Accident Analysis and Prevention 40 (2008) 1524–1530

Contents lists available at ScienceDirect

Accident Analysis and Prevention

journa l homepage: www.e lsev ier .com/ locate /aap

The development of a method to measure speed adaptation to trafficcomplexity: Identifying novice, unsafe, and overconfident drivers

Saskia de Craena,∗, Divera A.M. Twiska, Marjan P. Hagenziekera,Henk Elffersb, Karel A. Brookhuisc,d

a SWOV Institute for Road Safety Research, PO Box 1090, 2260 BB Leidschendam, The Netherlandsb Netherlands Institute for the Study of Crime and Law Enforcement NSCR, PO Box 792,2300 AT Leiden, The Netherlandsc Experimental and Work Psychology, Faculty of Behavioural and Social Sciences, University of Groningen,Grote Kruisstraat 2/1, 9712 TS Groningen, The Netherlands

tion,

perfotuatio(almo. As

icipaned beomplAdaped drrivingtive a

sessm

d Faculty Technology, Policy and Management, Transport Policy and Logistics’ OrganisaDelft University of Technology, Jaffalaan 5, 2628 BX Delft, The Netherlands

a r t i c l e i n f o

Article history:Received 16 February 2007Received in revised form 21 March 2008Accepted 30 March 2008

Keywords:Young novice driversAdaptation to task demandsSelf-assessment of skillsCalibration

a b s t r a c t

To monitor novice driveradaptation to the traffic siscenes presented in twothe situation’s complexityto previous pictures, partdifference in reported speadapt their speed to the cperformed worse on thesituation) than experiencidentified in an on-road dthe Adaptation Test is effecand that incorrect self-as

effect on speed adaptation.

1. Introduction

To explain the high accident risk of novice drivers, many con-tributing factors have been identified. Basically, these factors canbe grouped into two categories, namely factors related to task per-formance processes which are relevant for all novice drivers, andfactors related to motivational factors like intentional risk takingand sensation seeking (e.g. Jonah, 1997; Ulleberg, 2002), whichare dominant in relatively small sub groups. Of the factors relatedto task performance, the high risk of novice drivers is frequentlyattributed to the limited automation of driving routines (Engstromet al., 2003; Fuller, 2002; Groeger, 2000; Rasmussen, 1986), whichleads to a higher mental workload for novice drivers compared toexperienced drivers (De Waard, 2002; Detweiler and Schneider,1991; Patten et al., 2006).

∗ Corresponding author. Tel.: +31 70317 3333; fax: +31 70320 1261.E-mail address: Saskia.de.Craen@swov.nl (S. de Craen).

0001-4575/$ – see front matter © 2008 Elsevier Ltd. All rights reserved.doi:10.1016/j.aap.2008.03.018

rmance in the first years of solo driving, a test aimed at assessing speedn was developed and evaluated. The Adaptation Test consisted of 18 trafficst) identical photographs, which differed in one single detail, increasingthe pictures were presented randomly and participants could not returnts were kept unaware of the complexity differences in the pictures. The

tween the two pictures was used as an indication of drivers’ inclination toexity of the traffic situation. Results showed that novice drivers (n = 434)tation Test (i.e. less often reported a lower speed in the more complexivers (n = 173). In addition, unsafe drivers and overconfident drivers, asassessment, performed worse on the Adaptation Test. This indicates thatt measuring adaptation of driving speed to the complexity of the situation;ent, and overestimating driving skills in particular, may have a negative

© 2008 Elsevier Ltd. All rights reserved.

Although cognitive overload has shown to have a significantcontribution, an extra factor needs to be taken into account. Asthe driving task is self-paced (Fuller, 2005), task demands can bemodified by the driver. Thus, in theory, novice drivers can decreasetask demands to fit their (deficient) level of automated driving,for example by reducing speed or increasing headway. Whetherthe driver can be successful in this, is dependent on three factors:(a) the correct assessment of personal driving skills; (b) the cor-rect assessment of the complexity of the driving task; and (c) thecorrect selection of behaviours that change task demands effec-tively. In a psychological context, calibration refers to the degreeto which the driver perceives task demand and his/her capabilityaccurately (Kuiken and Twisk, 2001; Mitsopoulos et al., 2006). Fig. 1shows a schematic representation of the calibration process, whichis modified after Brown’s (1989) ‘model of subjective safety’. Obser-vations of driving behaviour indicate that novice drivers are poorlycalibrated, as they tend to drive with higher speeds and shorterheadways than experienced drivers (Engstrom et al., 2003; Quimbyand Watts, 1981). In addition, they are more likely, compared toother age groups, to engage in secondary behaviour (e.g. making a

is and

S. de Craen et al. / Accident Analys

Fig. 1. Working model of the calibration process.

telephone call) while driving (Sayer et al., 2005), thereby increasingtask demands.

