No. 985
The Potential of Driver Education to Re‐
duce Traffic Crashes Involving Young
Drivers
Anders Nyberg Division of Social Medicine and Public Health Science, Department of Health and Society, Linköping University, SE‐581 83 Linköping, Sweden Linköping 2007
The Potential of Driver Education to Reduce Traffic Crashes Involving Young Drivers © Anders Nyberg, 2007 Cover picture by Jonna Nyberg Published papers were reprinted with the permission of the copyright holder. Printed by LiU‐Tryck, Linköping 2007 ISBN: 978‐91‐85715‐69‐5 ISSN: 0345‐0082
CONTENTS
ABSTRACT ... 1 LIST OF PAPERS ... 3 LIST OF FIGURES AND TABLES... 4 DEFINITIONS... 5 1 BACKGROUND ... 7 1.1 TRAFFIC FATALITIES AND INJURIES—A SERIOUS PUBLIC HEALTH PROBLEM... 7 1.2 INVOLVEMENT OF YOUNG DRIVERS IN CRASHES... 8 1.2.1 Fatalities and severe injuries... 8 1.2.2 Crash types and circumstances related to injury crashes involving young novice drivers... 13 1.3 FACTORS ESSENTIAL FOR UNDERSTANDING THE PROBLEMS OF YOUNG DRIVERS... 17 1.3.1 Age and experience... 17 1.3.2 Gender differences ... 19 1.3.3 The impact of experience on development of skills ... 19 1.3.4 Individuals’ own estimations of their competence as drivers ... 22 1.3.5 Individual and social preconditions ... 23 1.4 MEASURES AVAILABLE TO REDUCE THE PROBLEMS OF YOUNG DRIVERS... 26 1.5 THE POTENTIAL OF DRIVER TRAINING TO REDUCE THE PROBLEMS OF YOUNG DRIVERS... 27 1.5.1 The three cornerstones of driver education... 27 1.5.2 Goals of driver education ... 28 1.5.3 Licensing systems... 31 1.5.4 The Swedish system for obtaining a class B driving license ... 36 2 AIMS... 39 3 MATERIALS AND METHODS ... 41 3.1 DESCRIPTION OF THE MATERIALS AND METHODS... 41 3.2 GENERAL COMMENTS ABOUT THE MATERIALS AND METHODS... 41 3.3 PRACTICING IN RELATION TO THE OUTCOME OF THE DRIVING LICENSE TEST (PAPER I) ... 44 3.4 PRACTICING FOR AND PERFORMANCE ON DRIVER’S LICENSE TESTS IN RELATION TO GENDER DIFFERENCES IN CRASH INVOLVEMENT AMONG NOVICE DRIVERS (PAPER II)... 453.5 SIXTEEN YEARS AGE LIMIT FOR LEARNER DRIVERS IN SWEDEN—AN EVALUATION OF SAFETY EFFECTS (PAPER III)... 47
3.6 ACCIDENT INVOLVEMENT AMONG LEARNER DRIVERS—AN ANALYSIS OF THE CONSEQUENCES OF SUPERVISED PRACTICE (PAPER IV)... 49
3.7 SAFETY HALLS—AN EVALUATION (PAPER V)... 51
4 RESULTS... 55
4.1 PRACTICING IN RELATION TO THE OUTCOME OF THE DRIVING LICENSE TEST (PAPER I) ... 55
4.2 PRACTICING FOR AND PERFORMANCE ON DRIVER’S LICENSE TESTS IN RELATION TO GENDER DIFFERENCES IN CRASH INVOLVEMENT AMONG NOVICE DRIVERS (PAPER II)... 58
4.3 SIXTEEN YEARS AGE LIMIT FOR LEARNER DRIVERS IN SWEDEN—AN EVALUATION OF SAFETY EFFECTS (PAPER III)... 61
4.4 ACCIDENT INVOLVEMENT AMONG LEARNER DRIVERS—AN ANALYSIS OF THE CONSEQUENCES OF SUPERVISED PRACTICE (PAPER IV)... 63
5 GENERAL DISCUSSION... 67
5.1 VISION ZERO AND THE SWEDISH DRIVER LICENSING SYSTEM... 67
5.2 UTILIZATION OF THE REFORM ALLOWING PRACTICING AT THE AGE OF 16 ... 69 5.3 AMOUNT OF DRIVING PRACTICE... 70 5.4 STRUCTURE OF THE LEARNING PERIOD... 72 5.5 THE DRIVING LICENSE TESTS... 76 5.6 MANDATORY DRIVER EDUCATION... 78 5.7 RESTRICTIONS ON NOVICE DRIVERS... 82 5.8 APPLICATION OF THE PRESENT RESULTS TO LICENSING SYSTEMS IN COUNTRIES OTHER THAN SWEDEN ... 84 5.9 FUTURE RESEARCH... 85 6 CONCLUSIONS AND SUGGESTIONS ... 89 ACKNOWLEDGEMENTS ... 91 REFERENCES ... 93
ABSTRACT
Traffic fatalities and injuries among young drivers as a result of road crashes constitute a serious public health problem. The ultimate goal of traffic safety work in Sweden has been formulated in Vision Zero, which includes an image of a future in which no one will be killed or seriously injured in traffic crashes. Therefore, it is unacceptable that young learner and novice drivers are in‐ volved in road crashes that result in fatalities or severe injuries. Driver educa‐ tion is an important tool to increase the probability that young drivers actually take their share of the responsibility for Vision Zero by obeying traffic rules and driving as safely as possible.
The general aim of the work underlying this thesis was to determine the po‐ tential of driver education to reduce road traffic crashes involving young driv‐ ers, particularly in Sweden. Paper I examined the relationship between the way in which the education is carried out and the outcome of the driving test. Paper II explored whether there are any gender‐related differences regarding driving practicing, the outcome of the license tests, and involvement in crashes during the first year of licensure. Paper III evaluated the reform that made it possible for learner drivers to start practicing from 16 years of age in terms of its effects on crashes involving young novice drivers. In paper IV, the focus was on investigating crashes during practice and comparing the results with the corresponding situation for novice drivers during their first two years of licensure. Paper V assessed an insight‐based educational approach aimed at inducing young drivers to make better use of vehicle‐related safety equip‐ ment.
