors)
Eco-cultural-socio-political level *economy, climate, geography, demography, national culture, and characteristics
Road engineering/ infrastructure
National level *Government, authorities, traffic safety regulations, political climate, public awareness
Exposure
Group level *vehicle types, informal rules, identitites
Automotive engineering/ vehicles
Accidents and its consequences
Organizational/Company level *market and financial conditions, management, organizational safety culture
Individual level *age, sex, personality, attitudes, motives, perceptual-motor, and cognitive abilities
FIGURE 14.1
Road users *behaviour *performance
The framework of the multilevel sociocultural and technical environment of traffic. Source: O¨zkan (2006).
Chapter | 14
Person and Environment: Traffic Culture
a “forced-paced” task, and professional drivers’ risk can also be determined by their companies. Furthermore, when all other conditions and situations are constant, each individual driver might have a different general driving style and accident liability. Because driving is to some extent a “self-paced” task and drivers determine their risk by their own choices (Na¨a¨ta¨nen & Summala, 1976), individual factors such as “extra motives,” personality, sex, and age influence an individual driver’s behaviors, performance, and accident risk (Elander et al., 1993). It can be assumed, therefore, that accident risk and differences between road users and/or countries in traffic safety may be the result of how these internal and external factors are operating within levels and between levels in the whole system. Logically, this overall structure would probably work differently among countries (and even among individual road users). It is well-known that road traffic accidents are a major problem throughout in the world. However, regional differences in traffic safety between countries are considerable. In 2002, for example, the World Health Organization’s (WHO) Western Pacific Region and South-East Asia Region accounted for more than half of the absolute number of road traffic fatalities that occurred annually throughout the world. The WHO
FIGURE 14.2
185
African Region (including the Middle East) had the highest fatality rate with 28.3 per 100,000 population, which was closely followed by the low- and middle-income countries of the WHO Eastern Mediterranean Region with 26.4 fatalities per 100,000 population (WHO, 2004). The vast differences in traffic fatalities among countries are remarkable in the world in general and in Europe and its close neighbors (e.g., the Middle East) in particular. In the European Union, approximately 40,800 people were killed in traffic accidents in 2000, and an additional 11,600 people were killed in the accession countries (European Transport Safety Council, 2003). As presented in Figure 14.2, eastern/ southern (Mediterranean) Europe (e.g., Greece and Turkey) has the highest accident rates compared to northern/western Europe. In 2003, 7.6 Finns and Britons and 7.7 Dutch per 1 billion vehicle-kilometers were killed in traffic accidents, whereas the corresponding figures for Greeks and Turks were 26.7 and 73 in 2001, respectively (International Road Traffic and Accident Database, 2005). Traffic fatalities were reported to be much higher in Middle Eastern countries (i.e., Iran) than in European countries (i.e., Turkey) (Raoufi, 2003). It could therefore be hypothesized that the vast differences among countries in traffic culture and level of safety
Road fatalities in some European countries per 1 billion vehicle-kilometers on all roads in selected years. Source: O¨zkan (2006).
186
would be reflected in drivers’ driving behaviors. As ¨ zkan, Lajunen, Chliaoutakis, Parker, and hypothesized, O Summala (2006b)dby comparing British, Dutch, Finnish, Greek, Iranian, and Turkish driversdshowed that drivers in “safe” western/northern countries scored higher on ordinary violations, especially speeding on the motorway, whereas drivers in “dangerous” southern European/Middle Eastern countries had higher driving error and aggressive driving scores. It was also suggested that the higher level of aggressive driving and errors in southern European and Middle Eastern drivers was due to higher levels of conflict attributed to less developed infrastructure, less respect for traffic rules, and higher levels of driver stress. According to the conclusions, the higher frequency of speeding reported by drivers in “safe” countries reflected the level of enforcement. In addition, it was claimed that the concept of being a “safe driver” is culture dependent and, therefore, understood differently in different countries. On the other hand, based on the overall complex multilevel sociocultural and technical environment of traffic, the countries may be similar to some extend with regard to safety. However, the “software” or “traffic culture” defining the main goals, values, norms, practices, and mechanisms must then be taken into account.
2. TRAFFIC CULTURE: GOALS AND MECHANISMS Levia¨kangas (1998) labeled all the factors (probably all those shown in Figure 14.1) that directly and/or indirectly influence a country’s level of traffic safety as “traffic culture” (AAA Foundation for Traffic Safety, 2007). According to him, traffic culture is the sum of all factors that affect skills, attitudes, and behaviors of drivers as well as vehicles and infrastructure. However, the term traffic culture has not been conceptualized comprehensively. Therefore, this section uses “traffic culture” as a framework of reference for studying the goals and mechanisms of traffic culture. It is well-known that practices overwhelmingly aim to achieve the goals of safety (i.e., decreasing the number of accidents and near accidents) and promoting mobility (i.e., reaching the destination in terms of the amount of travel and trip time in traffic) (Evans, 2004). However, mobility and safety are often, but not always, in conflict. The primary goal of a traffic system in a country is mostly mobility, which should be achieved by minimizing the risk of the unwanted by-productdaccidents (Evans, 2004). In addition, some subgoals, such as environment-friendly, comfortable, cost-effective transportation, are becoming increasingly important to policy makers and the public. It can be assumed that traffic culture in a country or in a region is formed and maintained mostly by formal and informal rules, norms, and values, which are the center of
PART | II
Key Variables to Understand in Traffic Psychology
the mechanism of traffic culture. They define the acceptable and necessary road user behaviors and performance and choices of engineering practices. Whereas formal rules are mostly applied and enforced by authorities, including education, enforcement, and engineering, road users mostly share informal rules, norms, and values as a result of exposure and interaction with other road users. Exposuredthat is, the degree to which a driver exposes him- or herself to traffic and the probability of being involved in an accidentdis “a systematic process affecting the crash system” (Chapman, 1973) and, therefore, one of the main reasons for the overrepresentation of a particular driver group in accident statistics (Laapotti, 2003). In addition, exposure can be supposed to be the main quantitative (i.e., the amount of driving) and qualitative (i.e., why, when, where, with whom, and in what kind of weather and road conditions) predictor of driving and the of the interaction among internal and external factors, risky general driving style, and accident involvement (Laapotti, 2003). For instance, the average male driver drives more miles than does the average female driver (Stradling & Parker, 1996). Drivers who drive frequently, violate traffic rules more often than those who drive less frequently. They also tend to commit more aggressive driving behaviors than young female drivers and older drivers (Lawton et al., 1997). In addition, driving experience is associated with confidence in one’s own driving skills but negatively related to concern for safety (Lajunen & Summala, 1995). Also, the relationship between mileage and accidents seems not to be linear but, rather, a negatively accelerating curve, with a smaller increase in accident rate at a higher level of mileage (Maycock, Lockwood, & Lester, 1991). Other road users are studied as a source of information, communication, imitation, and as a reference group (Bjo¨rklund, 2005; Zaidel, 1992). Cultural and environmental factors define acceptable and “normal” behaviors, which in turn influence the definition of violations, not only simply in the strict legal sense (Manstead, 1998) but also informally (Bjo¨rklund, 2005). In addition, they might influence appraisals of the intentions and behaviors of other road users, which in turn could influence attribution of intentionality, controllability, and responsibility of driver behaviors and potential reactions (i.e., retaliation). Moreover, these factors might lead to different evaluations of risk, one’s own and other’s performance and behaviors across countries, and interpersonal conflicts in traffic ¨ zkan, Lajunen, Parker, (Bjo¨rklund, 2005). For example, O Su¨mer, and Summala (2010) found that “others” was a critical component of safe driving among British, Dutch, Finnish, and Turkish drivers. It was found that symmetric interpersonal aggression between aggressive warnings and hostile aggression and revenge factors of “self” and “others” created a serious risk of road accident involvement for drivers in every country except among British male and
Chapter | 14
Person and Environment: Traffic Culture
Finnish female drivers. The statistically symmetric interaction between aggressive warnings and hostile aggression and revenge factors also indicated that aggressive warnings might have a potential to release anger and escalate aggression both “within drivers” and “between drivers.” Their study also showed that both situational and cultural factors are important for understanding the role of anger and aggression in driving as well as the symmetric interaction between “self” and “others.” Traffic culture is also a result of both the larger cultural heritage and the current state of the environment, including the economy and political climate (Levia¨kangas, 1998). Similar to the culture of a country (Hofstede, 2001), ecological (e.g., economy and geography), societal, and cultural factors seem to lead to the development and pattern maintenance of institutions or political bodies (e.g., legislation, engineering, and educational systems). Once these institutions are established, the societal norms and values and formal and informal rules will be reinforced, and the boundaries of road user behaviors will be determined. Thus, the traffic culture of a country is formed and continues based on the functions of the large number of factors and practices at the multilevels or layers. In summary, traffic culture of a country can be redefined as the sum of all external factors (ecocultural sociopolitical, national, group, organizational, and individual factors) and practices (e.g., education, enforcement, engineering, economy, and exposure) for the main goals of mobility and safety to cope with internal factors (road users, roads, and engineering) of traffic.
2.1. Traffic Safety Culture The conceptualization of traffic culture seems to be broad and sometimes equivalent to the traffic system as a whole. Traffic culture and traffic system are, in fact, mutually inclusive and the main contributors to the differences in traffic safety between countries. However, they are based on different principles. Traffic (or “hardware”) is mainly based on tangible things such as roads, traffic signs, infrastructure, vehicles, tools, and equipment. On the other hand, traffic culture is defined as the sum of all external factors and practices for mainly the goals of mobility and safety to cope with internal factors of traffic. In addition, basic assumptions, formal and informal rules, values, norms, perceptions, and attitudes are the center of the mechanism of traffic culture; in other words, the “software” of the traffic culture is traffic safety culture and/or climate. The safety culture concept emerged after the Chernobyl accident and several reports prepared by the International Atomic Energy Agency. The concept of safety culture was defined for the first time by the Advisory Committee on the Safety of Nuclear Installations (International Nuclear Safety Advisory Group, 1991) as follows: “Safety culture is
187
the product of individual and group values, attitudes, competencies, and patterns of behavior that determine the commitment to, and the style and proficiency of, an organization’s health and safety programs.” Zohar (1980) defined the safety climate as “a summary of molar perceptions that employees share about their work environments (safety climate).” The development of these concepts seems to be successive rather than parallel: “The minor substantive differences between culture and climate may prove to be more apparent than real” (Glick, 1985). As presented in Table 14.1, however, it might be difficult to replace them with each other, and it might be difficult to separate these concepts in practicedeven the concepts are very novel in the traffic literature (Antonsen, 2009; Guldenmund, 2000; Wiegmann et al., 2007). “Traffic culture and/or climate” and “traffic safety culture and/or climate” have remained mostly a notion in the literature without attempts to measure it empirically. One special problem related to measuring traffic culture can be seen in studies measuring “safety culture”: The “traffic culture” seems to largely overlap with the concept of “traffic climate,” and sometimes these concepts are used interchangeably. However, they are different concepts while being mutually inclusive (Antonsen, 2009; Guldenmund, 2000). Wiegmann and colleagues (2007), for example,
TABLE 14.1 Features and Differences between Culture and Climate Culture
Climate
Beliefs and values about people, work, the organization, and the community that are shared by most members within the organization
Common characteristics of behavior and expression of feelings by organizational members
More qualitative approach
More quantitative approach
Research focused on the dynamic process, creating and shaping culture An enduring aspect of the organization with traitlike properties
Reflection and manifestation of cultural assumptions
The perception of a coherence of numerous processes by all the members in the organization
The underlying meaning given to this coherence
Not easily changed and relatively stable
Shaped by interactions
Multiple dimensionality; holistic, mutual, and reciprocal; and shared by people
Tension between reductionism and holistic
Exists at a higher level of abstraction than climate
188
PART | II
gave 13 and 12 example definitions for safety culture and safety climate, respectively. However, the definition of traffic safety culture/climate remained unexplored. As in the literature on safety culture and climate, traffic safety culture can thus be defined as the product of exposure and interaction of road users and the set of formal and informal rules, norms, basic assumptions, attitudes, values, habits, and perceptions in relation to safety and/or to conditions considered risky, dangerous, or injuries. As presented in Table 14.1, safety climate will then be the surface features of the safety culture (Mearns, Flin, Gordon, & Fleming, 1998) or the temporal state measure of culture (Cheyne, Cox, Oliver, & Thomas, 1998). In addition, climate exists at a lower level of abstraction than culture (Guldenmund, 2000) and provides a limited set of variables that can be operationalized and measured (Cox & Flin, 1998). Thus, climate research is conducted mostly using quantitative methods (e.g., questionnaires) dealing with the members’ perceptions and practices and how these practices and perceptions are categorized into the analytical dimensions defined by the researchers (Guldenmund, 2000). ¨ zkan and Lajunen (under review) used and defined O “traffic climate” as preferred metric and the manifestation of traffic culture discerned from the road users’ attitudes and perceptions at a given point in time (Cox & Flin, 1998). Traffic climate is therefore defined as the road users’ (e.g., drivers’) attitudes and perceptions of the traffic in a context ¨ zkan & Lajunen, (e.g., country) at a given point in time (O ¨ under review). Ozkan, Lajunen, Walle´n Warner, and
Key Variables to Understand in Traffic Psychology
Tzamalouka (2006) found that compared to Swedish and Finnish drivers, Turks and Greeks perceived their traffic climate to be more dangerous, dynamic, fast, dense, unpredictable, chaotic, and free flowing, thus requiring more patience. In contrast, compared to Turks and Greeks, Swedes and Finns perceived their traffic climate to be more harmonious, safe, functional, enforced (including the use of preventive measures), dependent on mutual consideration, planned, and mobile. It can be claimed that the vast differences among countries (i.e., Greece, Finland, Sweden, and Turkey) in traffic safety also reflect their drivers’ perceptions of the traffic climate. The set of formal and informal rules, norms, basic assumptions, attitudes, values, habits, and perceptions can operate in different layers of traffic safety culture and climate. For example, there are some basic assumptions, core values and norms, and goals that are underlying factors of traffic safety culture at each level of the traffic culture. In addition, there are some espoused values and artifacts (e.g., attitudes, habits, and perceptions) at the upper layers of traffic safety climate for each level of the traffic culture (for a multilevel model of culture including basic assumptions, espoused values, and artifacts, see Schein, 1992). It is desirable in relation to safety that these layers operate consistently and harmoniously to minimize the exposure of road users and, sometimes, members of the public to conditions considered dangerous or to injuries at each level of traffic culture. In addition, levels of traffic culture should operate consistently and harmoniously as
FIGURE 14.3 “Swiss cheese” model. Source: Adapted from Reason (1990).
Chapter | 14
Person and Environment: Traffic Culture
well. It can be assumed that any simultaneous latent or active failures either within layers and/or levels or between layers and/or levels could result in risky or dangerous acts or injuries in due course. Like the “Swiss cheese” model (Figure 14.3), traffic culture and traffic safety culture/climate focus on the interaction between latent and active conditions/failures within and between layers (i.e., traffic culture and traffic climate) and/or levels (i.e., ecocultural, sociopolitical, national, group, organizational, and individual) and unsafe acts and their contribution to accidents. Safety is therefore the responsibility of actors at all layers and/or levels of the system, especially in the absence of “defense barriers” (e.g., enforcement). In addition to the integrative perspective, differentiation and fragmentation perspectives can also be applied in the open system (i.e., traffic; Antonsen, 2009). Salmon and colleagues (2010) stated that The Netherlands’ Sustainable Safety approach (Wegman, Aarts, & Bax, 2008), for example, highlights the fact that the fallibility of human operators is underpinned by the assumption that the responsibility for safety is shared among actors across all levels of the complex sociotechnical system (e.g., regulators, policy makers, designers, line managers, manufacturers, supervisors, and front-line operators). It is not just the responsibility of front-line operators (i.e., road
189
users) alone. In contrast to a closed system (e.g., factories), the levels of traffic and traffic culture influence and are influenced by each other and are underpinned by traffic safety culture/climate (e.g., basic assumptions, espoused values, and artifacts). Then, they are reflected in individual road user behavior, which in turn influences the likelihood of being in a traffic accident and thereby affecting the content and development of the other levels (Figure 14.4).