This pattern of inadequate calibration is assumed to be relatedto the high crash risk of young novice drivers (Brown and Groeger,1988; Brown et al., 1987; Gregersen, 1995; Mayhew and Simpson,1995). The fact that young drivers’ performance improves withpractice may be due to increased automation in combination withimproved self-assessment of skills, improved perception of the sit-uation’s complexity, and as a result, adequate adaptation to taskdemands (see Fig. 1).

In The Netherlands, a longitudinal study was initiated to studythe development of calibration over time. In this study, self-assessment of skills (left-hand side of Fig. 1) is measured bycomparing the drivers’ opinion of their driving skills to on-roaddriving performance (see also De Craen et al., 2007). To measurethe effect of perceived complexity on adaptation to task demands

Fig. 2. Two examples of a simple (left) and comp

Prevention 40 (2008) 1524–1530 1525

a new method was developed. This paper reports the results of thedevelopment of this method: the Adaptation Test.

The Adaptation Test consisted of 18 traffic scenes presented intwo (almost) identical photographs, which differed in one singledetail, thereby increasing the complexity of the situation (right-handside of Fig. 1). As the pictures were presented randomly and par-ticipants could not return to previous pictures, participants werekept unaware of the varying level of complexity. With the use ofphotographs, task complexity can be manipulated, which is fairlyimpossible in a more natural environment. In addition, the easyimplementation of photographs on a website makes this test very

cost-effective and applicable in a longitudinal study with manyparticipants.

To study the extent to which complexity affects adaptation totask demands (bottom Fig. 1) we used driving speed, because reduc-ing speed is the most straightforward way to decrease task demands(e.g. Quimby and Watts, 1981). The respondents were asked toassess at what speed they would drive in the depicted situations.Neither in the instructions nor in the pictures was explicit referencemade to the legal speed limit. The situations were selected in sucha manner that the “extra” element would increase the complex-ity of the situation, without legally obliging the driver to lower hisspeed. In Fig. 2, for instance, a driver is allowed to drive at the samespeed, and has right of way, in both situations. Although absolutespeeds are relevant, speed assessment on the basis of photographsis unreliable. Therefore, only the direction of the difference betweenreported speed in the simple and complex version was used to testfor differences between groups. A response was considered ‘cor-rect’ if the reported speed was lower in the complex situation thanin the corresponding simple situation. A higher speed in the com-plex situation was regarded to be equally unwanted as no speedchanges; for both cases such a response was considered ‘incorrect’.

lex (right) situation in the Adaptation Test.

is and

1526 S. de Craen et al. / Accident Analys

Table 1Design

n October 2005

GroupExperienced 173 6 situationsNovice I 297 6 situationsNovice II 137

A basic concept in the development of the test was not to mea-sure a driver’s hazard perception skill, but the adaptation of hisspeed as a result of the specific situation and his beliefs about hisown driving skills. For example, a driver may spot the danger (orhazard), but not decrease speed because he believes to be capableof handling the situation. Because of this, there was no time con-straint for participants to respond, as is often the case in (classic)hazard perception tests (see e.g. Groeger, 2000). As it was importantthat participants did not become aware of the mechanism behindthe Adaptation Test and consciously compare pairs of photos, thephotos were randomly distributed throughout a web based ques-tionnaire, without the possibility to scroll back to previous photos.A pilot study, using a similar photo judging task, revealed that par-ticipants are not aware of the differences between the photographswhen they are presented in random order. This was not checkedagain in the present study, because the Adaptation Test is part ofa longitudinal study and had to be used again with the same par-ticipants. Questioning the participants on the differences betweenphotographs may expose the underlying mechanism.