The findings of two of the studies (papers III and IV) showed that, in Sweden, taking advantage of the possibility to start practicing behind the wheel from the age of 16 years had a beneficial effect seen as reduced crash involvement among those young drivers. In paper V, it was revealed that using an insight‐ based educational approach can have a positive influence on learner drivers’ knowledge of and attitudes towards the use of car safety equipment (e.g., safety belts). In paper I, it was found that it is difficult to explain why 18–24‐ year‐olds pass or fail the driving test on the basis of background variables and information concerning how they had practiced driving. Paper II showed that, for females, training pursued in a more structured manner did not seem to be beneficial for the outcome of the license tests, and that males aged 18–24 were involved in 1.9 more injury crashes per 1,000 drivers than females during their
first year of licensed driving. Suggestions are given that can be used to de‐ velop the Swedish licensing system in a way that will increase the potential of driver education to reduce traffic crashes among young drivers. These ideas comprise aspects such as the following: persuading the youngest learner driver population to start practicing as early and as much as possible; the learning period should be better organized, which includes improved agree‐ ment between the goals of the national curriculum, the content/process of driver education, and the design of the license tests; professional instruction of learners in both the theory and the practice of driving should be a more prominent component of driver education; and parts of the Swedish licensing system should be made mandatory to help solve the problems of young driv‐ ers and to fulfill the goals of the national curriculum.
Keywords: driver education; driver training; learner drivers; novice drivers; young
drivers; evaluation; public health; traffic safety; supervised practice; 16 year age limit; safety halls; Sweden; questionnaires; attitudes; knowledge; experience; gender; driv‐ ing tests; crashes; crash involvement; crash types; road traffic.
LIST OF PAPERS
The thesis is based on the following papers, which will be referred to in the text by their Roman numerals: I Nyberg, A., Gregersen, N.P., Wiklund, M., 2007. Practicing in rela‐ tion to the outcome of the driving test. Accident Analysis and Preven‐ tion, 39(1), 159‐168. II Nyberg, A., Gregersen, N.P., 2007. Practicing for and performance on driver’s license tests in relation to gender differences in crash involvement among novice drivers. Journal of Safety Research (in press), doi:10.1016/j.jsr.2007.01.001
III Gregersen, N.P., Berg, H.Y., Engström, I., Nolén, S., Nyberg, A., Rimmö, P.A., 2000. Sixteen years age limit for learner drivers in Sweden – an evaluation of safety effects. Accident Analysis and Pre‐ vention, 32(1), 25‐35.
IV Gregersen, N.P., Nyberg, A., Berg, H.Y., 2003. Accident involve‐ ment among learner drivers – an analysis of the consequences of supervised practice. Accident Analysis and Prevention, 35(6), 725‐730.
V Nyberg, A., Gregersen, N.P., Nolén, S., Engström, I., 2005. Safety halls – an evaluation. Journal of Safety Research, 36(4), 429‐439.
LIST OF FIGURES AND TABLES
Figures
Figure 1. Driver fatalities per 100,000 license holders (private vehicles) in different age groups in Sweden 1994–
2005... 9
Figure 2. Severely injured drivers per 100,000 license holders (private vehicles) in different age groups in
Sweden 1994–2005... 9
Figure 3. Driver fatalities per 100,000 license holders (private vehicles) in different age groups and according to
gender in Sweden 1994–2005... 11
Figure 4. Severely injured drivers per 100,000 license holders (private vehicles) in different age groups and
according to gender in Sweden 1994–2005. ... 11 Tables
Table 1. Goals for Driver Education (GDE) matrix. ... 29 Table 2. Summary of materials and methods used in the five studies. ... 41 Table 3. Full model explaning the outcome of the driving test. ... 57 Table 4. Differences in crashes per 1,000 drivers and crashes per 10 million km driven between the 16-year-old
group and the two controlgroups with and without controlling for confounding factors... 62
DEFINITIONS
Accident/Crash: In this thesis, the terms accident and crash mean the same thing,
namely, a road traffic crash leading to fatalities or personal injury. Evans (1991) recommends the word crash, because in his opinion the term accident may cause people to think that such events occur due to fate and therefore are unpredictable and unpreventable. However, both terms are still used in literature, as well as in everyday language. Accident/Crash risk: Number of accidents/crashes per kilometre driven (unless oth‐ erwise defined). Driver education/Driver training: In this thesis these terms are used synonymously
and involves formal and informal education/training as well as both theoretical studies and behind‐the‐wheel practice.
Health risk: Number of accidents/crashes per license holder (unless otherwise de‐
fined).
Lay instructor: A non‐professional person with permission to supervise a learner
driver during practice, usually the father and/or mother of the learner.
Learner driver: A driver without a full license but with permission to practice under
supervision.
Novice driver: A person that recently qualified for a driving license, regardless of
biological age.
Professional instructor: A person specially educated to teach driving and usually
employed at a driving school or at a skid track. Safety hall: A specially designed facility at some skid tracks designed to give learn‐ ers knowledge about the various types of safety equipment in motor vehicles, why they should be used, and how to use them in a correct manner. Skid track: An enclosed, specially designed area where learners perform the part of the mandatory risk awareness education that involves low‐friction driving. Young driver: A driver 18–24 years old (unless otherwise defined). OLY/VITS: The Swedish National Road Administration Information System for Traf‐ fic Safety, which is a national crash database. NTF: The Swedish National Society for Road Safety. STAPRO: Database comprising, among other things, results of the license tests. STRADA: Swedish Traffic Accident Data Acquisition, which is also a national crash database. S(N)RA: Swedish (National) Road Administration VINNOVA: The Swedish Agency for Innovation Systems VTI: Swedish National Road and Transport Research Institute
1
BACKGROUND
This thesis concerns the potential of driver training to reduce road traffic crashes and severe injuries among young drivers, with special focus on Swe‐ den. The present research was conducted in order to draw attention to the fact that traffic fatalities and injuries among young drivers constitute a serious public health problem (OECD, 2006). It also examins the principle manifested in the Swedish Vision Zero, which anticipates a future in which no one will be killed or seriously injured in road traffic crashes. Vision Zero further states that the responsibility for traffic safety is shared by those who design and those who use the road transport system. Nonetheless, the burden is primarily on the former group in this context, because if the road users do not abide by the laws and regulations, it is the task of the system designers to solve that problem (SRA, 2006; Tingvall, 2005). It can be allotted to the system designers to draw up plans for the road environment and the laws and rules to be ap‐ plied both in the traffic system and by those who provide driver education. Therefore, driver training is an important tool to increase the probability that young drivers actually take their share of the responsibility for Vision Zero by obeying traffic rules and driving as safely as possible.