3. CONCLUSION An accident has been defined either as independent or a combined outcome of behavioral factors, vehicle-related factors, and road environment in the literature. However, as presented in this chapter, behavioral factors, vehicle-related factors, and road environment are actually embedded in a larger system (see Figure 14.1). An accident is therefore either an independent or a combined outcome of internal factors of the multilevel sociocultural and technical environment of traffic. Briefly, a road users’ and/or country’s level of safety in traffic is mostly determined by how and to what extent external factors (i.e., individual, organizational, group/community, national, and ecocultural sociopolitical levels) influence either directly or indirectly internal factors (i.e., behavioral factors, vehicle-related FIGURE 14.4 Multilevel model of “traffic safety culture and climate”
190
factors, and road environment), which in turn affect exposure, risk, and accident involvement. All these factors are interactive and (see Figures 14.3 and 14.4) operate simultaneously in daily life. Each component of traffic culture (see Figure 14.1) has its own weight on safety (or unintentional injuries), and that weight depends on its relevance and importance in time and space of an event. Compared to nonprofessional drivers, many factors (e.g., time schedule, working shifts, and hours) can increase professional drivers’ task demands. A company’s culture and/or climate, including safety policy and practices, can also largely determine how, why, when, and where they drive. It is also possible that the interaction among these levels will influence traffic safety. Gaygisiz (2010) found that among the low WGI countries, increasing hierarchy was positively related to road fatalities, whereas in the high WGI countries hierarchy was not related to road fatalities. Similarly, among the low WGI countries, increasing mastery was positively related to road fatalities, whereas the effect was smaller in the moderate and high WGI countries. This chapter provided the first definitions of traffic culture/climate, the multilevel approach, and a comprehensive framework for traffic safety. Note, however, that traffic is one of the most open systems of all. Therefore, applying the traffic culture framework in such a system is not as easy as applying safety culture in closed ones (e.g., industrial companies). In addition, theoretically and potentially, any road user of the system can trigger a factor within one of the levels for another road user embedded in the same system. In other words, there is a shared, continuous, active and interactive environment for all road users. It should also be noted that it is not easy to change some of these factors (e.g., societal and cultural factors). On the contrary, they are external factors to the traffic system and, therefore, it is very likely that internal factors (i.e., engineering and road user factors) might buffer or facilitate their effects on traffic safety. It seems that in addition to the traditional three E’s in injury prevention (i.e., engineering, enforcement, and education), economy and exposure (including interaction with other road users) should be added as the fourth and fifth E’s (see Chapter 1). For example, economic incentives should be used to encourage injury prevention (e.g., monetary incentives for purchasing safety equipment) and structural modifications. Economic resources might also be efficiently spent not only on traffic, road, and automotive engineering but also on education, enforcement, and emergency services to develop a more predictable, certain, interpretable and preventive traffic system. In addition, GNP also correlates with both culture dimensions and values, which influences their relationship with unintentional injuries. It is likely that economic incentives could be used as a tool for developing more
PART | II
Key Variables to Understand in Traffic Psychology
safety-minded cultures and values. Moreover, up-to-date highway codes and applicable legislative interventions (i.e., enforcement) and education should also target driver behavior and performance of everyday traffic. Furthermore, because the balance between safety and mobility in traffic is important and initiated by the decisions of policy makers and planners, goals for traffic should be carefully evaluated at each level of the traffic system for all road users. The priority goal also has the potential to determine the dominant group in traffic. For example, high-mobility priority might put too much weight on drivers and cars compared to pedestrians, motorcyclists, and others. Despite having some principles, assumptions, and frames for road safety strategies, the model may have a number of flaws that limit its utility somewhat in terms of practical road transport applications. First, there are no structured road transport-specific methodologies associated with the model. Valid data collection and analysis methods for road transport are also necessary to test the model’s assumptions. Second, because the model is generic, it also lacks a clear definition of the different failures residing at each of the levels and the role of the components of safety culture and climate within the model. Finally, the model is currently rather descriptive. On the other hand, we hope the proposed framework supports accident occurrence as a product of intellectual curiosity to reduce the number of accidents. In addition, the framework, including definitions of traffic culture and traffic safety culture/climate, seems to some extent merge person (i.e., the role of behavioral factors in traffic accidents) and environment perspectives (i.e., the structure of the complex multilevel sociocultural and technical environment of the traffic, its goals, and its mechanisms) in “the fourth age of safety” (i.e., traffic safety culture).
REFERENCES AAA Foundation for Traffic Safety. (2007). Improving the traffic safety culture in the United States: The journey forward. Washington, DC. Adams, J. G. U. (1987). Smeed’s law: Some further thoughts. Traffic Engineering and Control, 28, 70e73. Adams, J. G. U. (1995). Risk. London: UCL Press. Allstate Insurance Company. (2006, May). The 2006 Allstate America’s best drivers report. Northbrook, IL. Antonsen, S. (2009). Safety culture: Theory, method, and improvement. Surrey, UK: Ashgate. Arnold, J. (2005). Work psychology: Understanding human behaviour in the workplace (4th ed.). London: Prentice Hall. Baxter, J. S., Macrae, C. N., Manstead, A. S. R., Stradling, S. G., & Parker, D. (1990). Attributional biases and driver behaviour. Social Behaviour, 5, 185e192. ¨ zkan, T., Al-Bast, D. A. E., Bener, A., Al Maadid, M. G. A., O Diyab, K. N., & Lajunen, T. (2008). The impact of four-wheel drive on risky driver behaviour and road traffic accidents.
Chapter | 14
Person and Environment: Traffic Culture
Transportation Research Part F: Traffic Psychology and Behaviour, 11, 324e333. Bjo¨rklund, G. (2005). Driver interaction: Informal rules, irritation and aggressive behaviour (Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Social Sciences 8). Uppsala, Sweden: Doctoral dissertation, Acta Universitatis Upsaliensis. Blockey, P. N., & Hartley, L. R. (1995). Aberrant driving behaviour: Errors and violations. Ergonomics, 38, 1759e1771. Caird, K. J., & Kline, T. J. (2004). The relationships between organizational and individual variables to on-the-job driver accidents and accident-free kilometres. Ergonomics, 47(15), 1598e1613. Chapman, R. (1973). The concept of exposure. Accident Analysis and Prevention, 5, 95e110. Cheyne, A., Cox, S., Oliver, A., & Toma´s, J. M. (1998). Modelling safety climate in the prediction of levels of safety activity. Work & Stress, 12, 255e271. Cox, S., & Flin, R. (1998). Safety culture: Plihosopher’s stone or man of straw? Work & Stress, 12, 189e201. Doherty, S. T., Andrey, J. C., & MacGregor, C. (1998). The situational risks of young drivers: The influence of passengers, time of day and day of week on accident rates. Accident Analysis and Prevention, 30(1), 45e52. Elander, J., West, R., & French, D. (1993). Behavioural correlates of individual differences in road traffic crash risk: An examination of methods and findings. Psychological Bulletin, 113, 279e294. Elvik, R. (1996). Accident theory (Liikenneturvallisuusalan tutkijaseminaari No. 21). Helsinki: Gustavelundissa Tuusulassa. European Transport Safety Council. (2003). Assessing risk and setting targets in transport safety programmes. Brussels: Author. Evans, L. (1991). Traffic safety and the driver. New York: Van Nostrand Reinhold. Evans, L. (2004). Traffic safety. Bloomfield Hills, MI: Science Serving Society. Gaudry, M., & Lassarre, S. (2000). Structural road accident models. Oxford: Pergamon. Gaygisiz, E. (2010). Cultural values and governance quality as correlates of road traffic fatalities: A nation level analysis. Accident Analysis and Prevention, 42, 1894e1901. Groeger, J. A. (2000). Understanding driving: Applying cognitive psychology to a complex everyday task. East Sussex, UK: Psychology Press. Glick, W. H. (1985). Conceptualizing and measuring organizational and psychological climate: Pitfalls in multilevel research. Academy of Management Review, 10(3), 601e616. Guldenmund, F. W. (2000). The nature of safety culture: A review of theory and research. Safety Science, 34, 215e257. Hale, A. R., & Hovden, J. (1998). Management and culture. The third age of safety. A review of approaches to organizational aspects of safety, health, and environment. In A. M. Feyer, & A. Williamson (Eds.), Occupational injury, risk prevention and intervention. London: Taylor & Francis. Hatakka, M., Keskinen, E., Laapotti, S., Katila, A., & Kiiski, H. (1992). Driver’s self-confidencedThe cause of the effect of mileage. Journal of Traffic Medicine, 21(Suppl.), 313e315. Hofstede, G. (2001). Culture’s consequences: Comparing values, behaviours, institutions, and organizations across nations (2nd ed.). Thousand Oaks, CA: Sage. Hovden, J., Albrechtsen, E., & Herrera, I. A. (2010). Is there a need for new theories, models and approaches to occupational accident prevention? Safety Science, 48, 950e956.
191
International Nuclear Safety Advisory Group. (1991). Safety culture (Safety Series No. 75-INSAG-4). Vienna: International Atomic Energy Agency. International Road Traffic and Accident Database. (2005, September). Selected risk values for the year 2001. Paris: Author. http://www. internationaltransportforum.org/irtad/index.html. Jaeger, L., & Lassarre, S. (2000). The TAG-1 model for France. In M. Gaudry, & S. Lassarre (Eds.), Structural road accident models. Oxford: Pergamon. Jones, E. E., & Nisbett, R. E. (1972). The actor and the observer: Divergent perceptions of the causes of behaviour. In E. E. Jones, D. E. Kanouse, H. H. Kelley, R. E. Nisbett, S. Valins, & B. Weiner (Eds.), Attribution: Perceiving the causes of behavior. Morristown, NJ: General Learning Press. Laapotti, S. (2003). What are young female drivers made of? Differences in attitudes, exposure, offences and accidents between young female and male drivers. Turku, Finland: University of Turku. Lajunen, T., Corry, A., Summala, H., & Hartley, L. (1998). Cross-cultural differences in drivers’ self-assessments of their perceptual-motor and safety skills: Australians and Finns. Personality and Individual Differences, 24, 539e550. Lajunen, T., & Summala, H. (1995). Driver experience, personality, and skill and safety motive dimensions in drivers’ self-assessments. Personality and Individual Differences, 19, 307e318. Lawton, R., Parker, D., Manstead, A. S. R., & Stradling, S. G. (1997). The role of affect in predicting social behaviours: The case of road traffic violations. Journal of Applied Psychology, 27, 1258e1276. Levia¨kangas, P. (1998). Accident risk of foreign driversdThe case of Russian drivers in south-eastern Finland. Accident Analysis and Prevention, 30, 245e254. Lewin, I. (1982). Driver training: A perceptual-motor skill approach. Ergonomics, 25, 917e924. Lewin, K. (1952). Field theory in social science: Selected theoretical papers. London: Tavistock. Manstead, A. S. R. (1998). Aggressie op de weg. Leidschendam, The Netherlands: Unpublished paper presented at the SWOV Institute for Road Safety Research. Manstead, A. S. R., Parker, D., Stradling, S. G., Reason, J. T., & Baxter, J. S. (1992). Perceived consensus in estimates of the prevalence of driving errors and violations. Journal of Applied Social Psychology, 22, 509e530. Maycock, J., Lockwood, C. R., & Lester, J. F. (1991). The accident liability of car drivers. (Report No. 315). Crowthorne, UK: Transport Research Laboratory. Mearns, K., Flin, R., Gordon, R., & Fleming, M. (1998). Measuring safety climate on offshore installations. Work & Stress, 12, 238e254. Na¨a¨ta¨nen, R., & Summala, H. (1976). Road-user behaviour and traffic accidents. Amsterdam/New York: North-Holland/Elsevier. ¨ z, B., O ¨ zkan, T., & Lajunen, T. (2009). Professional and nonO professional drivers’ stress reactions and risky driving. Transportation Research Part F: Traffic Psychology and Behaviour, 13(1), 32e40. ¨ z, B., O ¨ zkan, T., & Lajunen, T. (2010). An investigation of the relaO tionship between organizational climate and professional drivers’ driver behaviours. Safety Science, 48, 1484e1489. ¨ z, B., O ¨ zkan, T., & Lajunen, T. (under review). An investigation of O professional drivers: Transportation companies’ climate, driver behaviours and performance, and accidents.
192
¨ zkan, T. (2006). The regional differences between countries in traffic O safety: A cross-cultural study and Turkish case. Helsinki: University of Helsinki. http://ethesis.helsinki.fi/julkaisut/kay/psyko/vk/ozkan/ theregio.pdf. ¨ zkan, T., & Lajunen, T. (2005). A new addition to DBQ: Positive Driver O Behaviour Scale. Transportation Research Part F: Traffic Psychology and Behaviour, 8(4e5), 355e368. ¨ zkan, T., & Lajunen, T. (2007). The role of personality, culture, and O economy in unintentional injuries: An aggregated level analysis. Personality and Individual Differences, 43, 519e530. ¨ zkan, T., & Lajunen, T. (under review). Traffic climate scale. O ¨ zkan, T., Lajunen, T., Chliaoutakis, J. E., Parker, D., & Summala, H. O (2006a). Cross-cultural differences in driving skills: A comparison of six countries. Accident Analysis and Prevention, 38(5), 1011e1018. ¨ zkan, T., Lajunen, T., Chliaoutakis, J. E., Parker, D., & Summala, H. O (2006b). Cross-cultural differences in driving behaviours: A comparison of six countries. Transportation Research Part F: Traffic Psychology and Behaviour, 9, 227e242. ¨ zkan, T., Lajunen, T., Parker, D., Su¨mer, N., & Summala, H. (2010). O Symmetric relationship between self and others in aggressive driving across gender and countries. Traffic Injury Prevention, 11(3), 228e239. ¨ zkan, T., Lajunen, T., & Summala, H. (2006). Driver behaviour quesO tionnaire: A follow-up study. Accident Analysis and Prevention, 38, 386e395. ¨ zkan, T., Lajunen, T., Walle´n Warner, H., & Tzamalouka, G. (2006, O July). Traffic climates and driver behaviours in four countries: Finland, Greece, Sweden, and Turkey. Athens, Greece: Paper presented at the 26th International Congress of Applied Psychology. Page, Y. (2001). A statistical model to compare road mortality in OECD countries. Accident Analysis and Prevention, 33, 371e385. Parker, D., West, R., Stradling, S. G., & Manstead, A. S. R. (1995). Behavioural characteristics and involvement in different types of traffic accident. Accident Analysis and Prevention, 27, 571e581. Pelzman, S. (1975). The effects of automobile safety regulation. Journal of Political Economy, 83, 677e725. Prince Market Research. (2006, April). In the driver’s seat: 2006 AutoVantage road rage survey. Nashville, TN: Affinion Group. http:// www.pmresearch.com/downloads/09RoadRageReport.pdf. Raoufi, M. (2003). Iran road traffic accidents. http://www.payvand.com/ news/03/jul/1138.html. Reason, J. (1990). Human error. Cambridge, UK: Cambridge University Press. Reason, J., Manstead, A., Stradling, S., Baxter, J., & Campbell, K. (1990). Errors and violations on the roads. Ergonomics, 33, 1315e1332. Ross, L. (1977). The intuitive psychologist and his shortcomings: Distortions in the attribution process. In L. Berkowitz (Ed.), Advances in experimental social psychology. New York: Academic Press. Ross, L., Greene, D., & House, P. (1977). The “false consensus effect”: An egocentric bias in social perception and attribution processes. Journal of Experimental Social Psychology, 13, 279e301. Salmon, P. M., Lenne, M. G., Stanton, N. A., Jenkins, D. P., & Walker, G. H. (2010). Managing error on the open road: The contribution of error models and methods. Safety Science, 48, 1225e1235.
PART | II
Key Variables to Understand in Traffic Psychology
SARTRE. (1998). The attitude and behaviour of European car drivers to road safety. Report on principal results. (Project on SARTRE 1, 2, and 3). http://sartre.inrets.fr/documents-pdf/repS3V1E.pdf. Schein, E. H. (1992). Organizational culture and leadership. San Francisco: Jossey-Bass. Schwartz, S. H. (1992). Universals in the content and structure of values: Theoretical advances and empirical tests in 20 countries. Advances in Experimental Social Psychology, 25, 1e65. Schwartz, S. H. (1999). A theory of cultural values and some implications for work. Applied Psychology: An International Review, 48(1), 23e47. Schwartz, S. H. (2004). Mapping and interpreting cultural differences around the world. In H. Vinken, J. Soeters, & P. Ester (Eds.), Comparing cultures: Dimensions of culture in a comparative perspective (pp. 43e73). Leiden, The Netherlands: Brill. Smeed, R. J. (1949). Some statistical aspects of road safety research. Royal Statistical Society Journal (A), 112(Part 1, Series 4), 1e24. Spolander, K. (1983). Bilfiirares uppfattning om egen kiirjormdga [Drivers’ assessment of their own driving ability]. (Report No. 252). Linko¨ping: Swedish Road and Traffic Research Institute. Stradling, S. G., & Parker, D. (1996). Violations on the road: Bad attitudes make bad drivers. Birmingham, UK: Paper presented at the International Conference on Road Safety in Europe. ¨ zkan, T. (2002). The role of driver behaviour, skills, and Su¨mer, N., & O personality traits in traffic accidents. Turkish Journal of Psychology, 17(50), 1e22. Svedung, I., & Rasmusen, J. (1998). Organisational decision making and risk management under pressure from fast technological change. In A. Hale, & M. Baram (Eds.), Safety management the challenge of change. Oxford: Pergamon. United Nations. (1997). Statistics of road traffic accidents in Europe and North America. Geneva: Author. Wegman, F., Aarts, L., & Bax, C. (2008). Advancing sustainable safety: National road safety outlook for The Netherlands for 2005e2020. Safety Science, 46(2), 323e343. Wiegmann, D. A., von Thaden, T. L., & Gibbons, A. M. (2007). A review of safety culture theory and its potential application to traffic safety. In Improving the traffic safety culture in the United States: The journey forward. Washington, DC: AAA Foundation for Traffic Safety. World Health Organization. (2002). Burden of Disease Project: Global burden of disease estimates for 2001. Geneva: Author. http://www. who.int/healthinfo/global_burden_disease/estimates_regional_2001/ en/index.html. World Health Organization. (2004). World report on road traffic injury prevention. Geneva: Author. Zaidel, D. M. (1992). A modeling perspective on the culture of driving. Accident Analysis and Prevention, 24(6), 585e597. Zhang, W., Huang, Y., Roetting, M., Wang, Y., & Wei, H. (2006). Driver’s views and behaviours about safety in China: What do they not know about driving? Accident Analysis and Prevention, 38, 22e27. Zohar, D. (1980). Safety climate in industrial organizations: Theoretical and applied implications. Journal of Applied Psychology, 65(1), 96e102.