1.1. Hypotheses

The objective of this study is to evaluate the Adaptation Test.Poor calibration may explain the high crash rate of young, novicedrivers, so it is assumed that novice and/or unsafe drivers willincorrectly assess their driving skills more often, will underesti-mate the demands of the driving task, and will consequently showless adaptation to task demands. The Adaptation Test should meetthe following basic criteria:

1. In general, drivers adapt their speed to the complexity of the sit-uation (report lower driving speeds in the more complex trafficsituations);

2. Inexperienced, overconfident, or unsafe drivers perform worseon the Adaptation Test than more experienced, well-calibratedand safe drivers.

To study this, we compared experienced and novice drivers ontheir self-reported skills, their objective skills as assessed in an on-road driving assessment, and their performance on the AdaptationTest.

2. Method

2.1. Participants

Two groups of drivers, novice and experienced drivers, wereselected for this study. In collaboration with the CBR, the DutchDriving Test Organisation, all drivers under the age of 25, who hadpassed their driving test during the last 2 weeks in September 2005,were invited to participate in this project. During these weeks wecontacted 553 drivers of whom 509 agreed to participate. This is aresponse rate of 92% for the novice drivers. RDW, the Dutch Vehi-cle Technology and Information Centre, provided the names of 999experienced drivers who held their licence for more than 10 years,and were not older than 50. Of the drivers that responded positivelyto an invitation to participate in this project only 179 participants

Prevention 40 (2008) 1524–1530

January 2006 April 2006 May 2006

6 situations Driving assessment (n = 47) 6 situations6 situations Driving assessment (n = 46) 6 situations

Driving assessment (n = 37) 6 situations

could be included in the project, for financial reasons. As a result,the response rate for experienced drivers was 18%.

After some initial drop-out (12%) the total sample consisted of607 drivers; 173 experienced drivers (drop-out = 3%) and 434 novicedrivers (drop-out = 15%). The experienced drivers had a mean age of41 (S.D. = 5.6 years), and 49% of the experienced drivers were male.The novice drivers had only 2 weeks of driving experience sincepassing their driving test when they filled in the first questionnaire;they had a mean age of 20 (S.D. = 1.8 years), 52% being male.

2.2. Design

The results reported here were collected within a larger lon-gitudinal study, in which novice and experienced drivers fill inquestionnaires during 2 years. With this design, it is possible thatdrivers who, on a regular basis, fill out questionnaires about trafficbehaviour and traffic safety will become more aware of the risksof driving than drivers who perhaps never even think about trafficsafety. More specific for the Adaptation Test, it is possible that par-ticipants figure out that each situation is depicted in pairs, and tryto find the differences. That is, they ‘learn’ how to successfully takethe Adaptation Test. In order to rule out these and other learningeffects (see also Bouchet et al., 1996) as alternative explanationsfor our results, the novice drivers were randomly divided into twogroups. A total of 297 novice drivers (Novice I) started filling in ques-tionnaires and diaries in October 2005, the remaining 137 novicedrivers (Novice II) were selected for participation in October 2005,but did not start filling in the questionnaires until May 2006 (seeTable 1). Additionally, 130 drivers (83 novice drivers and 47 experi-enced drivers) were randomly drawn from the sample, and invitedto participate in an on-road driving assessment.

2.3. Instruments

2.3.1. QuestionnaireThe participants in the current study completed questionnaires

on a website. Besides the assessment of the traffic situations,

they also reported background information (age, occupation, etc.),responses to items of the Driver Behaviour Questionnaire (Parkeret al., 1995), attitudes towards traffic safety, and own drivingbehaviour; these results will be reported elsewhere. The partic-ipants assessed six new traffic situations in three subsequentsessions (see Table 1). The photographs of traffic situations wererandomly distributed throughout the questionnaire, without thepossibility to scroll back, under two conditions: (1) in half of thesituations the participants viewed the complex situation first andin the other half they viewed the simple situation first; and (2)a simple and complex version of the same traffic situation neverimmediately succeeded each other. In this study, the eighteen traf-fic situations collected in the first three questionnaires were anal-ysed as one. In the final three questionnaires, planned in 2006/2007,the participants will assess the same eighteen traffic situations.