The background of the problems connected with young drivers is introduced in this chapter, along with ideas about how these issues might be solved through driver training. Also, the Swedish system for obtaining a class B driv‐ ing license is described, because, in one way or another, all the papers in‐ cluded in this thesis give information about how that system works or how it might be improved.
1.1
Traffic fatalities and injuries—a serious public
health problem
About 127,000 people are killed and more than two million are injured in road traffic crashes each year in the European Region of the World Health Organi‐ zation, WHO. This problem is especially pronounced in the age group 15–29 years, in which traffic crashes are the leading cause of death. Traffic crashes and related fatalities and injuries represent a major public health problem and lead to substantial costs to society, not only in Europe, but in all parts of the world (Racioppi et al., 2004). Notably, in the OECD countries, such crashes arethe primary cause of death among 18–24‐year‐olds, and they account for 35% of all deaths among 15–24‐year‐olds (OECD, 2006). Expressing the public health problem as disability‐adjusted life years (DALYs) has shown that road traffic injuries ranked ninth among all public health prob‐ lems in the world in 1990, and it has been predicted that they will be in third place by 2020 (Murray & Lopez, 1996, Krug, 1999; WHO, 2004).
1.2
Involvement of young drivers in crashes
1.2.1
Fatalities and severe injuries
In Sweden, during the period 1994–2005 an average of 518 people were killed annually in road traffic crashes (crash deaths due to disease excluded). Among people aged 18–24, the corresponding average number of deaths was 82, in‐ cluding 42 killed while driving a car and 20 as passengers. Also during the same time period, an average of 4,102 people were severely injured in road traffic crashes; 756 of those individuals were 18–24 year olds, and of them 336 had been driving the vehicle during the crash and 191 had been passengers. This means that, on average, a little more than 75% of all deaths and almost 70% of all severe injuries due to road traffic crashes among 18–24‐year‐olds occurred while the individuals were traveling in a private motor vehicle (Brüde, 2005; Forslund, 2006).
Figure 1 shows that traffic fatalities are a major problem among 18–24‐year‐old drivers. Comparing that group (the youngest) with drivers aged 25–64 indi‐ cated that the probability of being killed as a driver was on average 2.95 times higher (min. 1.98, max. 3.68) in the former group. In addition, compared to the oldest age group, individuals in the youngest group were 1.96 times more likely to be killed as a driver (min. 1.01, max. 2.97).
As can be seen in Figure 2, severe injuries due to traffic crashes are also a ma‐ jor problem among 18–24‐year‐old drivers. Comparing that age group with those aged 25–64 shows that the probability of being severely injured as a driver was on average 2.92 times higher (min 2.25, max 3.50) in the former group. Also, the probability was 4.11 times higher (min. 2.53, max. 5.98) for the youngest drivers when comparing them with the oldest age group.
0 2 4 6 8 10 12 14 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 D ri v er f a ta li ti es p e r 10 0, 0 00 li c e n se h o ld e rs
Total 18–24 Total 25–64 Total 65+
Figure 1. Driver fatalities per 100,000 license holders (private vehicles) in different age
groups in Sweden 1994–2005. No data regarding license holders were available for the year 1999. All crash fatalities due to disease were excluded. Sources: Forslund (2006) and Lekander (2006). 0 20 40 60 80 100 120 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 S everel y i n ju red p e r 100, 000 l icen se h o ld e rs
Total 18–24 Total 25–64 Total 65+
Figure 2. Severely injured drivers per 100,000 license holders (private vehicles) in different
age groups in Sweden 1994–2005. No data regarding license holders were available for the year 1999. Sources: Forslund (2006) and Lekander (2006).
Figure 3 shows that the traffic fatality problem is especially pronounced for the youngest male drivers, who were on average at 2.98 times higher risk (min. 1.84, max. 3.70) of being killed as a driver compared to male drivers aged 25–64. Compared to the oldest males, the risk was 2.01 times higher (min 1.06, max 3.15) in the younger group. Also, the probability of being killed as a driver was on average 4.09 times higher (min 2.65, max 6.21) for young male drivers than for their female counterparts. However, even though young fe‐ male drivers were much less crash prone compared to males, the probability of being killed as a driver was still on average 2.59 and 2.14 times higher for females aged 18–24 than for females in the age groups 25–64 and 65+ years.
Data from a number of different OECD countries show that male drivers aged 18–24 are involved in fatal road crashes three times more often (per million population) than female drivers in the same age group (OECD, 2006). Even when exposure is accounted for, the gender differences related to young nov‐ ice drivers persist, as has been shown by, among others, Kweon and Kockel‐ man (2003) in the United States and other investigators analyzing data col‐ lected in the Netherlands, Sweden, and the United Kingdom (Lynam et al., 2006). Kweon and Kockelman (2003) found that the crash risk per million miles was 1.2 times higher for male than for female drivers less than 20 years of age. Lynam et al. (2006) found that 18–24‐year‐old male drivers had over three times more fatal crashes per million kilometers driven.
With regard to severely injured drivers, the problem is again especially pro‐ nounced for young males. Considering the data in Figure 4, the probability of drivers being severely injured was on average 3.27 times higher (min. 2.52, max. 3.88) for the youngest males than for males aged 25–64. In addition, compared to the oldest males (65+ years), the likelihood of severe injury was 4.10 times higher (min. 2.51, max. 6.17) for the youngest male drivers. Also, the probability of being severely injured as a driver was on average 2.11 times higher (min 1.70, max 2.71) among young male drivers than among their fe‐ male counterparts. Notwithstanding, the probability of being severely injured was on average 2.26 and 4.51 times higher for the youngest female drivers compared to female drivers aged 25–64 and 65+ years, respectively.