Chapter 15
Human Factors and Ergonomics Ilit Oppenheim and David Shinar Ben-Gurion University of the Negev, Beer Sheva, Israel
1. INTRODUCTION “Human factors” or “ergonomic” aspects of traffic psychology refer to the implications of the road user’s physical, physiological, cognitive, personality, and social behavior concerns and considerations in the design of vehicles and roadways. Human factors/ergonomics (HFE) is a relatively new scientific discipline, with the first book in this area published by Chapanis, Garner, and Morgan in 1949. It is distinct from psychology, engineering, and design because the focus of analysis is on the interaction between people and technology rather than on people or technology independently from each other. This means that HFE requires an interdisciplinary approach. HFE scientists are concerned with human performance in technological systems with a view to optimization of the design of the system in terms of values such as effectiveness, safety, comfort, and well-being. Like all scientific disciplines, HFE is characterized by theoretical and methodological development together with empirical investigations. The latter tend to shift between real-world studies and laboratory studies (Stanton, Young, & Walker, 2007). The importance of HFE in highway traffic safety was highlighted in two landmark studies that were published at approximately the same time in the United States (Treat et al., 1977) and in the United Kingdom (Sabey & Staughton, 1975). The studies focused on the causes of traffic accidents, and they identified factors associated with large samples of driving crashes. The research groups, which were unaware of each other’s activities, obtained remarkably similar findings. The U.S. study found the road user to be the sole cause of 57% of crashes, the environment in 3%, and the vehicle in 2%. The corresponding values from the UK study were 65, 2, and 2%, respectively. Approximately half the crashes were caused by a combination of factors, in which had one not existed the crash would not have occurred. Thus, in conjunction with other causes, the road user was identified as a sole or contributing factor in 94% of crashes in the U.S. study and in 95% of crashes in the UK study. The road environment alone or with other factors was identified as a causal factor in Handbook of Traffic Psychology. DOI: 10.1016/B978-0-12-381984-0.10015-3 Copyright Ó 2011 Elsevier Inc. All rights reserved.
31% of crashes in the U.S. study and in 27% of crashes in the UK study (Figure 15.1). Although the driver is the main actor in the driving activity, driving is not an isolated activity. It takes place in a wider context in which the driver constantly interacts with his or her immediate environment and the vehicle. Although the human factor is more dominant than environmental or vehicle factors in the causation of accidents, the control of the road factorsdthat is, any external conditions and surroundings of the vehicle (e.g., road, traffic, and visibility conditions)dand the control of the vehicle characteristics (e.g., braking and steering performance and passenger protection) are often much easier than the control of the human factor. Moreover, by good design, it is even possible to compensate for various human failures and limitations and thus decrease the number of traffic accidents (Iyinam, Iyinam, & Ergun, 1997; Shinar, 2007).
Road environment
Road user
2
Vehicle
2
65
2
57
3
4
24
UK
6
27
US
1 3
1 1
28/34
95/94
8/12
FIGURE 15.1 Road user, environment, and vehicle contribution to crashes. Source: Reproduced from Rumar (1985) with kind permission of Springer Science and Business Media.
193
194
2. THE VIEW FROM THE DRIVER’S SEAT: DRIVER-CENTERED DESIGN IMPLICATIONS The human-centered approach to design places the user in the heart of the system. In driving, this means that all aspects of vehicle and environmental design considerations have to be subservient to the driver’s characteristics, needs, capabilities, and limitations. From these, we try to derive the most appropriate technologies. The human-centered approach was first applied to the humanecomputer interaction and has since been extended to other systems, such as automotive systems (Michon, 1993) and aviation (Cacciabue, Mauri, & Owen, 2003). HFE addresses the interaction between human beings and the devices or systems that they operate in various capacities. Good HFE design requires an understanding of the characteristics of the users and the tasks in which they are engaged. Therefore, the user-centered design philosophy is that an effective, safe, and accepted product must be designed around the user. It takes into account not only physical and perceptual capabilities (e.g., reaction time or visual acuity) but also behaviors, knowledge, motivations, and attitudes. Common tools that we use to understand how we interact with our environment (including our vehicle, other vehicles, the roadway, and other road users) are the theories and models. Their importance for improving highway safety was made very succinctly by Kantowitz et al. (2004, as cited in Shinar, 2007): Absent the theories, it is almost impossible to specify what new countermeasures might emerge. Thus, what is a standard operating procedure for many human factors researchers (using models) might require an act of faith from practicing highway engineers who do not normally invoke theories of human behavior. If aviation, nuclear power, and humanecomputer interaction can create better countermeasures through models, so can driving. (pp. 85e86)
According to Kantowitz (2000), a theory is the best practical human factors tool because (1) it fills in where data are lacking; (2) in a computational format, it can provide quantitative predictions needed by engineers; (3) it prevents us from reinventing the wheel by allowing us to recognize similarities among problems, such as the tendency of drivers to adopt inappropriate decision criteria in many situations; and (4) it is reusable. Once the investment has been made to build a model for a particular domain, the theory can be recycled inexpensively to answer many system design questions. The body of research that has accumulated on driving behavior is not just a collection of findings and conclusions but also a coherent picture that emerges out of these findings and conclusions. That picture is our theory of driving
PART | II
Key Variables to Understand in Traffic Psychology
behavior. Once we have a theory, we can continue gathering additional “facts” to fill the remaining gaps. The purpose of the models and theories of driver behavior is to make sense of it all. A theory and a model are not synonymous. A theory is a conceptual organization of concepts, mechanisms, and processes that are involved in the operation of a system, such as the driver in traffic. A model is less presumptive in the sense that it does not presume that these mechanisms and processes actually exist but only that if we posit them, then we can explain human behavior. Often, a model of human behavior is developed and then a search is made to determine if some of its mechanisms actually exist. A model can often serve as a basis for a theory. In general, unless there is independent evidence for the existence of specific processes and mechanisms, it is safer to talk of models of driver behavior than theories of driver behavior (Shinar, 2007).
2.1. Models of Driver Behavior The first attempts to model drivers’ behavior were made in the early 1960s (Delorme & Song, 2001) to improve driving safety and driver education (McKnight & Adams, 1971). Since then, a plethora of models have been suggested, and the best way to describe them is by presenting them within a conceptual framework. Figure 15.2 provides one such attempt (based in part on Weller, Schlag, Gatti, Jorna, & Leur, 2006). One approach is to describe driving behaviors within various driving tasks or what the driver does. The principal limitation of this approach is that it is purely descriptive and with very little predictive power. An alternative approach, the functional approach, is to model behavior relative to the driver’s tasks or functions. This approach attempts to predict driver behavior by focusing on why drivers do what they dodthat is, the situational and motivational factors that are involved in the risk management of driving. One advantage of these models is the potential to implement them, either by generating a simulation of the driver or by integrating them into some already
Descriptive models
Functional models
Hierarchical models
Motivational models
Control loop models
Informationprocessing models
FIGURE 15.2 Driver models. Source: Based on Weller et al. (2006).
Chapter | 15
Human Factors and Ergonomics
existing traffic simulation tools or driver assistance devices such as collision warning systems. Some of the variation among the models is due to the different applications for which the models are intended, and some of the variation is due to the part of the driving task they are intended to describe. Because driving encompasses so many tasks and subtasks at different levels, often performed by the driver simultaneously, it is perhaps not surprising that it is difficult to find any consensus in the literature on how the process of driving should be modeled. The scope of this chapter precludes a detailed description of the different models, and they are only described here in general terms as they relate to the different types listed in Figure 15.2. Descriptive behavioral models focus on what the drivers do. These models attempt to describe the entire driving task or some components of it in terms of what the driver does or has to do. The predictive power of such models is very limited because they do not take into account the forces that shape the different behaviors such as driver motivation, skills, capabilities, and limitations in different situations (Carsten, 2007). Despite this severe limitation, these models have provided a strong impetus to driving safety research (Lee, 2008; Michon, 1985; Parasuraman & Riley, 1997; Salvucci, 2006; Sheridan, 1970, 2004). The descriptive models can be divided into hierarchical models (e.g., Michon, 1985) and control loop models (e.g., McRuer & Weir, 1969). The hierarchical models describe behavior in terms of a hierarchy of three distinct types of behaviors, each building on the level below it. The lowest level is an operational, control level. At this level, most behaviors are automatic and consist of quick
195
responses to the changing environment (e.g., braking when a lead car slows down). The second level is a tactical, vehicle maneuvering level referring to how traffic situations are mastered. The behaviors are less reflexive and consist of conscious decisions in the driving, such as a decision to change lanes before exiting a highway. The third and highest level is a planning or strategic level, and it consists of long-term decisions such as which route to choose or even whether to drive at all. Thus, the three levels can be distinguished by the task requirements, the time frame needed to carry them out, and the cognitive processes involved at each level (Figure 15.3). The second type of descriptive models are the control loop models. These models describe the operation of the driving task in terms of inputs, outputs, and feedback. Control loop models deal primarily with the steering control aspect of driving in order to follow a specified route (McRuer, Allen, Weir, & Klein, 1977; as cited in Fastenmeier & Gstalter, 2007). These models of driving have traditionally been couched either in terms of guidance and control or in terms of human factors. Unfortunately, expanding these models to accommodate the rapidly growing complexity and sophistication of modern cars is a very daunting task. Within limits, due to their quantitative approach, such models can provide coherent and consistent ways of describing driver performance that help engineers develop and validate technical concepts for semi- and fully automated systems in cars (Hollnagel, Na˚bo, & Lau, 2003). Functional models, which include motivational models and information processing models, are the most likely to help understand complex tasks such as driving (Michon, 1985). These models focus on the mental activities
Knowledge-based behavior Identification
Decision
Planning
Route speed criteria
Rule-based behavior Recognition
Association
Strategic level
Stored rules
Maneuvering level
Feedback criteria Skill-based behavior Feature formation
Sensory input
Stimulus reaction automatisms
Action
Control level
FIGURE 15.3 Combination of performance levels according to Rasmussen (1986) and the hierarchical model according to Michon (1985) in Weller et al. (2006)
196
PART | II
involved in driving and attempt to explain why the driver undertakes certain actions. These models are necessary to understand human errors and difficulties and also to design driving assistance adapted to driver needs (Keith et al., 2005). These models strongly emphasize the driver’s cognitive state and have incorporated important psychological concepts such as motivation and risk assessment. Information processing models of the driving task belong to functional models because they involve interactions between different components (Michon, 1985). These models consist of different stages, which include perception, decision and response selection, and response execution. Each stage is assumed to perform some transformation of data and to take some time for its completion (Wickens, 1992). The driver in such models is described as a passive information transmission channel who performs different acts within capacity limitations. The system has two more crucial components: the attention allocation mechanism and a feedback loop. The feedback loop indicates that the process is an ongoing one that is continuously modified in accordance with new stimuli (Figure 15.4;Shinar, 2007). Much experimentation has been directed at determining which types of processing can occur simultaneously and which must occur sequentially. During the 1970s, a paradigm shift took place in the study of attention (Kahneman & Treisman, 1984). The shift involved a move away from determining the limits of processing and locus of the attentional bottleneck. Later, based in large part on theoretical advances by Schneider and Shiffrin (1977; Shiffrin & Schneider, 1977), research was directed at determining the characteristics and conditions under which automaticity develops. This work influenced research in HFE and began to influence theory in driving behavior (Ranney, 1994). Rasmussen’s model of information processing (Rasmussen, 1986), which has been proven to be heuristically fruitful in many HFE applications, serves as a starting point. Furthermore, a feedback loop was integrated into the model
Own car Pedestrians
of driverevehicleeenvironment, helping to adapt task difficulty or the desired amount of strain experienced by coping with the stressors that originate from drivers’ behavior, showing the influence of the drivers’ actions on the future situations with which they have to cope. Generic information processing models of the human driver can provide useful generalizations. They have value in systems engineering and in seeking to predict asymptotic limits of human performance. Even relatively simple models can have value as engineering tools by providing structure and drawing attention to possible limitations in the performance of humans in systems. However, as a means of understanding why specific individuals on a particular day in a particular set of circumstances behaved (or failed to behave) in a particular way, or predicting how an individual might behave in a given set of unexpected circumstances, such models are very limited and have little to offer. Motivational models of driving emerged in the 1960s and 1970s. Motivational models focus on “what the driver actually does” in a given traffic situation rather than on the level of driving skill. The main assumptions of these models are that driving is self-paced and that drivers select the amount of risk they are willing to endure in any given situation. The driver is seen as an active decision maker or information seeker (Gibson, 1966) rather than as the passive responder implicit in many information processing models. The importance of the situational factors resulted from the failure of earlier attempts to relate stable personality traits to accident causation. The risk associated with possible outcomes is seen as the main factor influencing behavior; however, these models also assume that drivers are not necessarily aware of the risks associated with other outcomes. Examples of motivational models include risk compensation models (Wilde, 1982), risk threshold models (Naatanen & Summala, 1976), and risk avoidance models (Fuller, 1984).
Automatic sensors
Automatic controls
Displays
Driver characteristics: personality, attitudes, experience, impairments, visual abilities, etc.
Road Other vehicles
Key Variables to Understand in Traffic Psychology
Driver sensory systems
Driver perceptual and attentional capacitites
Driver decisionmaking and response selection
Driverresponse capabilities
Vehiclecontrol dynamics
A simplified block diagram of the driver functions in the driver-vehicle-road system.
FIGURE 15.4
A limited capacity model of driver information processing. Source: Reproduced from Shinar (1978).
Chapter | 15
Human Factors and Ergonomics
Motivational models take into account interactions between general mechanism and individual differences. Although in most cases motivational models are equivalent to risk models, recent approaches view workload homeostasis as the most important motivational background (Fuller, 2005). Motivational models aim to describe how the driver manages risk or task difficulty (Carsten, 2007). For validation, all models have to rely on some dependent measure of driving safety. This could be an intermediate measure (e.g., the use of seat belts), but more typically it is based on crash-based measures, such as accident frequencies and rates or injury frequencies and rates. However, traffic safety is more than the mere absence of accidents (Ranney, 1994). Relative to the driver, we often strive to evaluate these measures in terms of driver performance and behavior measures manifested by errors and reaction time to various events. Various taxonomies of human error have been proposed, and three perspectives currently dominate the literature: Norman’s (1981) error categorization; Reason’s (1990) slips, lapses, mistakes, and violations classification; and Rasmussen’s (1986) skill, rule, and knowledge error classification. Here, we focus only on Reason’s classification of different types of unsafe acts (Figure 15.5). Slips and lapses are defined as behaviors related to attentional and memory failures that might impact driver safety (Wickens, Toplak, & Wiesenthal, 2008), and both are characterized by unintended behaviors (or failures to behave in some way). Slips relate more directly to psychomotor components of driving at the operational level
Error types Slip
197
of control and refer to events in which the planned action would have achieved the desired goal but the right intention is incorrectly executed (e.g., when a driver who plans to push the brake pedal to slow down inadvertently pushes the accelerator pedal; the intention was correct, but the execution was erroneous). On the other hand, lapses represent the failure to carry out any action at all. These are omission errors based on forgetfulness (e.g., a driver forgetting to turn off the lights when departing the car, although fully intended to do so). Lapses are of particular relevance to roadway accidents because they reflect deficiencies in skill-based automatic behaviors (Ranney, 1994; Reason, 1990). On the contrary, mistakes occur when a driver intentionally performs an action that is wrong (e.g., a driver decides to accelerate when the right action would have been to brake or slow down) as a result of limitations in perception, memory, and cognition. Mistakes originate at the planning level rather than at the execution level, and they are likely to precipitate inappropriate maneuvering decisions. Although both rule- and knowledge-based mistakes characterize intentions that are not suitable for the situation, there are some differences between the two. Rulebased mistakes tend to be made with confidence (misapplication of a good procedure; e.g., performing a task that has been successful before in a particular context), whereas knowledge-based mistakes are more likely to occur in a situation in which rules are not applicable and the operator becomes less certain (e.g., performing a task that is “unsuitable, inelegant, or inadvisable” at the most basic
Attentional failures - Intrusion - Omission - Reversal - Misordering - Mistiming
Unintended action Memory failures Lapse
- Omitting planned items - Place-losing - Forgetting intentions
Unsafe acts
Mistake
Rules-based mistakes -Misapplication of good rule -Application of bad rule Knowledge-based mistakes -Many variables forms
Intended action Intentional noncompliance Violation
-Routine violations -Exceptional violations -Acts of sabotage
FIGURE 15.5 Reason’s (1990) classification of unsafe acts. Source: Reproduced from Weller et al. (2006).