2.3.2. On-road driving assessmentTo obtain an objective measure of driving skills (as opposed

to their own assessment of their driving skills), the participantswere assessed in an on-road driving assessment. These assessmentswere carried out by three professional driving examiners from two

is and Prevention 40 (2008) 1524–1530 1527

Table 2

S. de Craen et al. / Accident Analys

Driving Test Organization locations in the cities of Rotterdam andRijswijk. In both cities the participants drove for half an hour on afixed route with different road types. The routes in both cities wereas similar as possible. On a few occasions an alternative route wasdriven to avoid traffic jams. In order to make the trip resemble anordinary one, the participants were instructed to follow signpostsalong part of the route (e.g. “Use the signposts to find your wayto the station”). For the remaining part, “go left/go right” instruc-tions were used to give directions to the driver. The examiners wereasked to rate the drivers on “the ability to drive safely”, on a scalefrom 0 to 10; 5.5 being the pass–fail mark in a real driving test.

The on-road driving assessment was chosen as a measure ofdriving skills, because, for example, the number of accidents as anindicator would have been too few for reliable analyses. The drivingassessment relies on the quality of the assessment of the examiners,for which doubts have been expressed with respect to the reliabilityand validity. For example, Baughan et al. (2005) argues that there isa difference between how a driver acts during an assessment withan observer, and how a driver behaves in everyday traffic, i.e. with-out an observer. An empirical study (Quimby et al., 1999) in whichdriving behaviour during an assessment was compared to drivingbehaviour in a natural setting, indicated that a driving assessment isa good reflection of actual driving style. Because the driving assess-ment in the present study is similar to the one in Quimby et al.(1999), we consider it an appropriate tool to measure driving skills.

The driving examiners were aware of the objective of the study.As this could bias their assessment (e.g. young drivers are ‘mostlikely novice drivers’), examiners were explicitly informed that thegroup consisted of different types of drivers, some older who rarelydrove, and some younger who drove on a regular basis (e.g. as aprofessional courier). In addition, the participants were instructednot to discuss their prior driving experiences with the examiner.

2.4. Data analysis

The direction of the difference between reported speed in thesimple and complex version was used for between group analyses.One point was assigned when the reported speed was lower in thecomplex situation than in the corresponding simple situation. Ahigher speed in the complex situation was regarded to be equallyunwanted as was no speed change, and no points were assigned.So, a participant could have a maximum score of 18 in the case thatall situations were assessed correctly. When for one of the photosno speed was reported, or the speed was unlikely to be correct (e.g.

555 km/h), the score was coded as ‘missing’ (in 0.3% of all entries).Because not all participants had 18 values (i.e. the Novice II groupassessed only six situations), the percentage of correct responseswas used for analysis and not the total sum of scores. This adapta-tion score (i.e. percentage of correct responses) was only calculatedfor participants with at least six valid scores and was used to testfor differences between groups.

Analysis of variance of the adaptation score was used to testfor significant differences between groups (˛ = .05). For the evalu-ation of the situations, repeated measures analysis with reportedspeed on the simple and complex situation (i.e. not the adaptationscore) as the within-subjects-factor was used (˛ = .05). Besides sig-nificance of the results, the effect size (Partial eta squared, �2

P) isalso considered with �2

P = .01 as a small, �2P = .06 as a medium,

and �2P = .14 as a large effect size (Cohen, 1988).

3. Results

No significant differences were found between the Novice I andNovice II groups concerning the questionnaire completed in May2006, indicating that there is no evidence of a learning effect in

Mean and standard deviation (S.D.) of the reported speed (km/h) in the eighteensituations of the Adaptation Test

Situation Legal limit Simple Complex

Mean S.D. Mean S.D.

1 50 41.9 8.1 37.6 9.22 50 38.2 8.5 36.0 9.83 50 38.9 10.6 31.1 12.34 50 32.2 8.5 30.5 9.45 50 43.6 8.3 41.5 8.86 120 102.1 14.9 92.6 18.07 50 46.5 7.4 44.3 8.38 80 82.5 12.3 65.7 13.49 30 31.5 8.4 29.0 7.7

10 50 47.7 8.7 46.3 10.311 50 40.0 9.0 28.6 11.412 80 74.4 16.1 74.2 14.913 80 66.4 13.6 61.1 14.014 50 31.8 10.0 32.9 9.715 50 46.1 8.2 44.5 8.316 50 44.7 9.8 35.6 11.917 50 36.1 9.1 34.6 8.918 50 48.2 8.9 42.1 10.7

this study. Thus, the Novice I and Novice II group were consideredas one group of novice drivers for the following analyses. The adap-tation score was less sensitive for the Novice II group compared tothe Novice I group (i.e. based on six situations instead of 18). As aresult of this, the significance tests became more conservative (i.e.smaller chance of finding a significant result when there is, in fact,a difference).