0 2 4 6 8 10 12 14 16 18 20 22 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Dr iver f a ta li ti es p e r 100, 000 l icen se h o ld e rs
Males 18–24 Females 18–24 Males 25–64
Females 25–64 Males 65+ Females 65+
Figure 3. Driver fatalities per 100,000 license holders (private vehicles) in different age groups and according to gender in Sweden 1994–2005. No data regarding license holders were available for the year 1999. All crash fatalities due to disease were excluded. Sources: Forslund (2006) and Lekander (2006). 0 25 50 75 100 125 150 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 S e verel y i n ju red p e r 100, 000 l icen se h o ld e rs
Males 18–24 Females 18–24 Males 25–64
Females 25–64 Males 65+ Females 65+
Figure 4. Severely injured drivers per 100,000 license holders (private vehicles) in different
age groups and according to gender in Sweden 1994–2005. No data regarding license holders were available for the year 1999. Sources: Forslund (2006) and Lekander (2006).
To further illustrate the extent of the problem, it can be mentioned that during the years 1994–2005, 19% of all those who were killed and 21% of those who were severely injured in road traffic crashes involving private vehicles were young novice drivers. These rates are very high, since on average only 8% of all licensed drivers during the mentioned period were young novice drivers (Forslund, 2006; Lekander, 2006).
The description so far has focused on the outcome for the young drivers them‐ selves. However, the problem associated with these vehicle operators is even more preponderant when considering that they are involved as drivers in many more fatal crashes in which they are not the ones who are killed. On av‐ erage, 94 young drivers aged 18–24 were involved in 85 fatal crashes in Swe‐ den over the years 1996–2005. In those crashes, an average of 100 people were killed, and 42 of them were the young drivers themselves (Björketun, 2006; crash fatalities due to disease excluded). This means that for every young per‐ son who was killed in a crash as a driver, on average 1.38 other road users died. During the same time period, on average 618 drivers aged 18–24 were involved in 544 crashes that led to severe injuries in 950 people, and 345 of those individuals were the young drivers themselves (Björketun, 2006). This indicates that for every young person who was severely injured when in‐ volved in a road crash as a driver, on average 1.75 other road users were se‐ verely injured. These figures are similar to data collected in the Netherlands showing that for every young driver (18–24 years of age) that was killed in a crash, about 1.33 other road users died; and for every young driver who was severely injured, about 1.8 other road users sustained severe injuries (OECD, 2006). Crashes that occur during driving practice have not been investigated to any significant extent, and the study reported in paper IV was one of the first at‐ tempts to distinguish between crashes involving learners and newly licensed drivers. However, in a Swedish study by Berg et al. (2004) which investigated the circumstances of occurrence of crashes during driving practice, it was found that efforts should be taken to reduce the youngest learners’ crashes in rural areas on high speed roads. Also, the EU project designated BASIC (Hatakka et al., 2003) has reported that surveys covering all crashes during practicing (even those involving minor vehicle damage) that have been con‐ ducted in Finland, Norway, Sweden, and the United Kingdom have shown that learner drivers are involved in very few crashes. This agrees with esti‐ mates published by Forsyth et al. (1995) indicating that the crash rate is at least
20 times higher during the first year of licensed driving than during the learn‐ ing period.
To summarize, the over‐involvement of young drivers (especially young males) in traffic crashes is a considerable problem, both from the perspective of public health and in terms of road safety (OECD, 2006).
1.2.2
Crash types and circumstances related to injury
crashes involving young novice drivers
Young novice drivers are overrepresented in most types of road crashes. However, involvement of this group is more extensive in certain kinds of crashes, and, in many countries, the highest overrepresentation is found in single‐vehicle crashes and loss‐of‐control crashes. For example, in the US state of Maryland, single‐vehicle crashes have been found to constitute 25.6% of all vehicle crashes involving 16‐year‐old drivers (Ballesteros & Dischinger, 2000). In Sweden, Gregersen and Nyberg (2002) found that during the years 1994– 2000, 27% of the traffic crashes involving 18–19‐year‐old drivers were single‐ vehicle crashes, whereas the corresponding rate for all other ages groups was 14%. Considering single‐vehicle crashes involving fatalities, the rate was 32% in the youngest age group compared to 24% among drivers of all other ages. Also, Clarke et al. (2001) have observed that in the United Kingdom 22% of road crashes involving 17–19‐year‐old drivers have been single‐vehicle crashes. The overrepresentation of young novice drivers is also evident when consider‐ ing loss‐of‐control crashes. For example, Harrison et al. (1999) found that 18% of causal injury crashes involving young drivers entailed loss of control of the vehicle. Similar findings have been reported by Laapotti and Keskinen (1998) and Clarke et al. (2002). It is also well known that young novice drivers are overrepresented in crashes that occur at night on weekends, as observed by, for instance, Gregersen and Nyberg (2002) in Sweden, Laapotti and Keskinen (1998) in Finland, Williams (1985) in the United States, and the OECD (2006) in Australia. As an example, 32% of crashes involving 18–19‐year‐old drivers in Sweden in 1994–2000 hap‐ pened in darkness (Gregersen & Nyberg, 2002), whereas the corresponding rate for drivers in all other age groups was 22%. Clarke et al. (2006) found that young drivers 17–25 years old were particularly prone to crashes caused by
loss of control of the vehicle on curves and in darkness. Those investigators also noted that young male drivers were more often responsible for rear‐end crashes, whereas young females were more frequently the drivers of the cars that were hit. Clark and colleagues also noted that the identified problems were greatest for 17–19‐year‐old males. Furthermore, they observed that, with regard to the likelihood of crashes, there was generally an improvement over time, that is, the older the driver the less the chance of being involved in a road crash. It is well established that speeding is a major factor related to car crashes in‐ volving young novice drivers—the higher the speed the greater the risk of los‐ ing control of the vehicle. Moreover, since younger drivers also tend to have more passengers in the car, if an crash does occur, more people will be injured and the outcome of the crash will be more severe (e.g., Evans, 1991; Twisk, 1994). In the United Kingdom, speeding has been found to be by far the most common offense among young drivers, as reported by Forsythe et al. (1995). Also, observational studies performed by Waylen and McKenna (2002), showed that younger license holders drove significantly faster than those who were older. McKnight and McKnight (2000) have estimated that speeding ac‐ counts for 20% of all crashes in the US states of California and Maryland. In Australia, Harrison et al. (1999) observed that speeding was more frequently done by young drivers (especially males) than by those who were older. Simi‐ larly, in a large‐scale study conducted in Europe by Goldenbeld (1999), young male drivers in particular reported that, compared to older drivers, they drove faster and also more often over the legal speed limit. In an investigation by Neyens and Boyle (2007), it was found that young drivers who were speeding or had been drinking alcohol were more likely to be involved in crashes with fixed objects than with other vehicles. Data covering 1983–2002 obtained in the United States have revealed that individuals younger than 25 represent the largest proportion of drivers known to be speeding when fatal crashes oc‐ curred (OECD, 2006). In that study, it also became apparent that the problem of speeding is clearly most pronounced among young male drivers, although speeding does seem to be a male problem at any age.