198
level). The knowledge-based mistakes will also involve much more conscious effort, and the chances of making a mistake while functioning at this level are higher than they are at a rule-based level because there are many more ways in which information acquisition and integration may fail (Reason, 1990). The final category, violation, involves “deliberate deviations from those practices deemed necessary to maintain the safe operation of a potentially hazardous system” (Reason, 1990, p. 195). In the case of driving, this would be deliberate deviations from accepted procedures, standards, and rules of safe driving (i.e., speeding). Research has shown that violations are positively statistically associated with crash involvement (Lindgren, Brostro¨m, Chen, & Bengtsson, 2007). Comparing violations with errors, Reason states that errors should be related to the individual cognitive processes, whereas violations concern the social text in which they occur. Errors may therefore be minimized by retraining, memory aids, and better humanemachine interfaces. Violations, on the other hand, should possibly be dealt with by trying to change users’ attitudes, beliefs, and norms and by improving the overall safety culture (Lindgren et al., 2007). When measuring behavior in terms of continuousdrather than discretedmeasures, we tend to use driver response time to various events. Response time is typically composed of at least three components: (1) the time required for the driver to perceive the sensory input and to decide on a response, often labeled perceptione reaction time; (2) the time used to perform the programmed movement, such as lifting the foot from the accelerator and touching the brakedoften labeled movement time; and (3) the time the physical device requires to execute its response, such as the time needed for the brakes to engage once the brake pedal has been depressed. Because the mental processing involved in the perception reaction time is an internal quantity that cannot be measured directly and objectively without a physical response, it is usually measured jointly with movement time (Setti, Rakha, & El-Shawarby, 2007), although not always (WarshawskyLivne & Shinar, 2002). Brake reaction time (RT) is a parameter of driving behavior that has not only attracted the interest of researchers but also is of great importance in road design and the accident litigation process. Brake RT is used in assessing stopping sight distance, durations of the amber phase in traffic signals, recommended distance in car following (recommended headways), etc. In accident litigation, the legal outcome often hinges on whether or not the participant driver reacted to the impending collision within an “acceptable” time, where acceptability is established from a certain percentile of RT distribution thought to represent the driver population (or relevant fraction of it) in relevant conditions (Summala, 2000).
PART | II
Key Variables to Understand in Traffic Psychology
2.2. Behavior (TypicaldAffected by Needs) Versus Performance (Capabilities) In modeling the driver’s interactions with the vehicle and the environment, we must also distinguish between driver performance and driver behavior (Naatanen & Summala, 1976; Shinar, 1978). Performance refers to “best behavior,” or what drivers are capable of doing in a given situation (limits of maximal behavior), and it is typically measured in a controlled experiment. Behavior refers to “typical behavior,” or what drivers actually do in most of their driving, and it is more difficult to measure. Driver performance is the end product of what a driver can do, given human limitations and given vehicle and environmental constraints. Driver behavior is what the driver actually does given the previously discussed limitations and constraints and given the driver’s needs, motivation, level of alertness, and personality. This distinction helps clarify differences between the major research paradigms used to study driving behavior. Models designed to predict driver performance are based on cognitive and physiological psychology and most often depict the driver as a limited capacity information processor. Models designed to explain and predict the more complex real on-road behavior are based on theories of personality, social psychology, and organizational behavior and assume that actual driving behavior represents the style and strategy drivers adopt to achieve their goals. In reality, our behavior on the road is a combination of both typical behavior (most of the time we drive) and maximal performance (when we find ourselves in very demanding situations). Thus, both models are complementary, and they are useful in rather different contexts (Shinar, 2007). The distinction is useful in analyzing factors that contribute to crashes. The contributing factors that have been studied include cognitive abilities, social context, emotion, personality, experience, and hazard perception skills. The situation or the context in which the driver drives plays a crucial role in determining the type of actions a driver is likely to take. Without the context, the validation of driving behavior models in real driving situations would be difficult (Rakotonirainy & Maire, 2005). How do drivers make their choices in concrete situations? To what extent are they affected by various potential competing needs and constraints? The answer is not a simple one. For example, to demonstrate how different situations and needs can affect speed choice, Shinar (2001) conducted roadside interviews of 225 Israeli drivers. The interviews were conducted at gas stations along three types of interurban roads with three different speed limits: twolane undivided roads without hard shoulders with a speed limit of 80 km/h (50 mph), improved two-lane roads with hard shoulders posted at 90 km/h (56 mph), and four-lane limited-access divided highways posted at 100 km/h
Chapter | 15
Human Factors and Ergonomics
199
115 SP_Limit 80
Estimated speed
110 105
107
SP_Limit 90
104
SP_Limit 100
100 95
98 96
98 97
97
94
109 108
97
94 90
89
90
87
87 84
85 80
98
Usual speed
Economic speed Speed w/family
Fun speed Safe speed
Legal speed
Speed category FIGURE 15.6 Average speed in kilometers/hour that drivers report they would select to maximize different goals relative to their actual speed and to the speed limit. Source: Reproduced with permission from Shinar (2001).
(62 mph). At each site, the drivers were asked about the speed that they usually maintain on that road, the speed they select when driving with their family, the speed they would choose if their primary goal was to save on fuel and vehicle wear and tear, the speed they considered safe, and the speed they would choose if there were no enforcement at all on the roaddthe one they would select to maximize pleasure or enjoy the “fun” of driving. Finally, they were tested on their knowledge of the actual speed limit on that road. The results, plotted separately for each road type, are displayed in Figure 15.6. First, note that in general, on all choice dimensions, peoples’ choices are primarily affected by the roadway, with roadways posted at and designed for slower speeds generating lower speed choices. Second, different motives lead to different speed choices. The desire for fun motivates people to drive at the highest speed, whereas frugality and safety motivate them to drive at the lowest speed. Interestingly, the drivers’ perceptions of the “safe” speed were sensitive to the nature of the road, but on the roads with the lower speed limits, the perceived safe speed was slightly higher than the legal speed. This is despite the fact that nearly all drivers on all roads correctly identified the legal speed limit. Finally, the actual speed drivers report driving on the road seems to be a compromise among the various motives, road design constraints, and enforcement, although it does seemdat least in the Israeli driving culturedto be much closer to the “fun” speed than the “safe” speed.
3. DRIVER VARIABLES AFFECTING THE DRIVEReVEHICLE INTERACTION Research on driver behavior and performance has identified at least six relatively independent psychological dimensions
that are likely to impact driver behavior when faced with advanced automobile automation (Stanton et al., 2007): locus of control, mental models, mental workload, situation awareness, stress, and trust. These factors are likely to interact with each other and with automation in a complex and unpredictable way. Useful models should include these dimensions in order to be able to predict driver behavior. For example, by postulating how a driver manages attention to multiple inputs, we can predict the negative effects of underload or overload on performance. This is applicable to support systems that are designed to reduce driver information load (e.g., adaptive cruise control, hazard detection systems, and lane deviation warnings). Thus, reducing mental demand will not necessarily mean that drivers will have spare attentional capacity. Rather, it suggests that reduction in mental demand in the driving task may lead to either (1) corresponding reductions in attentional resources allocated to the driving, and ultimately task-induced fatigue, or (2) engagement in additional tasks (e.g., talking on the phone or text messaging) that may negate the benefits of the original load reduction by inducing driver distraction. Both outcomes are probably counterintuitive predictions for those without HFE training and illustrate the importance of considering HFE in vehicle design. Intuitive introductions of driver support systems to future vehicles without careful considerations of HFE principles and knowledge will more likely lead us into an age of automation nightmare than automation utopia (Stanton et al., 2007). A critical factor in automation is that it shifts the emphasis in the role of the driver from the lower level “vehicle control” to a much higher level of “driving control.” The implication of this shift is that automation not only frees the driver from select physical tasks (e.g.,
200
maintaining the vehicle at a constant speed) but also eliminates some of the driver’s cognitive tasks (e.g., deciding whether to brake or accelerate in response to other road users). The foundation of automation philosophy comes from aviation, in which time constants are often very short and cognitive overload is a real risk. The direct application without careful consideration of the differences between the two systems may not be appropriate. Some solutionsdbased on the distinction between hard automation (in which the pilot/driver cannot overrule the automated response) and soft automation (in which the pilot/ driver can to various degrees override the automated response)dhave been offered (Young, Stanton, & Harris, 2007), but the field is still in an embryonic stage.
3.1. Vision and Perception: Physical and Psychological Variables If the eyes are our windows to the mind, then observing eye behavior is a natural tool to use to understand how the mind acquires and processes visual information (Shinar, 2008). The extensive research that has been conducted on eye movements in the past half century clearly establishes that eye movement data reflect moment-to-moment cognitive processes (Rayner, 1998), and that eye movements are closely linked to attention: People tend to direct their gaze and fixations to the objects of their attention (Hoffman & Subramaniam, 1995). Thus, an alert driverdunlike a fatigued or drugged driverdhas a very active eye movement pattern in which the eyes constantly jump (in what is labeled as saccadic movements) from one area of interest to another (collecting information in what is labeled fixations). The information gained from our studies of eye movement behavior can be appreciated when we compare the fixation patterns of novice drivers to those of experienced drivers. Experienced drivers move their gaze among various sources of information, directing most of their fixations ahead on the road where hazards are likely to first appear. In early seminal research, Mourant and Rockwell (1970, 1972) showed that experienced drivers’ fixations are widely dispersed with a modal location slightly above and to the right of the roadway (for U.S. drivers, who drive on the right side of the road), where most signs tend to be concentrated and from where pedestrians may enter the driving lane. As drivers become more familiar with the route, they attend to fewer signs, their fixations become more concentrated, and their modal position drifts closer to the driving lane and farther down the road, where drivers can obtain information with maximal lead time to respond to it. Their studies also revealed that experienced drivers rarely fixate on the lane markers, suggesting that monitoring vehicle position within the lane is accomplished through peripheral visiondand therefore the cues, such as lane markings, should be
PART | II
Key Variables to Understand in Traffic Psychology
conspicuous. In contrast, beginning drivers have to learn “how to look at their surroundings” just as they have to learn how to control their car. Unfortunately, it takes longer to learn how to acquire information than how to control the car. Thus, novice drivers are much more dependent on their fixations for maintaining their car in the lane, and their fixations are distributed over a much smaller part of the visual scene, much closer to the car and on the lane markings. Also, the novice driversdwho are probably visually overloadeddfixate on their rearview and side mirrors much less than do experienced drivers. In short, novice drivers are much less efficient at obtaining the necessary cues to drive safely and consequently are deficient in their ability to anticipate hazardsda key to safe defensive driving (Borowsky, Shinar, & Oron-Gilad, 2010). Similarly, an analysis of their precrash behavior indicates that they have inadequate visual scanning for potential obstacles and inattention to the driving task in generaldor not looking at the right place at the right time (McKnight & McKnight, 2003). Eye movements are also a powerful tool for assessing time-sharing and workload (Kiger, Rockwell, & Tijerina, 1995; Mourant, Rockwell, & Rackoff, 1969; Rockwell, 1988), for evaluating sign design and placement (Bhise & Rockwell, 1973), for showing impairments from alcohol and marijuana (Papafotiou, Stough, & Nathan, 2005), and for assessing the demands of various road geometries (Shinar, McDowell, & Rockwell, 1977). In approximately the past 10 years, eye movement research findings have also been incorporated into models of the effects of novel in-vehicle systems on visual search ¨ stlund, 2005; Reingold, (Engstro¨m, Johansson, & O Loschky, McConkie, & Stampe, 2003; Salvucci, 2006; Salvucci, Zuber, Beregovaia, & Markley, 2005). This type of modeling can set the stage for what is probably the most intriguing application of eye movements in driving: as triggers for actuating in-vehicle alerting and control systems. For example, eye movement behavior (in combination with driving performance measures) can be used to detect visual driver distraction in real time (Victor, Harbluk, & Engstro¨m, 2005; Zhang, Smith, & Witt, 2006). A note of caution is appropriate in interpreting and using eye movement data. Visual fixations are not synonymous with attention. Although people tend to move their eyes to the targets of their attention, the converse is not true: The location of our fixations does not always reveal the target of our attention. The common phenomenon of “looked but did not see” that precedes many crashes is a testament to that (Stutts, Reinfurt, Staplin, & Rodgman, 2001; Treat et al., 1977). The open eyes always fixate somewhere in space, but the mind is still free to roam and concentrate on nonvisual stimuli, such as auditory inputs (e.g., when talking on the cell phone) and deep thoughts. In fact, in such situations, there is a suppression of much of the
Chapter | 15
Human Factors and Ergonomics
201
saccadic eye movements, and the apparent concentration of the driver’s gaze on the road ahead is misleading, making target detection performance poorer (Victor et al., 2005). Focusing the eyes on the right object at the right time is often necessary to perceive safety-related information. However, once focused, how welldor more technically speaking, with what acuityddo we need to be able to see each object? It is a fact that the blind cannot drive. On the other hand, it is not at all obvious how well we need to see in order to drive. As typically measured and defined, visual acuity is a measure of a person’s ability to resolve details when they are presented under optimal illumination (meaning high levels of illumination and no glare), in the middle of the observer’s visual field (meaning when the observer is directly staring at it), with both the target and the observer being static (meaning neither one of them is moving), and under no time constraints (meaning the person can take as long as he or she needs to decide what the detail is). This highly constrained measure of visual performance is very relevant to reading from the board in a classroom (for which it was originally developed by Snellen) or to deciphering the name of a street when stopped at an intersection in the middle of the day. However, driving involves a very different visual task. In driving, none of the previous conditions apply most of the time: The driver is moving relative to the visual environment, the lighting conditions are often far from optimal (night, fog, and glare), emerging dangers typically first appear off to the side of the visual field and not where the driver is necessarily looking (e.g., when a pedestrian darts into the street or a car converges from an adjacent lane), and the driver has very little time to perceive and respond to many hazardous situations. Finally, seeing small details may not be the skill we need at all. Pedestrians and vehicles with which we might collide are not small details, and we do not collide with them because we are unable to read the words on a T-shirt or the license plate number of the car. Measures other than static visual acuity that appear to be much more relevant to driving safety (and licensing) include contrast sensitivity, dynamic visual acuity, and useful field of view (Owsley et al., 1998; Shinar & Schieber, 1991).
of our memory and biases in decision making are critical to our safety, and much of the current vehicle and environmental design is geared toward adjusting for these (as discussed later). Automation is one potential aid to memory and decision making. A simple example is the automatic railroad crossing gates that come down whenever a train reaches a threshold distance from the crossing that would not enable some drivers to cross safely. The effectiveness of such systemsdas with most automated support systemsd depends on their perceived validity. To be safe for nearly all drivers, the flashing lights or gates are often activated when the train is 2.5 minutes away. However, because some drivers consider this a very long lead time, they are tempted to cross the rails despite the flashing lights or the lowered gates because the information provided is incongruent with their direct perceptions (that they still have enough time to cross) (Shinar & Raz, 1982). The alternative approachd that of providing drivers with more information to facilitate their decisionsdalso does not ensure greater safety (e.g., providing pedestrians with countdown signals that indicate the time left to crossing an intersection). Because all of these aids have some levels of errors (either false alarms or misses), driver compliance with the warnings, on the one hand, and reliance on the system, on the other hand, is an issue that has to be resolved empirically for each specific system (Maltz & Shinar, 2004). Consequently, we cannot assume that partial automation of the driving taskdmostly in the form of assistive information and control devicesdis inherently beneficial. Two notable examples are the antilock braking system (ABS) and electronic stability control (ESC). ABS has not been conclusively proven effective in reducing car crashes, despite its demonstrated effectiveness in controlled testing (Farmer, 2001). However, it does seem to be effective in reducing motorcycle fatal crashes (Teoh, 2010). In contrast, ESC has been shown to be very effective (Lie, Tingvall, Krafft, & Kullgren, 2006), and it is now mandatory in all passenger cars in the United States and European Union countries.