3.1. Evaluation of the situations

Table 2 shows the mean and standard deviation of the reportedspeed for the eighteen situations in the Adaptation Test. Repeatedmeasures analysis, with reported speed on the simple and complexsituation as within-subjects-factor, was carried out for all situa-tions and for each of the eighteen situations separately. The resultsshowed a main effect of complexity; the mean reported speed wasgenerally lower in the complex situations than in the simple situa-tions (F1,603 = 655.10; p < .001), with a large effect size (�2

P = 0.52).Repeated measures analysis for each situation separately

showed that the effect of complexity on adaptation was signif-icant (p < .01) in sixteen out of eighteen situations. In situation12 (see Table 2), depicting a rural road (speed limit 80 km/h), no

significant difference was found between the complex and sim-ple situation whereas in situation 14, depicting an urban situation(speed limit 50 km/h), a significant difference was found in theunexpected direction. These two situations were not excluded fromthe analyses, because no definite explanation could be given as towhy these two situations produced a different result than expected.The two situations cause the Adaptation Test to be less sensitive,and therefore making the statistical tests more conservative.

3.2. Experience

Fig. 3 shows the adaptation score for the experienced and novicedriver groups, for both male and female drivers. The solid lines indi-cate the median, the boxes the interquartile range, and the whiskersindicate the minimum and maximum values. The analysis of vari-ance shows that there is a significant main effect of both experience(F1,603 = 26.77; p < .001) and gender (F1,603 = 8.42; p < .01) on theadaptation score. Experienced drivers performed better (45% cor-rect responses) than novice drivers (36% correct responses). Malesperformed better than females (respectively 43% and 38% correctresponses). The effect size is considered small for both experience

1528 S. de Craen et al. / Accident Analysis and

into safe and unsafe drivers based on the examiners’ safe driving

Fig. 3. Effect of experience level and gender on adaptation score; higher percentagereflects better adaptation.

level (�2P = .04) and gender (�2

P = .01). No significant interactionwas found between experience level and gender.

3.3. Driving skills

Pearson correlation coefficient between the examiners’ “abil-ity to drive safely” score and the adaptation score from thephotograph assessments indicated a moderate but significant cor-relation (rn=130 = .31; p < .01). Drivers who were rated higher bythe examiners performed significantly better on the AdaptationTest.

Based on the scores provided by the examiners, two types ofdrivers were distinguished, safe and unsafe drivers. Drivers whoscored less than 5.5 were considered to be unsafe drivers, and wouldhave failed if the driving assessment was an actual driving test. Notethat the distinction between ‘safe’ and ‘unsafe’ drivers is based onthe assessment; there is no information whether these drivers actu-ally portray safe or unsafe driving behaviour in every day life. Usingthis criterion, 28 drivers were considered unsafe drivers and 102drivers were considered safe drivers. The unsafe drivers respondedcorrectly in 28% of the situations. The safe drivers gave correctresponses in 42% of the situations. Fig. 4 shows the median and(interquartile) range of the adaptation score for safe and unsafe

Fig. 4. Effect of safe driving on adaptation score.

Prevention 40 (2008) 1524–1530

Table 3Calibration groups

On-road driving assessment Total

Passed Failed

ConfidenceHigh 75 (58%) 20 (15%) 95 (73%)Low 27 (21%) 8 (6%) 35 (27%)

Total102 (79%) 28 (21%) 130 (100%)

drivers. The analysis of variance showed that there was a mod-erate (�2

P = .09) and significant difference (F1,128 = 12.45; p < .01)between safe and unsafe drivers concerning the way they assessedthe depicted traffic situations. No interaction effect between genderand experience was found.

3.4. Self-assessment of skills

In the questionnaire, the participants judged, on a five-pointscale, how confident they were about their own driving skills. Basedon these scores, participants were divided into two groups: (a) highconfidence drivers (drivers who said they were (very) confident);and (b) low confidence drivers (drivers who responded with neutralconfidence or no confidence at all). Table 3 shows the cross tabula-tion of the drivers’ own assessment of skills and the categorization

scores. This cross tabulation results in three ‘calibration’ groups:(a) 83 (75 + 8) well-calibrated drivers (share the examiner’s opin-ion about their driving skills), (b) 20 overconfident drivers (wereconfident about their driving skills, but failed the driving assess-ment) and (c) 27 insecure drivers (were less confident about theirdriving skills, but passed the driving assessment).