Alcohol has also been found to play an important role in crashes involving young drivers. This is very interesting when considering that it has not been established that young drivers are more inclined than older drivers to operate a vehicle while under the influence of alcohol (Forsman & Gustafsson, 2004). Nonetheless, young drivers (typically young men) are overrepresented in al‐
cohol‐related crashes, usually in combination with high speed, night‐time driving, and the presence of passengers (e.g., Brorsson et al., 1993; Schulze, 1996; OECD, 2006). Data compiled in Sweden in 2002 have shown that the drivers in 63 car crashes that led to fatalities had been drinking alcohol, and 35% of those drivers were 18–24 years old, even though this age group ac‐ counted for only 7% of the licensed drivers in the country at that time (SRA, 2004a). A plausible explanation for that observation is that, with each drink consumed, the risk of fatal injuries increases more rapidly in young people than in those who are older (Glad, 1985; Keall et al., 2004). Beirness and Simp‐ son (1991) and Shope (1997) have reported that, compared to drivers who have not had any crashes, those who have been involved in crashes are overrepre‐ sented with regard to factors such as the use of alcohol and other intoxicating substances. Corfitsen (1994), Pack et al. (1995), and Horne et al. (2002) have all concluded that tiredness is an important cause of many crashes involving young male drivers. Furthermore, findings reported by Åkerstedt and Kecklund (2001) in‐ dicate that the crash risk for drivers aged 18–24 is 5–10 times higher at night compared to before noon. A logical explanation for these observations is that young drivers have little experience and knowledge of how to cope with fa‐ tigue compared to older, more experienced drivers (Summala & Mikkola, 1994). Based on reviews of research on factors such as sleep loss and sleeping habits, Groeger (2006) hypothesized that drowsiness may have a major impact both on crash involvement among young drivers during darkness and the early hours of the morning and on problems associated with learning to drive. This hypothese was based on the findings that the youngest age group exhib‐ its the largest difference between the actual amount of sleep they get and the amount that they really require and/or desire. It is often emphasized that the presence of passengers, particularly passengers of a similar age, increases the risk of crash involvement in young novice driv‐ ers. It has been suggested that this problem is sometimes due to the driver be‐ ing affected by group pressure from the passengers and that at other times it is the result of the driver being distracted by the passengers. In an instrumented vehicle study conducted by Engström (2003), it was found that young male passengers tried to influence a young male driver (by egging on or mocking) to act dangerously in different ways (e.g., to drive faster, to overtake another vehicle, or to test the performance of the car). The results showed that in most cases the driver resisted the pressure from the passengers and that the mean
speed was higher when the driver was alone than when there were passengers in the vehicle. In another study, Engström et al. (2006) observed that passen‐ gers had a protective effect, regardless of the age or the gender of the driver, and that influence was lowest on drivers (particularly males) who were 18–24 years of age, although even in this age group the crash risk was higher for those driving without passengers. These findings are contradicted by other research results, especially analyses of data collected in the United States showing that the crash risk is higher in the presence of same‐age passengers or increasing numbers of passengers (e.g., Preusser et al., 1998; Williams, 2000; Chen et al., 2000). This discrepancy might be explained by the cultural differ‐ ences between the United States and Sweden, or that the drivers in the young‐ est age group in Sweden were a little older than those included in the US stud‐ ies (who were 16–19 years old). Obviously, there are also many other factors that might divert the attention of young drivers. For instance, Gregersen and Falkmer (2003) have pointed out that in‐vehicle support systems may have a distracting and negative impact on traffic safety, particularly for young novice drivers, since it is reasonable to expect that that group will be affected by problems such as a higher mental workload, a tendency to overestimate the in‐vehicle support systems, or other motives associated with their way of operating a motor vehicle. These authors concluded that the support systems may very well be beneficial for the safety of young licensees, although little research has been done to address that is‐ sue. Therefore, to be able to exploit in‐vehicle aids in the best way possible as a means of enhancing traffic safety for young drivers, further studies will have to be carried out to determine how drivers in this age group make use of such systems (e.g., regarding level of acceptance, utilization, and overconfidence). This reasoning is in line with explanations put forward by Sarkar and Andreas (2004) stating that young drivers tend to overestimate their ability to handle new technological devices while driving. Also, Neyens and Boyle (2007) found evidence that the rate of crashes with fixed objects will increase among young drivers confronted with in‐vehicle distractions (e.g., adjusting controls and eating). Furthermore, those researchers noted that rear‐end collisions are more likely to happen if the driver is cognitively distracted or if there are distur‐ bances related to passengers or a cell phone.
The above‐mentioned overrepresentation of young novice drivers in crash sta‐ tistics is magnified by the fact that a disproportionately large number of those individuals are included in the group of drivers who do not use a seat belt.