3.2. Decision Making and Memory
The separation of the driverevehicleeenvironment system into its three major components is a convenient way to organize the discussion of the various ways that HFE is applied to safety improvements. However, because we are dealing with an integrated system, many approaches to improving driver performance involve more than one component. To illustrate, new and emerging technologies to improve hazard detection, to increase vehicle control, and to shape driver behavior are applied to the interactions between the driver and the roadway and the driver and the vehicle, and often to the interaction among all three
Once a driver attends to and perceives some information, he or she must then decide what to do with it (see Figure 15.4), and that decision is largely dependent on the driver’s decision-making skills and past experience (memory). These decisions are most explicitly detailed in the various hierarchical models (as noted in the previous discussion of the alternative models), and their outcomes can then be described in terms of the time needed to make the decisions and their correctness or errors (see Figure 15.5). The limits
3.3. Response Properties: Anthropometry and Age
202
components. However, to avoid redundancies, the following discussion is organized by the primary component that is addressed by each approach.
4. VEHICLE VARIABLES AFFECTING THE DRIVEReVEHICLE INTERACTION Traditionally, issues of fitting the car to the driver focused on the drivers’ dimensions (anthropometry) and physical abilities (biomechanics). However, in the past three decades, HFE considerations in vehicle and road design have focused primarily on the drivers’ sensory and cognitive limits and capabilities and how they relate to various safety and communications technologies. Thus, HFE issues are at the center of the evaluation of current advanced driver assistance systems (ADAS); new in-vehicle information systems (IVIS) that are designed to support the driver in an appropriate, user-oriented way; and various infotainment systems, such as navigation and cell phones and e-mail communications, that are an increasing source of driver distraction. The primary orientation is usercentered design: optimizing performance through the satisfaction of human needs and performance limitations. These can determine the technical requirements, the usability of ADAS and IVIS, the operability of humane machine interface, behavioral adaptation and risk compensation, acceptance of innovations, and social impacts. To bring some order to the issue, various standards and guidelines for vehicle design include some HFE recommendations. No attempt is made to deal with all of these, but they are discussed briefly where they relate to safety. The objective is to acquaint the reader with the basic vehicle design features that affect driver performance and traffic safety. Thus, the following sections deal with select in-vehicle technologies, cockpit design, field of view and mirrors, intervehicle and driverevehicle communications, and automatic versus manual gear.
4.1. In-Vehicle Technologies Information and communication technologies are increasingly in use on the roads in so-called “intelligent transport systems” (ITS) and in safety-oriented electronic systems (e-Safety). Many of the current ITS applications focus on means to improve road users’ safety, and when the focus is elsewheredfor example, on comfort and traffic managementdthe implications for safety still have to be considered. In general, ITS should support the driving task so that fewer errors will be made and unsafe behavioral choices will be avoided. ITS have different functions: (1) to provide the driver with time-, situation-, and location-dependent information; (2) to provide warnings; and (3) to physically intervene
PART | II
Key Variables to Understand in Traffic Psychology
with the vehicle control in critical situations. ITS can also be classified based on their technical aspects: Vehicle systems with no outside interaction; roadside systems without interaction with vehicles; and the most “intelligent” categorydsystems with interaction between individual vehicles and other data sources, such as between vehicles or between the vehicle and the roadway. These systems can provide the latest situational information to an individual driverdfor example, weather conditions, temporary speed limits, or hazardous situations farther along the road. Some systems prevent unsafe driving in advance. Examples include alcohol ignition interlocks that prevent drivers from starting their cars if their blood alcohol content exceeds the legal (or other preset) limit and seat belt interlocks that warn drivers if seat belts are not fastened when the motor is running (or even intervene with the driving). Many cars already have a warning system, but the intervention systems go one step further by actually preventing driving in an unsafe mode. The second category of systems is those that prevent unsafe situations or actions while driving. Examples include systems that (1) offer support for vehicle control, such as ABS, ESC, adaptive cruise control (ACC), and lane-keeping support; (2) record and/or prevent intentional and unintentional errors, such as Intelligent Speed Adaptation; (3) offer support in observing, interpreting, and predicting traffic situations, such as forward collision warning; and (4) react to (temporarily) reduced driver capabilities, such as driver drowsiness monitoring and warning. Although in general the potential safety effects of ITS applications are large, the ultimate effects are definitely smaller than expected. This is because initial effectiveness assumes that “all other things remain the same,” whereas in fact everything we have learned of human behavior and the concept of the human-in-the-system indicates that when any part of the system changes, all other parts may change as welldespecially the human component. Possible unintended negative side effects of ITS and e-Safety can lead to the various types of errors discussed previously (see Figure 15.5) and include the following: l
l
Underload and diminished attention level: A common assumption (not always correct, however) is that road safety will benefit from reductions in mental workload on the driver. However, when driving tasks are partly replaced by ITS, the level of stimulation decreases to a point at which the driver is in an underload situation that is characterized by reduced attention (e.g., when fatigued or when driving on a very monotonous road). Information overload: ITS can also create additional information that the driver has to or chooses to deal with. For example, navigation systems can also provide
Chapter | 15
l
l
l
Human Factors and Ergonomics
information about speed, points of interest along the way, time and distance to arrival at target location, etc. All of these can attract attention and impair the driver’s ability to react to dangerous situations that are associated with rapid increases in information (e.g., when a car ahead suddenly stops). Thus, a good ITS requires judicious decisions concerning when and what information to present to the driver. Green (1999) recommended the use of a “15-second rule” for assessment of risks in in-vehicle navigation systems. The rule, which has been adopted by the American Society of Automotive Engineers, stipulates that a task that requires more than 15 seconds to perform while the vehicle is parked should be considered hazardous to driving. Note that this is a human factors performance-based recommendation rather than a vehicle design-based recommendation, acknowledging the value of human factors considerations as a guideline for vehicle-related design. Incorrect interpretation of information: The driver must be able to understand what the system does and what it means. The wrong interpretation of information can have an opposite than intended effect. For example, ABS was designed to improve braking and control performance provided that the driver also adopted a new emergency braking strategydone of continuous hard braking instead of the pumping action in which most drivers on the road today were trained. Thus, assuming that the ABS is a totally vehicle-based system and not one requiring a certain change in the driver’s behavior leads to less efficient rather than more efficient braking. Overreliance on the system: The driver’s expectations of a system must be realistic. Because no assistive system is 100% correct, complete trust can lead to errors and other inappropriate behaviors (e.g., not paying attention to the road). For example, overreliance on an obstacle detecting system when reversing may cause a fatal accident if a small child is behind the car because these systems often fail to detect small objects. Risk compensation: There is evidence that when the system’s benefits are obvious to drivers, drivers are often inclined to compensate for the reduced risk by taking risks they would not otherwise take (e.g., see Wilde’s (1982) theory of risk homeostasis). This can result in a smaller net effect, zero benefit, or even actual increased risk. Obvious examples of risk compensation are the tendency to increase speed on divided highways compared to two-lane roads and the tendency to engage in distracting tasks (e.g., dialing a phone number) in low-traffic situations. However, the issue is whether the drivers manage to retain the same risk level under all situations. Thus, per mile driven, even at the highest speed limits, most divided highways are still safer than
203
l
the two-lane rural roads that they replace. Drivers also increase their speeds when they drive with better tires on slippery (e.g., snowy) roads, but the net effect on crashes is still positive (Fridestroem, 2001). In contrast, ABS has not provided the actual safety benefitsdas measured in crashesdthat were predicted from experimental studies, and in some situations it actually seems to increase crashes (Delaney & Newstead, 2004). Effects on non-users: Especially when not all vehicles are equipped with a particular ITS system, it is possible that some drivers without such a system anticipate the supposed behavior of cars that do have that system. It is also possible that car drivers without such a system would behave as if they did have one (by mistake or imitation).
To improve safety, ITS, as with any application, should be designed in a manner consistent with the capabilities and behaviors of the range of people who will be using them. Furthermore, although the designer of the ITS may have a specific purpose in mind, realization of that objective through design should be consistent with the driver’s task as the intended users understand it. Many ITS applications are not designed to improve safety, but they may still have a road safety effect. Examples include ACC, which keeps a constant speed unless it detects an obstacle ahead, in which case it automatically reduces a car’s speed to avoid a collision. ACC can have both positive and negative effects. Positive effects are obtained when ACC functions as planned, but negative effectsdor even disastrous effectsdcan occur on rare occasions when it fails while the driver’s attention is not focused on the road ahead (e.g., when a pedestrian darts into the road). Driver GPS-based navigation systems are designed to help drivers in their navigation tasks (maneuvering level in Michon’s model; see Figure 15.3), but they have both positive and negative effects on safety. They reduce the driver load by eliminating the need to read street signs and cues for directions, but they distract the driver from the road whenever the driver gazes at them. An analysis estimated both types of effects and concluded that in sum they provide a net safety benefit (TNO, 2007). Another obstacle to predicting the safety effects of a system is its level of penetration into the overall traffic system. In general, however, theoretical analyses, simulation studies, and field operational tests confirm that the effects can be sizeable (SWOV, 2008). An interesting analysis of the potential benefits of various in-vehicle safety systems (not all involving the driver) was conducted as a part of an EU project, eIMPACT, and it provided concrete, unified estimates of traffic and safety benefits in terms of percent fatalities that could be reduced by each of the candidate systems. The findings are presented in Figure 15.7, and they provide both the potential benefits
204
PART | II
Key Variables to Understand in Traffic Psychology
potential −18% −16% −14% −12% −10% −8%
−6%
−4%
−2%
year 2020 high penetration rate
0% Electronic stability control, ESC Lane keeping support MAPS&ADAS Dynamic speed adaptation due to weather conditions, obstacles or congestion Emergency braking SASPENCE - Safe speed and safe following eCall Driver drowsiness monitoring and warning Local danger warning Cooperative intersection collision warning Night vision warning Lane change assistant and warning Pre-crash protection of vulnerable road users Full speed range ACC Post crash warning Reversible lanes due to traffic flow
Estimated safety impacts on fatalities (%) of 16 IVS. “Potential” assumed 100% fleet penetration of systems; “Year 2020 high penetration rate” assumed a promoted penetration in 2020.
FIGURE 15.7 Estimated safety effects of intelligent in-vehicle systems based on 80% market penetration and full 100% penetration. Source: Reproduced with permission from Rama (2009).
(assuming 100% penetration) and the “high penetration rate” benefits (assuming 80% penetration). Note that the two are not always correlated. For example, ESC (also known as electronic stability program) is a very effective system, and there is a strong relationship between 100% use and high penetration. In contrast, a lane-keeping support system has a high potential benefit, but at 80% penetration, the expected benefit is very small.
4.2. Cockpit Design The primary focus in vehicle occupant packaging is the driver’s workstation. Implementing ergonomic methods into the design of the driver workspace and interface is essential to ensuring a safer, healthier, and more comfortable driving experience. It is important to consider both drivers’ limitations and capabilities (in terms of vision, hand reach, anthropometry, and force) and design constraints (e.g., available space, collision safety, aesthetic design, and size of components). The driver “package” usually refers to the spatial dimensions of the intended user population when considering locations and adjustment ranges of the steering wheel, the seat, and the pedals; the physical locations of controls and displays with which the driver interacts; and the visual
field afforded by the widows, windshield, and mirrors (Parkinson & Reed, 2006). In fitting the car to the driver, it is not enough to consider what the driver can handle or see; how the driver prefers to interact with the vehicle interior must also be considered. Drivers with similar anthropometric dimensions may have different preferences in optimal locations of controls and displays (e.g., some like to sit erect, and some like to slouch). Failure to include variability in preference that is not attributable to anthropometry can produce misleading design recommendations that do not take into sufficient account the need for adjustability (Garneau & Parkinson, 2009). The process of designing a new vehicle involves fulfilling a large number of requirements. A list of recommended guidelines regarding occupant packaging for vehicle interior designs is provided by the Society of Automotive Engineers (Jung, Cho, Roh, & Lee, 2009).
4.3. Field of View and Mirrors Visibility from inside vehicles can be either direct (the environment is observed directly) or indirect (the environment is observed through mirror reflection). Direct visibility is determined mainly by the sizes and locations of
Chapter | 15
Human Factors and Ergonomics
the windows: A wider field of view includes more environmental cues and hence increases the ability to detect potential hazards. Indirect visibility is affected by the width, reflectance, and types of mirrors. For example, convex mirrors provide a much wider field of view allowing detection of objects that would otherwise be missed, but they create distortions in distance due to minification of the objects viewed, leading to errors such as overestimating the distance from these objects (Dewar & Olson, 2002). ITS-based systems are also improving the field of view with cameras to provide video views of difficult-to-see locations, such as immediately behind and in front of buses and trucks where pedestrians are otherwise invisible (Bota & Nedevschi, 2008; Nedevschi, Bota, & Tomiuc, 2009).
4.4. Intervehicle and DrivereVehicle Communications Icons can be used to communicate information to the driver in a language-free and space-efficient manner. Incomprehensible icons have the potential to affect safety (Campbell et al., 2004). Icons are generally preferred to text messages because there are no language barriers and a driver licensed to drive in any country can therefore drive almost any car in another country. However, icons are problematic when their meaning is not immediately apparent to all users. Icons representing rare events may not be comprehensible to all drivers; therefore, in some situations, text labels are preferable. In general, comprehension is improved when an icon is highly familiar, its use is standardized, and its graphics are compatible with the content it conveys (e.g., the image of a person with a shovel to indicate a work zone) (Ben-Bassat & Shinar, 2006). When these conditions are not met, it may be better to use a text label, despite its obvious limitations. Using the previously discussed principles makes most of the displays on car dashboards relatively easy to understand. However, with increasing possibilities of displaying all kinds of information, increasingly complex and crowded displays can compromise safety by distracting the driver from the road and the traffic. For example, the use of air-conditioning was traditionally associated with a single button. Today, there are multiple gages providing climate controls in different parts of the car. The basic idea behind ergonomically “good” in-car warning technology is to provide drivers with information that is otherwise not directly perceivable. Thus, there are indicators relating to the operational state of the vehicle (e.g., fuel level, oil pressure, and water temperature) and to the performance of the vehicle (e.g., speed). In the past, this information was typically displayed either in the form of (hard) dials or colored lights. Currently, much more sensitive and exact information on many more vehicle functions is displayed via digital soft displays but at the cost of more complicated
205
program controls. For critical information, some of the visual information is also augmented with auditory feedback, such as extreme low fuel and seat belt reminders. A difficulty encountered by all drivers is detecting passing and overtaking vehicles while they are in the driver’s “blind zones”dthose zones that are outside the forward field of view afforded by the windshield and front windows and the rearview mirrors. A recent aid is an alerting system that notifies the driver if there is a nearby vehicle in the adjacent lane whenever the driver signals an intention to switch lanes. A critical human factors issue is the nature of this warning. Should it be auditory or visual, and if it is visual, where should it be displayed. An auditory system that is gaining popularity is Mobileye (Mobileye, 2010). Reed and Flannagan (2003) suggested placing such a warning along with turning signals on the outside rearview mirrors. They found that it is more visible than conventional turn signals. Furthermore, these signals are closer to the driver’s line of sight when his or her vehicle is in or near the blind zone (Reed & Flannagan, 2003). Sivak, Schoettle, and Flannagan (2006) evaluated the effects of mirror-mounted turn signals on the frequency and the severity of turn signal-related crashes. The results indicated a tendency for vehicles with mirror-mounted turn signals to be less likely involved in relevant crashes, but the effect was not significant and could have been confounded by other safety-related factors. Furthermore, there are doubts regarding the possibility of a reduction in crash severity for vehicles with mirror-mounted turn signals.