Chi-square analysis indicated that the group of experienceddrivers consisted of significantly more ‘well-calibrated’ drivers(79%) than the group of novice drivers (55%) (X2

1,n=130 = 7.06;p < .01). Within the group of novice drivers 21% were overconfident,compared to 6% of the experienced drivers, while 24% were inse-cure, compared to 15% of the experienced drivers. There was nosignificant difference between males and females in the three cal-ibration groups. There was some overrepresentation of females inthe insecure group, but not in the overconfident group. These cali-bration groups were used in a further analysis of the assessment ofsituations depicted on the photographs.

Fig. 5 shows median and (interquartile) range of the adaptationscore for the three calibration groups. A moderate (�2

P = .09) andsignificant difference between the calibration groups (F2,127 = 6.46;

Fig. 5. Effect of self-assessment on adaptation score.

is and

S. de Craen et al. / Accident Analys

p < .01) was found. The well-calibrated drivers scored an average of43% correct responses, the insecure group scored 38%, whereas theoverconfident group of drivers scored worst with only 26% correctresponses. Post hoc Bonferroni tests show that the overconfidentdrivers significantly differed from the well-calibrated drivers andthe insecure group (p < .01). No significant interaction effect wasfound between calibration groups and gender.

The calibration groups were constructed using the scores on thedriving assessment and confidence level. So in theory it is possiblethat the difference in well-calibrated versus insecure and overcon-fident drivers can be explained completely by the performance onthe driving assessment, and that confidence level does not influ-ence this effect. In this case there would not be an interaction effectbetween driving assessment score and confidence level. Analysisof variance showed a main effect of passing or failing the drivingassessment on adaptation score (F1,126 = 6.36; p < .01) but no maineffect of confidence on adaptation score. A small (�2

P = .02) inter-action effect was found, however, due to low power (30%), only 8drivers failed the assessment and had predicted this, this effect wasnot significant.

4. Discussion

The objective of this study was to evaluate a new methodol-ogy, using traffic situations depicted on photographs, to measureadaptation of driving speed (the Adaptation Test). The results of thereported mean speed show that drivers reported a lower speed forthe complex situation in sixteen out of eighteen situations. It wasexpected that novice and/or unsafe drivers incorrectly assess theirdriving skills and the demands of the driving task more frequentlyand consequently adapt their speed less often to the situation thanexperienced and/or safe drivers. For the analysis of the differencesbetween these groups the adaptation score was calculated. If a par-ticipant reported a lower speed in the more complex situation, hegained one point, while no ‘negative’ points were assigned. Analy-sis of these adaptation scores indicate that novice drivers are worseat adapting speed to the situation than experienced drivers.

A sub-sample of 130 drivers (83 novice drivers and 47 expe-rienced drivers) participated in an on-road driving assessment. Asignificant positive correlation between performance on the on-road driving assessment and assessment of the traffic situationswith the photographs was found, indicating that drivers with a highscore for ‘safe driving’ adapt their speed more often to the situation.When distinguishing between safe and unsafe drivers, based on the

performance on the driving assessment, unsafe drivers respondedcorrectly (i.e. reported a lower speed in the more complex situation)to 28% of the situations, whereas, safe drivers responded correctlyto 42% of the situations.

The performance on the on-road driving assessment was usedto construct three self-assessment groups: well-calibrated drivers,overconfident drivers, and insecure drivers. The results on assessingtraffic situations were as expected: well-calibrated drivers per-formed best, the insecure group performed somewhat worse, whilethe overconfident drivers performed significantly worse on theadjustment to traffic situations. This is consistent with previousresearch where overconfident drivers, rather than insecure drivers,were described as a problem group (Gregersen, 1996; Mathewsand Moran, 1986). Because the calibration groups were constructedusing the scores from the driving assessment and confidence rat-ings, in theory, the results could have been caused by one of thesefactors (and not the combination). Because the power to test forinteraction effects was too low (i.e. too few participants in aggre-gated groups) this could not be tested.

Experienced drivers and safe drivers were expected to be betterat assessing their skills, i.e. more aware of the limits of their skills.