Again, this is seen primarily in young male drivers, and in Sweden the lower rate of seat belt wearing has been found both in observational studies (Ceder‐ sund, 2006) and in two in‐depth investigations of fatality crashes (SNRA, 2003; SRA, 2004b). In the SRA study from 2004, it was found that 40% of 18– 24‐year‐olds killed in car crashes had not worn a seat belt. Among 18–25 year olds in Sweden the average general rate of usage between the years 2000–2005 was 81% for males and 93% for females (Cedersund, 2006). Similar findings have also been made in other countries (e.g., Goldenbeld, 1999; NHTSA, 2001; Cvijanovich et al., 2001; Williams & Shabanova, 2002; Williams et al., 2003; McCartt & Shabanova Northrup, 2004). The seat belt is the piece of safety equipment known to most effectively minimize the severity of injuries caused by a crash. Hence the lower rate of seat belt wearing among young drivers is no doubt one reason why these individuals are overrepresented in crash‐ related injury statistics.
1.3
Factors essential for understanding the prob‐
lems of young drivers
This section gives a brief overview of aspects connected with decision making, driving behavior, and the involvement of young novice drivers in road crashes, especially with respect to factors behind many of the problems ex‐ perienced by this group of licensees. It is also important to point out that the problems affect all novice drivers regardless of gender, although they have a greater impact on males. It is generally assumed that the three main factors in this context are age, experience, and gender. However, it is also worth men‐ tioning that there is interaction and overlap between these three factors and between the underlying causes and processes (e.g., individual and social cir‐ cumstances) (OECD, 2006).
1.3.1
Age and experience
The effects of age and experience are very important for the outcome of both the learning period and the time as a novice driver. Studies have shown that the older driver is when she/he obtains a license, the lower the risk of initial crash involvement as a novice driver (e.g., Cooper et al., 1995; OECD, 2006). It has also been established that the process of acquiring experience has a sub‐ stantial impact on the outcome during the first period as a licensed driver. For example, Sagberg (2000) and Mayhew et al. (2002) have shown that there is a sharp reduction of crashes (by 40–50%) among novice drivers during their first
seven to eight months of licensed driving. In addition, Maycock, et al. (1991) found that both age and experience have an effect; more precisely, these re‐ searchers studied recent licensees and noted that the rate of crash involvement decreased with age and with increasing experience of the driver. Nevertheless, results vary in this context. It seems that factors related to age account for 30– 50% of the crash reduction during the first period of independent driving, and the corresponding rates for the factors associated with experience are in the range 50–70% (Gregersen, 2003). Notably, it is also clear that age‐related fac‐ tors have a more pronounced effect the lower the age limit is for obtaining a driver’s license (e.g., a person can obtain a license at the age of 16 years in mo‐ st US states). Even so, many reviews have revealed that the lack of experience is one of the primary factors contributing to crash involvement among novice drivers (e.g., Lynam & Twisk, 1995; Mayhew & Simpson, 1996; Engström et al., 2003; OECD, 2006). In the report published by the OECD (2006), it was con‐ cluded that the findings of many studies indicate that experience is more im‐ portant than age when trying to explain why young drivers have high crash rates during early licensure.
Therefore, it is important that aspiring drivers are given the opportunity to gain as much experience as possible during the learning period, before they get a license. This conclusion is supported by two studies in which it was found that driving skills improved as a power function of the amount of be‐ hind‐the‐wheel practicing done by a learner (Groeger & Clegg, 1994; Groeger & Brady, 2004), and also by an investigation showing that simply gaining driv‐ ing experience is an important part of learning to operate a motor vehicle (Maycock & Forsyth, 1997). The latter study showed that males that had actu‐ ally driven more during the learning period had lower rates of crash involve‐ ment compared to males that had been awarded a license within a relatively short period of time without much behind‐the‐wheel training. The importance and potential safety benefits of gaining experience constituted one of the main arguments behind the decision to lower the age limit for practicing in Sweden from 17½ to 16 years. Inasmuch as the age for licensing is still 18, this legisla‐ tive change resulted in the possibility of acquiring supervised experience be‐ hind the wheel for a period of two years before driving alone as a license holder. This reform was evaluated in the study reported in paper III.
1.3.2
Gender differences
Clearly, even though the crash risk is generally higher for young drivers than for those who are older and more experienced, the problem is most alarming for young males. It has been suggested that the gender difference in crash in‐ volvement among young novice drivers is associated with several factors, such as the following tendencies exhibited by many young males: being more prone to sensation seeking and therefore more reckless when driving (Jonah, 1996; Clarke et al., 2005; Groeger & Brown, 1989); overestimating their driving ability (Gregersen, 2003); having a higher degree of risk acceptance while driv‐ ing (Deery, 1999); driving in “difficult” situations, such as bad weather and late at night (Forsyth et al., 1995); driving faster (Waylen & McKenna, 2002); driving while intoxicated (Begg et al., 1999); having more passengers in the car, which increases exposure to peer pressure (Rolls & Ingham, 1992; Lapotti et al., 1998); being less safety‐oriented (Meadows & Stradling, 1999; Laapotti, 2003); being more prone in violating formal and informal traffic rules (Rimmö, 1999); having a lifestyle associated with risky driving behavior (Schulze 1990; Gregersen & Berg, 1994); having different goals for their driving and generally driving more frequently and consequently being affected by more situations (Gregersen, 1998; Laapotti, 2003).
1.3.3
The impact of experience on development of
skills
Research has shown that to be able to drive safely, it is necessary to become experienced in the tasks involved in operating a motor vehicle. A driver’s mental workload, visual search skills, and capacity for hazard perception are to a large extent connected with driving experience. Therefore, these factors have been shown to be problematic for learners and novice drivers (e.g., Gre‐ gersen, 2003), and it is also believed that they can partly explain why it is es‐ sential to have driving experience.
As a driver gains experience, more tasks—both motorized and visual—will become wholly or partly automatized, and his/her ability to perform an in‐ creasing number of exercises simultaneously will gradually improve. These processes can be described on the basis of a theory formulated by Rasmussen (1986), which considers human behavior in terms of the way that individuals control their own actions on three different levels in relation to the develop‐ ment of skills. The level that a person makes use of is decided by the experi‐
ence that she/he has of the task to be solved. In the context of driving a motor vehicle, an inexperienced driver or one who encounters unfamiliar situations acts on a knowledge‐based level, implying that actions are preceded by conscious decisions that resemble problem solving. On this level, the driver has no ready solutions at hand, and the decisions emanate from, for instance, the person’s own know‐how and previous experiences of other situations (e.g., riding a bi‐ cycle).