4.5. Automatic Versus Manual Gears During the twentieth century, the automobile evolved from a humble horseless carriage into one of the most technologically advanced mass market ubiquitous humane machine systems. Currently, we are starting to see a new generation of technologies enter the vehicle in the form of driving automation. Of course, vehicle automation has been around for some time, with electric starters and automatic gearboxes becoming widely available in the first half of the previous century. However, whereas traditional automatic systems sought to take over the lower level, operational components of vehicle control (see Figure 15.3), new technologies are taking over more tactical and even strategic aspects of driving (Ranney, 1994). Thus, Stanton and Young (2000) distinguish between two types of automation: vehicle automation that addresses low-level vehicle operations (vehicle control) and driving automation that addresses tactical (maneuvering) and strategic-level functions. An immediate by-product of automation is a change in the driver information processing load, often reducing the cognitive costs of driving. Shinar, Meir, and Ben-Shoham (1998) had novice and experienced drivers drive their own
206
cars along a previously selected course of urban streets and requested them to note whenever they saw a “Slowd Children” or a “No Stopping” sign. They found that the novice drivers with a manual gearshift detected significantly fewer signs (of both types) than novice drivers with automatic gearshift, whereas experienced driversdwho in general detected more signsddid not manifest this difference between the two gearboxes. Automation is gradually taking over the driver’s role, with some going so far as to predict full vehicle automation for British roads by 2030 (Walker, Stanton, & Young, 2001). Given the state-of-the-art today, driver needs (some people like to have control and choose manual gears), the highly dynamic environment of driving (a pedestrian may dart into the road), and the high stakes involved (people can get killed), we believe that this is highly unlikely. Also, although such a shift can reduce the driver’s load (as in the gearbox example), assuming the driver does not stay at home while the car goes shopping, it can also fundamentally change the driver’s role from that of controller to that of a monitor. This in turn creates new difficulties, most notably that of situation awareness (Endsley, 1995)d a situation in which the driver’s world differs from that of the automated system. In the context of driving, although ITS may behave in exactly the manner prescribed by the designers and programmers, it may lead to some scenarios in which the driver’s perception of the situation or the actual but unanticipated reality is at odds with the system operation (Stanton & Young, 2000).
5. ENVIRONMENTAL VARIABLES The road infrastructure conveys a wealth of informationdsuch as road signs and markings and implicitly by means such as road layoutdthat guides drivers’ activities. Although there are circumstances in which drivers have to react to some unexpected event, usually drivers execute planned actions that are shaped by their expectations of the road, traffic scenarios, and the reality they actually face. The role of the environment is secondary to human factors in the causes of accidents, but it is much more significant than that of vehicles (see Figure 15.1). Treat et al. (1977) found that view obstructions are the most frequent environmental cause of accidents, followed by slick roads. In an analysis of the causes of accidents (National Highway Traffic Safety Administration, 2008), the environment was listed as a critical reason for the crash of 16% of the vehicles. In these crashes, slick roads were listed in approximately 50%, glare was listed in 16.4%, and the weather was listed in 8.4%. In a naturalistic driving study, Dingus et al. (2006) monitored the crash involvement of 100 cars during a period of approximately 1 year. Although the number of accidents was low, the detailed monitoring of the vehicles
PART | II
Key Variables to Understand in Traffic Psychology
provided insights into the specific crashes that are not otherwise available in post-crash analysis. They defined 18 types of conflicts (e.g., conflict with either a lead, adjacent, or following vehicle, single-vehicle conflict, and conflict with an obstacle or with another road user), with each describing the factors that precipitated, contributed to, and were associated with the event. The environment was classified in terms of its static or dynamic nature, with the former consisting of the infrastructure and the latter consisting of the changing driving environment (including traffic, visibility, and weather). The infrastructure category includes the factors that are fixed and do not change with the environment: (1) Trafficway flowdone way and divided roadway; (2) traffic control devicedsignals and signs; (3) localitydinterstate and residential areas; (4) roadway alignmentdstraight, curve, level, or hillcrest; and (5) relation to junctiond intersection and entrance/exit ramp. The driving environment category consists of conditions that change on a daily or hourly basis: (1) surface conditiondwet or snowy; (2) lightingdstreetlamps or daylight; (3) traffic densitydstable flow and limited speed or flow; and (4) atmospheric conditionsdclear or rainy. For example, results for the single-vehicle crashes revealed that infrastructure and driving environment (e.g., weather and visibility, roadway alignment, and roadway delineation) were contributing factors in 29% of the crashes, in 23% of the single-vehicle near crashes, and in only 10% of the incidents. In the case of “lead-vehicle crashes,” when an interaction occurred between the subject vehicle and the vehicle directly in front of it, the environmental factors were not judged to be a strong contributing factor, with only one crash being due to weather and visibility. This is somewhat surprising considering that more than 40% of the crashes included inclement weather and wet or snowy surface conditions. Not surprisingly, traffic flow was fairly strongly associated with the lead-vehicle crashes and near crashes. The infrastructure associated with the crashes and near crashes was straight and level in most of the crashes (87%). Regarding the lead-vehicle incidents, none of the driving environment factors were identified as contributing, and only one crash infrastructure factor (i.e., roadway delineation) was identified as contributing. Roadway alignment may have played a role, with 42% of the crashes being on curves. Two-thirds of the crashes were intersection related. Inadequate road infrastructure or poor environmental conditions can potentially impact road user behavior and performance in a way that can potentially lead to road user errors. These inadequate conditions include confusing layout, misleading signage, poor road surface-related conditions, poor weather conditions, poor lighting conditions, time of day, and misleading or inappropriate rules and regulations (Stanton & Salmon, 2009).
Chapter | 15
Human Factors and Ergonomics
5.1. Roadway Design Driver workload can also be affected by the roadway design. Therefore, road design should meet driver expectations and take into account the abilities and limitations of all road users (Dewar, 2008). Long sight distances and divided highways with wide, high shoulders that enable high speeds ease the driver workload, whereas congested roads, frequent curves, narrow bridges, short sight distances, and in general situations that increase uncertainty increase workload. Here, too, current practices are designed with these goals in mind. Good examples are those of positive road guidance and self-organizing roads. The positive guidance approach enhances the safety of hazardous locations by providing drivers with direct visual cues of how to behave. For example, delineation and analogous curve signs that visually display the amount of curvature are preferable to text-based signs. This approach combines the highway engineering and human factors technologies to produce an information system matched to the characteristics of the location and the attributes of drivers (Alexander & Lunenfeld, 1979). The self-organizing roads are design principles that increase the probability that a driver will “automatically” select appropriate speed and steering behavior for the roadway without depending on road signs or enforcement. The geometric features of the road encourage the desired driver behavior, and thus compliance does not rely on the driver’s ability or willingness to read and obey road signs. A roundabout is a self-organizing road. The road geometry forces the driver to slow down. Pavement markings help the driver perceive this lower speed requirement. Also, intentionally narrowing the roadway and shoulders creates self-organizing features that instruct the driver to slow down. When there is a conflict between road features and road signs, drivers may often follow the speed implied by the roadway design rather than the speed instructed by the road sign. Another important example of a self-organizing road is the 2 þ 1 roadway, which is a three-lane road with the passing lane alternating on each side of the road in a regular manner. This organizes the driver’s expectations about being able to pass. These roads also work well for speed management. Studies of human reaction time helped formulate standards for geometric roadway design (Keith et al., 2005).
5.2. Roadway Illumination Roadway lighting is designed, fabricated, and installed for the public benefit at night. It is difficult to quantify its value because it rests not only on its concrete implementation and operation costs but also on its expected benefits, which are inherently difficult to estimate. Most studies have revealed that roadway lighting provides positive safety and security
207
benefits (i.e., crash prevention) for drivers and pedestrians (Rea, Bullough, Fay, Brons, & Van Derlofske, 2009; ROSPA, 2007). Roadway lighting has been suggested as a relatively low-cost intervention with the potential to prevent traffic crashes by improving drivers’ visual capabilities and ability to detect roadway hazards and by reducing contrast between headlight glare and the surrounding environment. However, it is also argued that roadway lighting could have an adverse effect on road safety; drivers may “feel” safer because lighting gives them improved visibility, which could result in drivers increasing speed and reducing concentration (Beyer & Ker, 2009). Hogema, Veltman, and Van’t Hof (2005) measured the effects of variations in motorway lighting on driver behavior and concluded that when the lighting was switched off, mental effort increased because heart rate and blink rate increased and speed decreased. When drivers have to deal with higher workload (i.e., task demand), they often have two options: They can increase their effort or they can try to reduce the task load, for example, by reducing speed. Thus, drivers have more time available to anticipate potential hazards, and this reduces the workload. The potential safety benefit of roadway lighting can be summarized as follows: l
l
l
Darkness (or the absence of lighting) results in an excessively large number of crashes and fatalities, particularly those involving pedestrians. Lighted intersections and interchanges tend to have fewer crashes than unlighted ones. Visibility is primarily associated with crashes involving pedestrians and intersections at night. Thus, in these scenarios, the roadway lighting might have the greatest effect with regard to the reduction of nighttime crashes.
5.3. Traffic Control Devices: Road Markings, Signs, and Signals Understanding drivers’ comprehension and predictive behavior of traffic control devices is critical. If the driver does not understand the message being presented, then his or her response may vary significantly and affect safety. Intersection traffic control devices are composed of signs, signals, roundabouts, or pavement markings that can be placed alongside the intersection. They are used to move vehicles and pedestrians safely and efficiently, consequently preventing collisions by providing the “rightof-way” principle assignment. The most extensively used devices for current traffic control include traffic signals, stop signs, and roundabouts. Improper placement of a traffic control device may decrease its efficiency. If drivers recognize the signal too late to safely react to the situation, an increase in the
208
number of accidents at the intersection may occur. One such example is placing the device too closely around the bend of a sharp curve. Catastrophic results will occur when drivers fail to stop in time. The primary goal of all traffic control devices is to maintain the safety of the drivers advancing through the intersection. Some conventional devices currently in use have significant shortcomings that can actually hinder safety. Traffic signal lights are one such example. When configuring a signal light, an engineer must be careful with regard to the duration of the green, red, and amber (yellow) phases. An extensively studied issue is that of the yellowlight dilemma. If the yellow phase is too short, drivers might have to slam on their brakes to avoid crossing the intersection before the light turns red. This could and does cause an increase in rear-end collisions. On the other hand, if the yellow light is on for too long, drivers might ignore it and speed up to cross the intersection, and this candand doesdincrease the likelihood of side crashes in the intersection (which are typically more severe than rear-end crashes) (Fadi & Hazem, 2009). The duration of the green and red phases also appears to affect driving behavior: Drivers tend to run red lights more when the green phase is short and the red phase is long than when the green phase is long and the red phase is short (Shinar, 1998), and they also tend to do so more when the green phases in consecutive intersections are not synchronized than when they are (Shinar, Bourla, & Kaufman, 2004).
6. CONCLUSION The aim of this chapter was to provide a brief understanding of the approaches to driverevehicle modeling and the impact of driver, vehicle, and roadway characteristics on highway traffic safety. As vehicles have evolved, there has been a gradual shift from a vehicle-centered approach to a user-centered approach to designing vehicles and roadways. This requires a scientifically based understanding of the motivations, capabilities, and limitations of people in operating their vehicles and in correctly perceiving the roadway and traffic demands. The need for this understandingdexpressed in models of driver behaviordis underscored by the fact that the overwhelming majority of traffic accidents involve a human error, most often a misperception of the situation or inappropriate human decision or action. Here, “to err is human” is not an empty statement. However, the role of human factors in highway safety is not simply to identify the human failings and recommend ways to correct them but, rather, to identify the elements in the driverevehicleeroadway that give rise to these errors. Thus, the aim is not to train drivers not to err but, rather, to prevent the situations that give rise to errors. In this chapter,
PART | II
Key Variables to Understand in Traffic Psychology
we discussed some of the areas in which human factors research has contributed to this goal. Because of space constraints and potential overlap with other chapters, many areas were excluded in this discussiondsuch as the effects of impaired or distracted driving, age, and individual differencesdbut the reader can find more about these areas in the other chapters. More information on human factors in highway safety can also be found in Dewar and Olson (2007) and Shinar (2007).
REFERENCES Alexander, G. J., & Lunenfeld, H. (1979). A users’ guide to positive guidance in highway control. Human Factors and Ergonomics Society Annual Meeting Proceedings, 23(6), 452e455. Ben-Bassat, T., & Shinar, D. (2006). Ergonomic guidelines for traffic sign design increase sign comprehension. Human Factors, 48(1), 182e195. Beyer, F. R., & Ker, K. (2009). Street lighting for preventing road traffic injuries. Cochrane Database of Systematic Reviews CD004728. Bhise, V. D., & Rockwell, T. H. (1973). Toward the development of a methodology for evaluating highway signs based on driver information acquisition. Highway Research Record, 440, 38e56. Borowsky, A., Shinar, D., & Oron-Gilad, T. (2010). Age, skill, and hazard perception in driving. Accident Analysis and Prevention, 42, 1240e1249. Bota, S., & Nedevschi, S. (2008). Multi-feature walking pedestrians detection for driving assistance systems. Intelligent Transport Systems, 2(2), 92e104. Cacciabue, P. C., Mauri, C., & Owen, D. (2003). The development of a model and simulation of aviation maintenance technician task performance. International Journal of Cognition, Technology & Work, 5, 229e247. Campbell, J. L., Hofmeister, D. H., Kiefer, R. J., Selke, D. J., Green, P., & Richman, J. B. (2004). Comprehension testing of active safety symbols. (Paper No. 2004-01-0450). Warrendale, PA: SAE International. Carsten, O. (2007). From driver models to modelling the driver: What do we really need to know about the driver? In P. C. Cacciabue (Ed.), Modelling driver behaviour in automotive environments (pp. 105e120) London: Springer. Chapanis, A., Garner, W. R., & Morgan, C. T. (1949). Applied experimental psychology. New York: Wiley. Delaney, A., & Newstead, S. (2004). The effectiveness of anti-lock brake systems; A statistical analysis of Australian data. Proceedings road safety research, policing and education conference. Perth: Road Safety Council of Western Australia. Delorme, D., & Song, B. (2001). Human driver model for SmartAHS. (California PATH Research Report No. UCB-ITS-PRR-2001-12). Berkeley: University of California Press. Dewar, R. E. (2008). The role of human factors in road safety audits. In Proceedings of the 18th Canadian multidisciplinary road safety conference. British Columbia: Whistler. June 8e11, 2008. Dewar, R. E., & Olson, P. L. (2002). Human factors in traffic safety. Tucson, AZ: Lawyers & Judges. Dewar, R. E., & Olson, P. L. (2007). Human factors in traffic safety (2nd ed.). Tucson, AZ: Lawyers & Judges.