Prevention 40 (2008) 1524–1530 1529

All three subgroups, experienced drivers, safe drivers, and driverswho correctly assessed their driving skills, indeed performed bet-ter at adapting speed to the situation. However, there was someinconsistency as well. Males were significantly better at adaptingspeed to the situation than female drivers. This is an unexpectedresult, since (young) males have a higher crash risk than (young)female drivers (OECD, 2006). In addition, there was no significantdifference between males and females in the calibration groups. So,although the difference between males and females on the adapta-tion score was small (�2

P = .01), there appears to be a gender effectthat cannot be explained by the self-assessment of skills. Unfor-tunately the sample was too small to find significant interactioneffects between self-assessment groups and gender.

With respect to the newly developed method, two pointsdeserve attention. First of all, the test uses ‘reported’ speed to mea-sure adaptation to task demands which is not the same as actualspeed behaviour. However, using reported speed enables completecontrol over the complexity level of the traffic situation. In the cur-rent study, we take into account that reported speed is not thesame as actual speed choice; therefore, the between group anal-yses are based on the difference between the reported speed in thecomplex and simple traffic situation, rather than absolute speed.Secondly, none of the groups performed very well on the test. Ofall the subgroups, experienced drivers performed best on the test,but still only scored correctly on 45% of the situations; in morethan half of the situations, even experienced drivers do not reporta decreased speed for the more complex situation. A closer lookat the situations revealed that often the same speed is reportedfor both the simple and complex situation. Despite explicit instruc-tions that the speed limit in the traffic situations should be ignored,reported speed often fluctuated around the speed limit. It seemsthat for some traffic situations the road design determines speedchoice more than the specific circumstances do (i.e. complexity ofthe situation). It can be argued that there was not enough differ-ence between the simple and complex photograph. Increasing thisdifference could indeed lead to more reported differences betweenthe complex and simple version of the traffic situation. However, itis not so obvious that this would improve the test in differentiatingbetween novice and experienced drivers. Because the complexitywould be more obvious, novice drivers would probably also be bet-ter at recognizing this difference. Moreover, the photographs wereconstructed with the restriction that in both photographs the prior-ity situation and legal limit was the same. With this restriction, thedifferences between the simple and complex photographs had to be

a subtle one. In any case, when a different speed was reported, thiswas almost always in the correct direction, i.e. a decreased speedfor the more complex situation.

Overall, the results indicate that the Adaptation Test is effectiveat measuring adaptation of driving speed to the situation; incorrectself-assessment of skills and especially overconfidence of drivershas an effect on this adaptation. The test is not very sensitive indifferentiating between individual drivers (i.e. even experienceddrivers do not report a decreased speed for the complex situation inmore than half of the situations). However, because of its consistentability to differentiate between groups of drivers (experienced ver-sus novice drivers, safe versus unsafe drivers, and well-calibratedversus overconfident drivers), the Adaptation Test is a promisingmethod for assessing calibration. The Test is currently being usedin a 2-year longitudinal study into the development of calibrationin young novice drivers.

Acknowledgement

This study was facilitated by the Dutch Driving Test Organisa-tion.

is and

Mayhew, D.R., Simpson, H.M., 1995. The Role of Driving Experience: Implications for

1530 S. de Craen et al. / Accident Analys

References

Baughan, C., Gregersen, N.P., Hendrix, M., Keskinen, E., 2005. Towards Europeanstandards for testing: Final report. Commission Internationale des Examens deConduite Automobile CIECA, Brussels.

Bouchet, C., Guillemin, F., Briancon, S., 1996. Nonspecific effects in longitudinal stud-ies: impact on quality of life measures. Journal of Clinical Epidemiology 49 (1),15–20.

Brown, I.D., 1989. How Can We Train Safe Driving? Traffic Research Centre, Universityof Groningen, Haren, The Netherlands.

Brown, I.D., Groeger, J.A., 1988. Risk perception and decision taking during thetransition between novice and experienced driver status. Ergonomics 31 (4),585–597.

Brown, I.D., Groeger, J.A., Biehl, B., 1987. Is driver training contributing enoughtowards road safety? In: Rothengatter, J.A., de Bruin, R.A. (Eds.), Road Users &Traffic Safety. Van Gorcum, Assen, The Netherlands, pp. 135–156.

Cohen, J., 1988. Statistical Power Analysis for the Behavioral Sciences. LawrenceErlbaum Associates, Hillsdale, NJ.