The more experienced a driver becomes, the greater the number of driving tasks and traffic situations that will be familiar and recognized. In other words, the driver gradually accumulates a repertoire of behavioral patterns that can be used to solve and handle traffic tasks and situations that resemble each other. That is, a certain driving situation will be associated with a particu‐ lar pattern of behavior. Decisions on this rule‐based level will be more or less unconscious and automatized.
After extensive driving experience, a driver can process a large amount of in‐ formation on an unconscious level. This means that many tasks are done automatically without the driver being aware of what information he/she has used to make decisions or take actions. This skill‐based level requires very little conscious attention on the part of the driver. It should be emphasized that the majority of drivers switch between all three levels while operating a motor ve‐ hicle, and it is the experience of the driver and the traffic system as a whole that determine the level on which most of the driving is performed.
Another important issue that is closely related to the discussion above is the extent to which tasks involved in driving make demands on the driver’s men‐ tal capacity. It has been suggested that one explanation why experienced driv‐ ers are better than inexperienced drivers at managing traffic situations is a de‐ creased mental workload (Gregersen, 2003), which can be achieved by more extensive practice behind the wheel. If decisions and actions are largely auto‐ matized because a driver is very experienced, then a limited amount of mental resources is needed to operate the vehicle. However, it is not only these me‐ chanical aspects that are of importance. The more experienced a driver is the easier it is for him/her to recognize and interpret different traffic situations and on the basis of that develop “mental models” that can facilitate processing of the information and use of the available mental resources. Thus, the size of the mental workload is influenced by the degree to which driving has become automatized and the particular mental models that have become established.
For instance, mental resources are required to solve each subtask correctly, and the driver’s capacity in that respect determines how many subtasks can be fulfilled simultaneously. Therefore, to drive safely, it is necessary to be able to make use of the controls in the vehicle (e.g., the steering wheel, clutch, gear lever, and brake and accelerator pedals) and at the same time interact with other motor vehicles and road users, adapt to changes in the road environ‐ ment, and take into account both formal and informal traffic rules. This means that a driver must be able to very rapidly process and interpret all pertinent information outside the vehicle and react in accordance with the specific situa‐ tion at hand (e.g., by applying the brakes or changing direction) to continue to be in control of what is happening. An inexperienced driver needs extensive mental resources to solve fairly “simple” tasks (e.g., shifting gears), which means that insufficient resources are left to solve all other tasks satisfactorily. This complexity of information processing in humans has been described by, for example, Wickens (1992).
Various approaches have been used to investigate the ways that drivers deal with the mental workload as they become more experienced. Patten et al. (2006) applied a secondary task method in a field study, and they were able to verify the above‐mentioned theory proposed by Rasmussen (1986) indicating that driving experience reduces the mental workload on the driver. Patten et al. (2006) suggests that this finding can be particularly important for learners, since the abilities needed to be able to drive safely must be acquired cogni‐ tively (e.g., in different traffic situations) and as motor skills (e.g., being able to handle the vehicle). Also, in a study performed in real traffic situations to evaluate the Swedish reform lowering the age limit for driving practice to 16 years (Gregersen et al., 2000; Nyberg & Nolén, 2001), it was observed that young novice licensees who had gained more driving experience as learners reported a lower mental workload compared to those who were less experi‐ enced. Similarly, Shinar et al. (1998) found that inexperienced drivers did not detect road signs as well as experienced drivers did when driving a car with a manual transmission compared to an automatic transmission. This finding was interpreted as showing that gear shifting is a task that becomes automa‐ tized over time, or, in other words, that a novice driver will have to use most of his/her mental resources to shift gears so that there will not be enough left to detect road signs. Falkmer and Gregersen (2005) have found that, compared to experienced driv‐ ers (i.e., those with more than 100,000 km of licensed driving), inexperienced (learner) drivers more often fixated on objects inside the vehicle, on things of
relevance in the traffic, and on entities constituting potential hazards in the traffic environment, and they spread their fixations less on the horizontal me‐ ridian. Furthermore, although some results are contradictory, research has shown that there are differences between novice and experienced drivers, at least with regard to visual search skills (e.g., Mourant & Rockwell, 1972; Miltenburg & Kuiken, 1990; Falkmer & Gregersen, 2001; Åberg, 1981; Chap‐ man & Underwood, 1998; Williams, 1985), and this is generally believed to be related to the mental workload.
Hazard perception entails the ability to discover, recognize, and react to po‐ tentially dangerous situations, and investigations on that subject have also in‐ dicated that mental workload is one of the most important aspects to consider when trying to understand the difficulties that novice drivers encounter in traffic. For instance, several studies have provided evidence supporting the notion that automation and mental workload can explain differences between novice and experienced drivers with regard to how fast they detect hazards in the traffic environment (McKenna & Crick, 1991; Drummond, 1995; Renge, 1998). However, McKenna and Farrand (1999) concluded that the opposite was true, because they found that experienced drivers performed less satisfac‐ torily than novice drivers did when required to carry out a secondary task at the same time as trying to detect hazards. Also, Sagberg and Bjornskau (2006) administered a video‐based hazard perception/reaction test to groups of driv‐ ers who had had licenses for different periods of time, and their results did not show that reaction times decreased significantly with increased driving ex‐ perience. Accordingly, those authors concluded that research findings indicat‐ ing a rapid reduction in crash risk during the first months of licensed driving can probably be explained to only a small extent by improvement in hazard perception skills.