Chapter | 15
Human Factors and Ergonomics
Dingus, T. A., Klauer, S. G., Neale, V. L., Peterson, A., Lee, S. E., Sudweeks, J., Perez, M. A., Hankey, J., Ramsey, D., Gupta, S., Bucher, C., Doerzaph, Z. R., Jarmeland, J., & Knipling, R. R. (2006). The 100-car naturalistic driving study, phase II: Results of the 100-car field experiment. Washington, DC: National Highway Traffic Safety Administration. Endsley, M. (1995). Toward a theory of situation awareness in dynamic systems. Human Factors, 37(1), 32e64. ¨ stlund, J. (2005). Effects of visual and Engstro¨m, J., Johansson, E., & O cognitive load in real and simulated motorway driving. Transportation Research F, 8, 97e120. Fadi, B., & Hazem, H. R. (2009). Collision avoidance system at intersections. (Report No. FHWA-OK-09e06). Edmond: Oklahoma Department of Transportation. Farmer, C. M. (2001). New evidence concerning fatal crashes of passenger vehicles before and after adding antilock braking systems. Accident Analysis and Prevention, 33(3), 361e369. Fastenmeier, W., & Gstalter, H. (2007). Driving task analysis as a tool in traffic safety research and practice. Safety Science, 45, 952e979. Fridestroem, L. (2001). The safety effect of studded tyres in Norway (Report No. TOI-RAP- 493/2000). Oslo: Norwegian Institute of Transport Economics. Fuller, R. (1984). A conceptualization of driving behavior as threat avoidance. Ergonomics, 11, 1139e1155. Fuller, R. (2005). Towards a general theory of driver behaviour. Accident Analysis and Prevention, 37, 461e472. Garneau, C. J., & Parkinson, M. B. (2009). Including preference in anthropometry-driven models for design. ASME Journal of Mechanical Design, 131(10). Gibson, J. J. (1966). The senses considered as perceptual systems. Boston: Houghton Mifflin. Green, P. (1999). The 15-second rule for driver information systems. In Proceedings of the ITS America ninth annual meeting, Washington, DC (CD-ROM). Hoffman, J. E., & Subramaniam, B. (1995). The role of visual attention in saccadic eye movements. Perception and Psychophysics, 57, 787e795. Hogema, J. H., Veltman, H. A., & Van’t Hof, A. (2005). Effects of motorway lighting on workload and driving behaviour. In G. Underwood (Ed.), Traffic and transport psychology (pp. 355e368). Oxford: Elsevier. Hollnagel, E., Na˚bo, A., & Lau, I. (2003). A systemic model for driverin-control. In Proceedings of the second international driving symposium on human factors in driver assessment, training, and vehicle design, Park City, UT. Iowa City, IA: University of Iowa, Public Policy Center. Iyinam A.F., Iyinam, S., & Ergun, M. (1997). Analysis of relationship between highway safety and road geometric design elements: Turkish case. Jung, M., Cho, H., Roh, T., & Lee, K. (2009). Integrated framework for vehicle interior design using digital human model. Journal of Computer Science and Technology, 24(6), 1149e1161. Kahneman, D., & Treisman, A. (1984). Changing views of attention and automaticity. In R. Parasuraman, & D. R. Davis (Eds.), Varieties of attention. Orlando, FL: Academic Press. Kantowitz, B. H. (2000). In-vehicle information systems: Premises, promises, and pitfalls. Transportation Human Factors, 2(4), 359e379. Kantowitz, B. H., Singer, J. P., Lerner, N. D., McGee, H. W., Hughes, W. E., Perez, W. A., & Ullman, G. L. (2004). Development
209
of critical gaps and knowledge base in support of the safety R & T partnership agenda. (Contract No. DTFH61-01-C-00049, Report to the Federal Highway Administration). Ann Arbor: University of Michigan Transportation Research Institute. Keith, K., Trentacoste, M., Depue, L., Granda, T., Huckaby, E., Ibarguen, B., Kantowitz, B., Lum, W., & Wilson, T. (2005). Roadway human factors and behavioral safety in Europe. (Publication No. FHWA-PL-05-005). Washington, DC: U.S. Department of Transportation. Kiger, S. M., Rockwell, T. H., & Tijerina, L. (1995). Developing baseline data on heavy vehicle driver visual workload. In Proceedings of the Human Factors and Ergonomics Society, 39th annual meeting (pp. 1112e1116). Santa Monica, CA: Human Factors and Ergonomics Society. Lee, J. D. (2008). Fifty years of driving safety research human factors. Journal of the Human Factors and Ergonomics Society, 50(3), 521e528. Lie, A., Tingvall, C., Krafft, M., & Kullgren, A. (2006). The effectiveness of electronic stability control (ESC) in reducing real life crashes and injuries. Traffic Injury Prevention, 7(1), 38e43. Lindgren, A., Brostro¨m, R., Chen, A., & Bengtsson, P. (2007). Driver attitudes towards advanced driver assistance systemsdCultural differences and similarities. In D. de Waard, G. R. J. Hockey, P. Nickel, & K. A. Brookhuis (Eds.), Human factors issues in complex system performance (pp. 205e215). Maastricht, The Netherlands: Shaker. Maltz, M., & Shinar, D. (2004). Imperfect in-vehicle collision avoidance warning systems can aid drivers. Human Factors, 46, 357e366. McKnight, A. J., & Adams, B. B. (1971). The development of driver education objectives through an analysis of the driving task. Chicago: Paper presented at the National Safety Congress. October, 1970. McKnight, A. J., & McKnight, A. S. (2003). Young novice drivers: Careless or clueless? Accident Analysis and Prevention, 35, 921e925. McRuer, D. T., Allen, R. W., Weir, D. H., & Klein, R. (1977). New results in driver steering control models. Human Factors, 17, 381e397. McRuer, D. T., & Weir, D. H. (1969). Theory of manual vehicular control. Ergonomics, 12(4), 599e633. Michon, J. A. (1985). A critical view of driver behaviour models. What do we know, what should we do? In L. Evans, & R. Schwing (Eds.), Human behaviour and traffic safety (pp. 485e525) New York: Plenum. Michon, J. A. (Ed.). (1993). Generic intelligent drive support. London: Taylor & Francis. Mobileye. (2010). Mobileye pedestrian protection. http://www.mobileye. com/sites/mobileye.com/files/brochures/PedestrianProtectionEnglish_0.pdf. Mourant, R. R., & Rockwell, T. H. (1970). Mapping eye-movement pattern to the visual scene in driving: An exploratory study. Human Factors, 12, 81e87. Mourant, R. R., & Rockwell, T. H. (1972). Strategies of visual search by novice and experienced drivers. Human Factors, 14, 325e335. Mourant, R. R., Rockwell, T. H., & Rackoff, N. J. (1969). Drivers’ eye movements and visual workload. Highway Research Record, 292, 1e9. Naatanen, R., & Summala, H. (1976). Road user behaviour and traffic accidents. New York: North-Holland.
210
National Highway Traffic Safety Administration. (2008). National motor vehicle crash causation survey. (Report to Congress No. DOT HS 811 059). Washington, DC: U.S. Department of Transportation. Nedevschi, S., Bota, S., & Tomiuc, C. (2009, September). Stereo-based pedestrian detection for collision-avoidance applications. IEEE Transactions on Intelligent Transportation Systems, 10(3). Norman, D. A. (1981). Categorization of action slips. Psychological Review, 88, 1e15. Owsley, C., Ball, K., McGwin, G., Sloane, M. E., Roenker, D. L., White, M. F., Overley, E., & Todd, O. D. (1998). Visual processing impairment and risk of motor vehicle crash among older adults. JAMA, 279, 1083e1088. Papafotiou, K., Stough, C. K., & Nathan, P. (2005). An evaluation of the sensitivity of the standardised field sobriety tests (SFSTs) to detect impairment due to marijuana intoxication. Psychopharmacology, 180, 107e114. Parasuraman, R., & Riley, V. (1997). Humans and automation: Use, misuse, disuse, abuse. Human Factors, 39(2), 230e253. Parkinson, M. B., & Reed, M. P. (2006). Optimizing vehicle occupant packaging (Technical Paper No. 2006-01-0961). SAE Transactions: Journal of Passenger Cars - Mechanical Systems. Rakotonirainy, A., & Maire, F. (2005). Context-aware driving behaviour model. In Proceedings of the 19th international technical conference on the enhanced safety of vehicles (ESV’19), Washington, DC. Rama, P. (2009). Safety effects of intelligent in-vehicle systems. (Nordic Road and Transport Research No. 3, p. 19). Linko¨ping, Sweden: VTI (Swedish National Road and Transport Research Institute). Ranney, T. A. (1994). Models of driving behavior: A review of their evolution. Accident Analysis and Prevention, 26(6), 733e750. Rasmussen, J. (1986). Information processing and humanemachine interaction: An approach to cognitive engineering. New York: Elsevier. Rayner, K. (1998). Eye movements in reading and information processing: 20 years of research. Psychological Bulletin, 124, 372e422. Rea, M. S., Bullough, J. D., Fay, C. R., Brons, J. A., & Van Derlofske, J. (2009). Review of the safety benefits and other effects of roadway lighting. Final report prepared for National Cooperative Highway Research Program Transportation Research Board of The National Academies. (Project No. 5-19). Washington, DC: Transportation Research Board of the National Academies. Reason, J. (1990). Human error. New York: Cambridge University Press. Reed, M. P., & Flannagan, M. J. (2003). Geometric visibility of mirrormounted turn signals. (Technical Report No. UMTRI-2003-18). Ann Arbor: University of Michigan Transportation Research Institute. Reingold, E. M., Loschky, L. C., McConkie, G. W., & Stampe, D. M. (2003). Gaze-contingent multiresolutional displays: An integrative review. Human Factors, 45, 307e328. Rockwell, T. H. (1988). Spare visual capacity in driving revisited. In A. G. Gale, M. H. Freeman, C. M. Haslegrav, P. Smith, & S. P. Taylor (Eds.), Vision in vehicles II (pp. 317e324). Amsterdam: NorthHolland. Royal Society for the Prevention of Accidents. (2007). Street lighting and road safety. Birmingham, UK: Author. Rumar, K. (1985). The role of perceptual and cognitive filters in observed behaviour. In L. Evans, & R. C. Schwing (Eds.), Human behaviour and traffic safety. New York: Plenum. Sabey, B.E., & Staughton, G. C. (1975). Interacting roles of road environment, vehicle and road user in accidents. Paper presented at the
PART | II
Key Variables to Understand in Traffic Psychology
5th International Conference on the International Association for Accident and Traffic Medicine, London. Salvucci, D. D. (2006). Modeling driver behavior in a cognitive architecture. Human Factors, 48(2), 362e380. Salvucci, D. D., Zuber, M., Beregovaia, E., & Markley, D. (2005). Distract-R: Rapid prototyping and evaluation of in-vehicle interfaces. Portland, OR: Paper presented at the Computer Human Interface annual conference. Schneider, W., & Shiffrin, R. M. (1977). Controlled and automatic human information processing: I. Detection search and attention. Psychological Review, 84. le54. Setti, J., Rakha, H., & El-Shawarby, I. (2007). Analysis of brake perceptionereaction times on high-speed signalized intersection approaches (CD-ROM). Paper presented at the Transportation Research Board 86th annual meeting. Washington, DC: Transportation Research Board, National Research Council. Sheridan, T. B. (1970). Big brother as driver: New demands and problems for the man at the wheel. Human Factors, 12, 95e101. Sheridan, T. B. (2004). Driver distraction from a control theory perspective. Human Factors, 46, 587e599. Shiffrin, R. M., & Schneider, W. (1977). Controlled and automatic human information processing: II. Perceptual learning, automatic attending and a general theory. Psychological Review, 84, 127e189. Shinar, D. (1978). Psychology on the road: The human factor in traffic safety. New York: Wiley. Shinar, D. (1998). Aggressive driving: The contribution of the drivers and the situation. Transportation Research, Part F: Psychology and Behavior, 1, 137e160. Shinar, D. (2001). Driving speed relative to the speed limit and relative to the perception of safe, enjoyable, and economical speed. In Proceeding of the conference on traffic safety on three continents, Moscow, Russia, September 19e21. Linko¨ping, Sweden: VTI (Swedish National Road and Transport Research Institute). Shinar, D. (2007). Traffic safety and human behavior. Oxford: Elsevier. Shinar, D. (2008). Looks are (almost) everything: Where drivers look to get information. Human Factors, 50(3), 380e384. Shinar, D., Bourla, M., & Kaufman, L. (2004). Synchronization of traffic signals as a means of reducing red-light running. Human Factors, 46, 367e372. Shinar, D., McDowell, E. D., & Rockwell, T. H. (1977). Eye movements in curve negotiation. Human Factors, 19, 63e72. Shinar, D., Meir, M., & Ben-Shoham, I. (1998). How automatic is manual gear shifting? Human Factors, 40, 647e654. Shinar, D., & Raz, S. (1982). Driver response to different railroad crossing protection systems. Ergonomics, 25(9), 801e808. Shinar, D., & Schieber, F. (1991). Visual requirements for safety and mobility of older drivers. Human Factors, 33, 507e519. Sivak, M., Schoettle, B., & Flannagan, M. J. (2006). Mirror-mounted turn signals and traffic safety. (Technical Report No. UMTRI-2006-33). Ann Arbor: University of Michigan Transportation Research Institute. Stanton, N. A., & Salmon, P. M. (2009). Human error taxonomies applied to driving: A generic driver error taxonomy and its implications for intelligent transport systems. Safety Science, 47(2), 227e237. Stanton, N. A., & Young, M. S. (2000). A proposed psychological model of driving automation. Theoretical Issues in Ergonomic Science, 1(4), 315e331.
Chapter | 15
Human Factors and Ergonomics
Stanton, N. A., Young, M. S., & Walker, G. H. (2007). The psychology of driving automation. International Journal of Vehicle Design, 45(3), 283e288. Stutts, J. C., Reinfurt, D. W., Staplin, L., & Rodgman, E. A. (2001). The role of driver distraction in traffic crashes. Washington, DC: AAA Foundation for Traffic Safety. Summala, H. (2000). Brake reaction times and driver behavior analysis. Transportation Human Factors, 2(3), 217e226. SWOV. (2008). Intelligent transport systems (ITS) and road safety. (fact sheet). Leidschendam, The Netherlands: Author. Teoh, E. R. (2010). Effectiveness of antilock braking systems in reducing motorcycle fatal crash rates. Fairfax, VA: Insurance Institute of Highway Safety. TNO. (2007). Key findings. Retrieved May 2010 from. http://www.tno.nl/ downloads/pb_2007_13_32324_tno_es_uk.pdf. Treat, J. R., Tumbas, N. S., McDonald, S. T., Shinar, D., Hume, R. D., Mayer, R. E., Stanisfer, R. L., & Castellan, N. J. (1977). Tri-level study of the causes of traffic accidents. (Indiana University Final Report No. DOT-HS-034-3-535-77-TAC). Washington, DC: National Highway Traffic Safety Administration. Victor, T. W., Harbluk, J. L., & Engstro¨m, J. A. (2005). Sensitivity of eyemovement measures to in-vehicle task difficulty. Transportation Research F, 8, 167e190.
211
Walker, G. H., Stanton, N. A., & Young, M. S. (2001). Where is computing driving cars? International Journal of Human Computer Interaction, 13(2), 203e229. Warshawsky-Livne, L., & Shiner, D. (2002). Effects of uncertainty, transmission type, driver age and gender on brake reaction and movement time. Journal of Safety Research, 33, 117e128. Weller, G., Schlag, B., Gatti, G., Jorna, R., & Leur, M.v.d. (2006). Human factors in road design state of the art and empirical evidence. (Report No. 8.1, RiPCORD-iSEREST). Brussels, Belgium: European Commission. Wickens, C. D. (1992). Engineering psychology and human performance (2nd ed.). New York: HarperCollins. Wickens, C. M., Toplak, M. E., & Wiesenthal, D. L. (2008). Cognitive failures as predictors of driving errors, lapses, and violations. Accident Analysis and Prevention, 40, 1223e1233. Wilde, G. J. S. (1982). The theory of risk homeostasis: Implications for safety and health. Risk Analysis, 2, 209e225. Young, M. S., Stanton, N. A., & Harris, D. (2007). Driving automation: Learning from aviation about design philosophies. International Journal of Vehicle Design, 45(3), 323e338. Zhang, H., Smith, M. R. H., & Witt, G. J. (2006). Identification of real-time diagnostic measures of visual distraction with an automatic eyetracking system. Human Factors, 48, 805e821.
This page intentionally left blank
Part III
Key Problem Behaviors 16. 17. 18. 19.
Factors Influencing Safety Belt Use Alcohol-Impaired Driving Speed(ing): A Quality Control Approach Running Traffic Controls
215 231 249 267
20. Driver Distraction: Definition, Mechanisms, Effects, and Mitigation 21. Driver Fatigue
275 287
This page intentionally left blank
Chapter 16
Factors Influencing Safety Belt Use Jonathon M. Vivoda and David W. Eby University of Michigan Transportation Research Institute, Ann Arbor, MI, USA
1. INTRODUCTION In 1763, the Frenchman Nicolas Joseph Cugnot invented the first “automobile” (Chambers, 1902; Sinclair, 2004). It was self-propelled using steam actuation and could travel only for approximately 10e15 min before stopping to build up more steam (Sinclair, 2004). Cugnot’s invention was difficult to control, and in 1771 he crashed it into a stone wall, creating the first recorded automobile crash in history (Figure 16.1; Bellis, 2010; Sinclair, 2004). As you might imagine, Cugnot’s vehicle was not equipped with safety beltsdit was not even equipped with brakesdbut luckily for him, it could travel only slightly faster than 2 miles per hour, and no one was hurt in the crash. Although there were no injuries in this very first crash, this story illustrates that even from the earliest development of automobiles, there has been a need for safety devices to help keep motorists safe. The seat belt, or safety belt, was invented in the 1800s by George Cayley, and the first patent was registered to Edward J. Claghorn in 1885 (Carter & Maker, 2010; Kett, 2009). One of the most important safety devices ever invented, its early application was primarily for use in aircraft to keep pilots in their seats during maneuvers. In the
1930s, a group of physicians in the United States recognized the potential value of safety belts in motor vehicles and began to urge automobile manufacturers to install them in cars as standard equipment. These doctors felt so strongly about this cause that they even took it upon themselves to equip their own cars with safety belts (School Transportation News, 2010). It was not until the 1950s that a few vehicle manufacturers began to include safety belts in vehicles, and during the 1950s and 1960s, several states began to require vehicles to have safety belt anchors installed (Kahane, 2004; School Transportation News, 2010). In 1968, a U.S. federal standard went into effect requiring auto manufacturers to include safety belts in all new vehicles (Kahane, 2004; National Highway Traffic Safety Administration (NHTSA), 1998).3
2. EFFECTIVENESS Today, safety belts are standard equipment for every seating position in every vehicle produced for sale in the United States. Since their early development, researchers have continued to assess the effectiveness of safety belts and make design improvements. Early safety belt designs FIGURE 16.1 Depiction of Nicolas Joseph Cugnot crashing his automobile into a wall. Source: Wikipedia.
Handbook of Traffic Psychology. DOI: 10.1016/B978-0-12-381984-0.10016-5 Copyright Ó 2011 Elsevier Inc. All rights reserved.