De Craen, S., Twisk, D.A.M., Hagenzieker, M.P., Elffers, H.E., Brookhuis, K.A., 2007. Doyoung novice drivers overestimate their driving skills? In: Proceedings of theYoung Researchers Seminar, CDV, Brno.

De Waard, D., 2002. Mental workload. In: Fuller, R., Santos, J.A. (Eds.), Human Factorsfor Highway Engineers. Elsevier Science Ltd., Oxford, pp. 161–175.

Detweiler, M., Schneider, W., 1991. Modeling the acquisition of dual-task skill in aconnectionist/control architecture. In: Damos, D.L. (Ed.), Multiple-Task Perfor-mance. Taylor & Francis, London, pp. 69–99.

Engstrom, I., Gregersen, N.P., Hernetkoski, K., Keskinen, E., Nyberg, A., 2003. Youngnovice drivers, driver education and training. Report no. 491A. Swedish NationalRoad and Transport Research Institute, Linkoping, Sweden.

Fuller, R., 2002. Human factors and driving. In: Fuller, R., Santos, J.A. (Eds.), HumanFactors for Highway Engineers. Elsevier Science Ltd., Oxford, pp. 77–97.

Fuller, R., 2005. Towards a general theory of driver behaviour. Accident Analysis &Prevention 27 (3), 461–472.

Gregersen, N.P., 1995. Prevention of road accidents among young novice car drivers.Dissertation Thesis. Linkoping University, Linkoping, Sweden.

Gregersen, N.P., 1996. Young drivers’ overestimation of their own skill—an experi-ment on the relation between training strategy and skill. Accident Analysis &Prevention 28 (2), 243–250.

Prevention 40 (2008) 1524–1530

Groeger, J.A., 2000. Understanding Driving—Applying Cognitive Psychology to aComplex Everyday Task. Psychology Press Ltd., Hove, East Sussex, UK.

Jonah, B.A., 1997. Sensation seeking and risky driving: a review and synthesis of theliterature. Accident Analysis & Prevention 29 (5), 651–665.

Kuiken, M.J., Twisk, D.A.M., 2001. Safe driving and the training of calibration. Reportno. R-2001-29. SWOV Institute for Road Safety Research, Leidschendam, TheNetherlands.

Mathews, M.L., Moran, A.R., 1986. Age differences in male drivers’ perception ofaccident risk: the role of perceived driving ability. Accident Analysis & Prevention18, 299–313.

the Training and Licensing of New Drivers. Insurance Bureau of Canada, Toronto,Ontario.

Mitsopoulos, E., Triggs, T., Regan, M., 2006. Examining novice driver calibrationthrough novel use of a driving simulator. In: Proceedings of the SimTecT 2006Simulation Conference: Challenges and Opportunities for a Complex and Net-worked World (SimTecT 2006), Melbourne, Australia.

OECD, 2006. Young drivers: the road to safety. European Conference of Ministers ofTransport ECMT/CEMT, Joint OECD/ECMT Transport Research Centre, Paris.

Parker, D., Reason, J.T., Manstead, A.S.R., Stradling, S.G., 1995. Driving errors, drivingviolations and accident involvement. Ergonomics 38 (5), 1036–1048.

Patten, C.J.D., Kircher, A., Ostlund, J., Nilsson, L., Svenson, O., 2006. Driver experienceand cognitive workload in different traffic environments. Accident Analysis &Prevention 38 (5), 887–894.

Quimby, A., Maycock, G., Palmer, C., Grayson, G.B., 1999. Drivers’ Speed choice: anin-depth study. TRL Report 326. Transport Research Laboratory, Crowthorne,Berkshire.

Quimby, A.R., Watts, G.R., 1981. Human factors and driving performance. LaboratoryReport 1004. Transport and Road Research Laboratory, Crowthorne, Berkshire.

Rasmussen, J., 1986. Information processing and human-machine interaction; anapproach to cognitive engineering. North-Holland, Amsterdam, The Nether-lands.

Sayer, J.R., Devonshire, J.M., Flannagan, C.A., 2005. The effects of secondary tasks onnaturalistic driving performance. Report no. UMTRI-2005-29. The University ofMichigan, Transportation Research Institute, Ann Arbor, Michigan, USA.

Ulleberg, P., 2002. Personality subtypes of young drivers. Relationship to risk-takingpreferences, accident involvement, and response to a traffic safety campaign.Transportation Research Part F: Traffic Psychology and Behaviour 4 (4), 279–297.