1.3.4
Individuals’ own estimations of their compe‐
tence as drivers
Many investigations have shown that overconfidence is a real problem among novice drivers (e.g., Spolander, 1983; Matthews & Moran, 1986; Gregersen, 1996; McKenna et al., 1991; McKenna, 1993; Deery, 1999; Nolén & Nyberg, 2001). The findings of those studies imply that new licensees have an incorrect perception of the level of their own ability to operate a motor vehicle, so that they believe they are much better drivers than they actually are. This false conception is especially widespread among young male drivers, which in turn
suggests that those individuals find it difficult to assess dangers and risks in traffic. Therefore, overconfidence may explain why novice drivers are more apt than long‐term licensees to commit violations such as speeding, not wear‐ ing a seat belt, and driving too close to the vehicle in front. One explanation for this difference is that the new license holders think they can get away with the transgressions, because in their own opinions they are such excellent driv‐ ers. For the same reason, they feel that even if something unexpected happens, they can manage to avoid a crash. However, Groeger (2006), among others, has argued that young drivers have a sounder apprehension of their driving ability than more experienced drivers do. This researcher means that the prob‐ lems in many questionnaire studies are associated with methodology issues, because subjects are usually asked to compare themselves with an “average” driver instead of a driver that is more similar to themselves. Not surprisingly, opinions vary as to how responses are affected by different ways of posing questions, although experimental investigations have shown that young li‐ cense holders do overestimate their ability to drive (e.g., Gregersen 1996; Nolén & Nyberg, 2001).
1.3.5
Individual and social preconditions
Factors such as social norms, lifestyle, personality, socio‐economic factors, and reasons for driving have been found to be associated with the problems ex‐ perienced by young drivers. Most of these aspects are related to life in general, but despite that they influence the attitudes, decisions, and behavior of indi‐ vidual drivers when driving a car (e.g., Hatakka et al., 2002; Engström et al., 2003; OECD, 2006).
When young people are first learners and later novice drivers, they are going through a period in life that involves a wide variety of changes. For instance, many of them are in the process of becoming independent, which includes testing limits and breaking free from influences and expectations from people whose values and norms have previously been very important in their lives (e.g., parents). Of course, these elements do still have a substantial impact (see below regarding the influence of parents), but in most cases other influences are of growing importance, including things like peer group norms and im‐ ages of the way young people ought to be that are portrayed by the media, popular culture, and society. These social standards have been found to have a pronounced impact on decision‐making process and thus they also influence
young people’s ways of behaving as drivers of motor vehicles (e.g., Engström et al., 2003).
What is known as “the theory of planned behavior” stresses the importance of social norms for the actual actions of an individual (Ajzen, 1991). The social norms are not considered to be some kind of standards that are valid for and accepted by all persons in a society, but instead as subjective norms that a per‐ son believes to exist in relation to significant others. This means that a person can feel pressured into exhibiting a particular behavior that is common to a group of people who are important to that individual. For instance, a driver who is closely associated with a group that likes speeding is also likely to speed, regardless of whether she/he knows that speeding is dangerous or really feels inclined to speed. The subjective norms have been found to have a substantial impact on young drivers, in other words their driving behavior is highly influenced by social pressures from friends or peer groups (Berg, 2001, Evans, 1987; Jessor, 1987, Møller, 2004). Even if many young people are in the process of breaking free from their par‐ ents at the time they start driving, there is also evidence that they are affected by their parents’ driving behavior. Wilson et al. (2006) found an increased risk of at‐fault collisions during the first three years of licensed driving by young people whose parents had been involved in such collisions or had committed speeding offences or other moving violations during the four years prior to their children’s licensure. Therefore, those investigators concluded that it is important to inform parents about their role in influencing their children’s driving behavior and risks. Similar findings indicating that the driving behav‐ ior of parents is passed on to their children have been reported by other re‐ searchers, such as Taubman‐Ben–Ari et al. (2005), Bianchi and Summala (2004) and Ferguson et al. (2001). Also, one study has shown that young drivers that have poor relationships with their parents and other adults (e.g., teachers) are overrepresented in crashes (Beirness & Simpson, 1991), whereas Shope (1997), among others, found evidence that living with both parents is a protective fac‐ tor. In Sweden, investigations conducted by Murray (1998), Berg et al. (1999), and Hasselberg and Laflamme (2003) have shown that knowledge of socio‐ economic factors is needed to understand the problems of young drivers. Murray (1998) observed that the level of crash involvement among young drivers was higher for those who had low grades in school, particularly in
theoretical subjects, and also for those who were members of farming and blue‐collar families, as compared to white‐collar families. Berg and coworkers (1999) found that a learner’s permit was obtained by more youngsters in white‐collar families than in blue‐collar families. Those authors concluded that teenagers in blue‐collar families are unfairly affected by the cost of driving practice, since most of the participants in their study had given responses indi‐ cating that they could not afford to obtain a learner’s permit. Hasselberg and Laflamme (2003) reported that young drivers whose parents were lower‐ income white‐collar workers, blue collar workers, farmers, or entrepreneurs suffered more injuries than did young drivers whose parents were middle‐ and high‐income white‐collar workers.
Another important aspect is that there are different motives and goals for driv‐ ing a motor vehicle, which influences the risk of being involved in crashes. Gregersen and Berg (1994) found that high‐risk drivers were characterized by a lifestyle that included driving behaviors such as showing off and sensation‐ seeking. In another study, Møller (2004) noted that risk‐taking behavior repre‐ sents one way of achieving status and position in the group that is of impor‐ tance to the driver. It has also been shown that motivation for driving is influ‐ enced by the situation in which the driving takes place, for example, a person can drive aggressively in the company of peers but more sensibly in the pres‐ ence of parents or children (e.g., Rolls & Ingham, 1992). In 1988, Klemenjak and Hutter (cited in Gregersen, 2003) found that “disco fans” drove less safely when they went to dances on the spur of the moment late in the evening than when they spent the whole evening at such nightspots. The reason for this dif‐ ference was assumed to be that the spur of the moment decisions were more often made after imbibing alcohol and also frequently entailed driving long distances to reach the disco. In agreement with that, are results reported by Schulze (1990). Clarke et al. (2006) suggest that crashes that occur during darkness and involve drivers 17–25 years of age are not related to poor visibil‐ ity but rather to the driving behavior of the young licensees under these cir‐ cumstances, meaning that many of the crashes are due to voluntary risk be‐ haviors displayed during ‘recreational’ driving. Personality is an aspect that has also been found to influence driving behavior and consequently the crash risk (e.g., Ulleberg, 2002; Gregersen, 2003; OECD, 2006). Extensive research has been conducted to discern consistent associations in this context. Factors such as social deviance, aggression, impulsiveness, hos‐ tility, emotional liability, and low altruism have been found to be related to