215
216
reduced injuries compared to riding unbelted, but they also had some serious weaknesses. For example, the first designs typically included only one belt strap (EDinformatics, 2010). Designs with a strap that only went across the waist (lap belts) left a motorist’s upper torso vulnerable to injury, and designs with only a sash (a strap that went over the shoulder) sometimes resulted in motorists sliding forward under the sash belt during a collision. A more effective design combined the two into one belt (often called a three-point belt), a variation of which is still used in most modern vehicles (Kahane, 2004). The modern three-point belt is a very effective safety device. Given a crash, these belts are estimated to reduce the likelihood of a fatality by approximately 40e45% and to reduce the likelihood of an injury by as much as 80%, depending on the type of crash and vehicle (Cummings, Wells, & Rivara, 2003; NHTSA, 2001). Between 1960 and 2002, safety belts were responsible for preventing 168,524 motor vehicle deaths in the United States (Kahane, 2004), and in 2008 alone, the estimated number of prevented fatalities was 13,250 (NHTSA, 2008). These numbers highlight the enormous public health impact safety belts, belt use legislation, and belt use intervention programs have had. When combined with a modern air bag, the likelihood of injury and death is reduced even more (NHTSA, 2001). Most cars today also include many other safety features as standard equipment (e.g., antilock brakes, crumple zones, and third brake lights), as well as optional safety equipment (e.g., adaptive cruise control, lane departure warnings, and rearview cameras). However, there is an important key difference between safety belts and other safety features such as air bags and third brake lights: motorists must use safety belts in order for them to be effective.
3. MEASUREMENT The gold standard for measurement of safety belt use is a methodology known as direct observation. As the name suggests, data collectors look into passing vehicles and visually observe the safety belt use of occupants in those vehicles. Direct observation studies are expensive and difficult to coordinate, but compared with other methods, they provide a less biased accounting of the true belt use rate for the area under observation. If sampled and conducted properly, these studies can provide an accurate “snapshot” of belt use that is representative of a much larger area, including a state, region, or even a country (Chaffe, Solomon, & Leaf, 2009; Eby, Vivoda, & Cavanagh, 2009; Pickrell & Ye, 2009a). Direct observation also has some important limitations, however. By definition, this method only allows for collection of data on phenomena that can be readily
PART | III
Key Problem Behaviors
observed. Generally, such studies only assess vehicle type, seating position, sex, age, race, and, of course, safety belt use. Understanding why people choose not to wear safety belts on a given trip, or identifying underlying factors that can influence belt use, cannot be assessed. In addition, because this technique is unobtrusive, data collectors must observe and make judgments about many of the vehicle occupants’ characteristics. The validity of these judgments is rarely assessed because the only way to do so would be to stop vehicles and ask occupants their age, race, etc.dan approach that would be disruptive and thus is rarely used. For a more detailed description of the benefits and problems associated with direct observation surveys, see Chapter 5. Aside from direct observation, several other methods are also used to assess safety belt use. One of the most popular is self-report. Chapter 4 provides a complete description of self-report instruments and methods; issues specifically related to safety belt use are described here. Self-reported belt use is usually obtained by using either telephone interviews or paper-and-pencil questionnaires. These types of surveys are much less expensive to administer than direct observation, but research has shown that they tend to be less accurate, with self-reported belt use nearly always higher than observed use (Hunter, Stewart, Stutts, & Rodgeman, 1993; Nelson, 1996; Streff & Wagenaar, 1989). This consistent difference is usually attributed to either social desirability bias or a mismatch between how people understand the survey questions and what the researchers actually intended. Social desirability bias refers to the tendency for participants to respond in a way that they believe will be viewed favorably by others (Paulhus, 2002). Because most people believe that wearing a safety belt is a good thing to do, to avoid being judged negatively by others, some people may respond affirmatively even when they do not wear a safety belt. The second common reason for higher self-reported belt usedthe misunderstanding of the question being askeddis best illustrated in the Motor Vehicle Occupant Safety Survey (MVOSS). In the MVOSS, participants were asked how often they wear a safety belt while driving. Immediately following this question, they were asked when was the last time that they did not wear a safety belt while driving. Of the 88% of drivers who reported wearing a safety belt “all of the time” while driving, 10% immediately reported not wearing their belt within the past month (Boyle & Lampkin, 2008). It is clear that many respondents did not interpret the meaning of “all of the time” to be literal but, rather, an approximation, thus creating measurement error within the survey results. For these reasons, direct observation methods are used when it is important to estimate the most accurate belt use rate possible, but self-report is used when a richer exploration of the factors that influence safety belt use is needed.
Chapter | 16
Factors Influencing Safety Belt Use
Self-report surveys can easily assess age and race, as well as many other factors that influence belt use (i.e., income and education), by simply including questions about them. This method can also explore why people choose to wear or not to wear safety belts in given situations. This method allows us to better understand how factors such as the expected length of a given trip or the purpose of a trip influence one’s decision to wear a safety belt. Because these different methods allow us to answer different questions about safety belt use, it is important to continue to use both to best understand all aspects of this behavior. Other data sources are also sometimes used to determine belt use rates, including police reports, crash databases, and motor vehicle fatality databases. These tend to be used less often because of the inherent biases associated with them. Safety belt citations from police records are often not aggregated across precincts, making analysis of population-wide belt use difficult. In addition, these records only contain information for motorists who actually received a citation, and these motorists were often originally stopped for some other reason, resulting in no viable comparison group of those who were in compliance with the law. Data from crash and fatality databases can also present problems largely because everyone represented in the database was involved in a crash (or fatal crash). These motorists are not representative of the larger population, thus making it difficult to generalize belt use to the motoring public. However, even given these inherent weaknesses, one valuable use for these data is to evaluate changes in belt use over time. For such an analysis, it is not necessary to generalize the results, and any biases that exist will presumably be constant at both points in time. Although direct observation methods offer the most accurate estimation of safety belt use, self-report methods, citation databases, and crash databases can all be valuable for assessing different aspects of safety belt use. Depending on the available funding and the questions that need to be answered, each of these sources has its place in a researcher’s repertoire. However, when using each of these methodologies, it is important to understand the strengths and weaknesses of the given approach.
4. INTERNATIONAL SAFETY BELT USE According to the World Health Organization (WHO, 2009), 1.27 million people died globally in traffic crashes during 2008. WHO estimates that by 2030, there will be 2.4 million yearly traffic crash-related fatalities throughout the world, and that traffic crashes will be the fourth leading cause of death. The same report also suggests that the regular use of safety belts could significantly reduce this projected fatality number.
217
Table 16.1 shows published safety belt use rates from a sample of countries. There are clear differences, with belt use varying from 11 to 96%. Note, however, that there are some challenges when comparing safety belt use rates. Table 16.1 also includes the data collection method, sample, area, and year(s) data were collected, and each of these factors could impact the reported rates. Data from studies of city or regional belt use may be quite different from countrywide data. Indeed, statewide belt use in the United States during 2009 ranged from 67.6% in Wyoming to 98.0% in Michigan (Chen & Ye, 2010). Despite these issues, safety belt use clearly differs among countries for a variety of reasons. One important reason is the presence of a mandatory safety belt law. A study of 178 countries found that only 88% had national or regional laws requiring use of a belt, and only 57% of countries required restraint use by all vehicle occupants (WHO, 2009). Furthermore, the economic condition of countries was related to the likelihood of a mandatory belt use law, with only 54% of low-income countries surveyed having a belt use law. Even with a law, belt use can be low if enforcement of the law is weak. The WHO (2009) report found that only 19% of countries reported strong enforcement of their law. In addition, use laws cannot be effective if vehicles are not equipped with safety belts. Only 29% of countries that manufacture vehicles had regulations requiring safety belts for all vehicle seating positions (WHO, 2009). China, for example, only started requiring all vehicles sold in that country to have belts for all seating positions in 2004 (Routley et al., 2008). International belt use rates also vary because of cultural differences among countries. Among university students studied in 15 different countries, attitudes toward the value of safety belts were significantly related to countrywide belt use rates (Steptoe et al., 2002). In the United Arab Emirates, cultural and attitudinal differences about safety belts may account for the low belt use within this high-income, developing country (Barss et al., 2008). Likewise, low safety belt use observed in Russia may be due in part to the lack of a culture around using belts (Akhmadeeva, Andreeva, Sussman, Khusnutdinova, & Simons-Morton, 2008). Other studies have addressed the effects of cultural beliefs in destiny (fatalism) on belt use and found that use of belts is lower among populations who hold those beliefs (Peltzer, 2003). From a global perspective, increasing the use of belts can have a profound impact on the reduction of unintentional injury. WHO (2009) recommends a 5-point strategy to increase belt use worldwide: (1) Require vehicle manufacturers to install belts in all seating positions; (2) improve laws to require safety belt use in all vehicle seating positions; (3) strengthen enforcement and ensure
218
PART | III
Key Problem Behaviors
TABLE 16.1 A Sample of Safety Belt Use Rates in Countries Throughout the World Country
Rate (%)
Method
Sample
Area
Year(s)
Reference
Argentina
86.0
Not reported
All vehicles
City
2004
Silveira (2007)
Australia
96.0
Direct observation
Not reported
Country
2009
Australian Automobile Association (2010)
Belgium
76.0
Self-report
University students
Country
2000
Steptoe et al. (2002)
Canada
92.5
Direct observation
Occupants, cars
Country
2006e2007 Transport Canada (2008)
China
55.1
Direct observation
Drivers, cars
Two cities
2005e2007 Routley et al. (2008)
Columbia
32.0
Crash data
Occupants, all vehicles
Country
2005-2006
O’Bryant (2008)
Costa Rica
82.0
Direct observation
Occupants, cars
Country
2004
FIA Foundation (2005)
England
95.0
Direct observation
Drivers, cars
Country
2009
Walter (2010)
France
91.5
Self-report
University students
Country
2000
Steptoe et al. (2002)
Germany
76.5
Self-report
University students
Country
2000
Steptoe et al. (2002)
Greece
57.5
Self-report
University students
Country
2000
Steptoe et al. (2002)
Hungary
73.0
Self-report
University students
Country
2000
Steptoe et al. (2002)
Iceland
84.0
Self-report
University students
Country
2000
Steptoe et al. (2002)
Ireland
85.5
Self-report
University students
Country
2000
Steptoe et al. (2002)
Italy
83.5
Direct observation
Drivers, cars
Region
2005
Zambon et al. (2008)
Jamaica
81.2
Direct observation
Front outboard, cars
City
2004
Crandon et al. (2006)
Native United States
55.4
Direct observation
Front outboard, cars
Tribal 2004 land, USA
Leaf and Solomon (2005)
Netherlands
86.0
Self-report
University students
Country
2000
Steptoe et al. (2002)
Poland
76.5
Self-report
University students
Country
2000
Steptoe et al. (2002)
Portugal
94.0
Self-report
University students
Country
2000
Steptoe et al. (2002)
Russia
77.9
Direct observation
Not reported
Territory
2006
Global Road Safety Partnership (2010)
Scotland
95.0
Direct observation
Drivers, cars
Country
2009
Walter (2010)
Slovenia
95.6
Self-report
Front outboard, cars
Country
2000
Bilban and Zaletel-Kragelj (2007)
South Africa
81.0
Crash data
Drivers, all vehicles
Country
2002
Olukoga and Noah (2005)
Spain
80.0
Self-report
University students
Country
2000
Steptoe et al. (2002)
Turkey
35.5
Self report
Convenience sample
City
2007
S¸ims¸eko glu and Lajunen (2009)
United Arab Emirates
11.0
Direct observation
Drivers, all vehicles
City
2003e2004 Barss, et al. (2008)
United States
83.0
Direct observation
Front outboard, cars
Country
2008
that enforcement is equal for all seating positions; (4) establish systems to collect data on rates of safety belt use; and (5) supplement all efforts with a mass media campaign that highlights the safety benefits of belt use, the increased likelihood of being cited for non-use, and the penalties for being cited.
Pickrell and Ye (2009a)
5. SAFETY BELT USE IN THE UNITED STATES In 1984, New York became the first state to enact a law requiring belts to be worn by occupants riding in the front seats of vehicles. Prior to this first law, the overall U.S. belt use rate was estimated to be between 16 and 18% (Hedlund,
Chapter | 16
Factors Influencing Safety Belt Use
219
Gilbert, Ledingham, & Preusser, 2008). Many other states quickly followed New York’s example, and belt use increased dramatically throughout the United States. By 1987, 20 states and the District of Columbia had enacted belt use legislation, and U.S. belt use had increased to more than 40%. By 1995, every state except New Hampshire had passed a law, with belt use estimated at approximately 70% (Hedlund et al., 2008). Achieving this level of belt use, and nearly universal belt use legislation, was quite an accomplishment, but closer analysis of both revealed an important limitation. When many of these belt use laws were passed, they were done so with what was termed “secondary enforcement.” Secondary enforcement means that a police officer cannot stop a motorist only because safety belt non-use is observed; the motorist must be stopped for some other infraction (e.g., speeding) and can then also be cited for failing to buckle up. Indeed, 42 of the original 50 jurisdictions enacted secondary laws (Hedlund et al., 2008). Because secondary laws are much more difficult to enforce, states with this provision generally have much lower belt use rates than primary enforcement jurisdictions. During the 1990s, several states began efforts to upgrade their enforcement provisions to primary, with California becoming the first to do so in 1992 (Hedlund et al., 2008). When a state changes its belt use law from secondary to primary enforcement, a specific pattern of use typically
emerges. Immediately following the legislative change, belt use increases dramatically, but then it generally decreases somewhat as the months continue. As the rate stabilizes, overall use remains at a significantly higher level than was observed before the change (Figure 16.2). The median observed safety belt use increase in states that have made this change is 14 percentage points (Shults, Elder, Sleet, Thompson, & Nichols, 2004), which translates to a substantial number of reduced injuries and deaths. NHTSA (2009) estimates that for each percentage point increase in U.S. belt use overall, approximately 270 lives are saved. However, as of 2010, only 30 states (and the District of Columbia) allow for primary enforcement, so there is still work to be done (Insurance Institute for Highway Safety (IIHS), 2010). Aside from legislative changes, interventions have also been very effective at increasing belt use. The most famous and most successful safety belt intervention is known as “Click It or Ticket” (CIOT; Hedlund et al., 2008). CIOT utilizes widespread media messages informing the public that police will be specifically focusing on safety belt use enforcement during a given time frame. These messages are supplemented with high-visibility police enforcement of the safety belt law. Coupling the media messages with enforcement serves to increase motorists’ perception of the likelihood of receiving a ticket for safety belt non-use. As
100.0 90.0 80.0
Use rate, percent
70.0 60.0 50.0 40.0 30.0 20.0 10.0
Safety belt law implemented
2010
2009
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
1989
1988
1987
1986
1985
1984
1983
0.0
Primary enforcement implemented Year
FIGURE 16.2 Example of safety belt use in Michigan, 1983e2009. Source: Based on data from Datta and Savolainen (2009); Eby, St. Louis, and Vivoda (2005); O’Day and Wolfe (1984); and Streff, Molnar, and Christoff (1993).
220
PART | III
a collective result of all these effortsdfrom the original safety belt laws to the primary enforcement upgrades and the many iterations of various safety belt intervention campaignsdsafety belt use in the United States is at an alltime high of 84% (Pickrell & Ye, 2009a). Although this is an impressive increase from where belt use began, there are still key groups that continue to ride unbelted. Males, young people, pickup truck occupants, motorists in rural areas, and several other groups use safety belts less often than their counterparts (Pickrell & Ye, 2009a, 2009b). The following section describes the factors that influence belt use among different groups and explains what is known about why these differences exist.
6. FACTORS THAT INFLUENCE SAFETY BELT USE What is currently known about why individuals within various “low belt use groups” wear safety belts less often than others is described in the subsections that follow. Figure 16.3 depicts these differences in belt use. This figure allows for comparisons within a given category (e.g., belt
Key Problem Behaviors
use by vehicle type), but it also gives the reader an idea of the scope of the differences in belt use across categories (e.g., pickup truck occupants compared to young people). However, because no single study collects data about all these factors, this figure was created using several different sources, so some comparisons should be made with caution. In addition, a given individual is likely to fit into several of these categories simultaneously (e.g., a young male driver in a rural area), and Figure 16.3 does not account for these potential confounders.
6.1. Part-Time and Non-Users Prevalence rates of safety belt use are often considered on an “all-or-nothing” basis. We tend to think of people in a given group as either more or less likely to wear a safety belt compared to people in other groups. Although there is value in taking this group-based approach, it is also important to consider the variation within a given group at both the situational and the individual level. Self-report surveys and qualitative research have identified the existence of three distinct types of safety belt users: (1) people
0-3
20.0
Some college college grad
>$100 K
Daytime
$15-30 K
Primary
Other White
Driver
Suburban
Urban
Female
25-69 70+
4-7 8-15
5.0
Cars Vans/SUVs
-20.0
Vehicle type
Age
Gender
Population Seating denstiy position
Race
Vehicle Law purpose type
Time of day
Income†