During their first several months with a solo license, novice teen drivers are overrepresented in crashes, particularly rear-end, intersection and run-off-road crashes. Their involvement in these crashes appears to be due to six poorly developed skills. These include both the tactical and strategic components of hazard anticipation, hazard mitigation and attention maintenance. No training program has addressed all six skills at one time using widely available software (PowerPoint, VBA) that could be downloaded from the internet and run on any computer or mobile device which had PowerPoint. No previous training program specifically targeted for training the exact three crash types in which novice drivers are most likely to be involved. And no previous training program has exposed drivers to the training more than once. The question this study addressed is whether learning could be accelerated so that the risky behaviors that are believed to lead to crashes among novice teen drivers during their first several months of solo licensure were reduced in a few hours. There is reason to believe that an omnibus training program that addressed all six skills across the three most critical crash types could have an impact not only on the behaviors known to reduce crashes, but also on actual crashes. The reason for this optimism in an area of research, novice driver training, where training program after training program has failed to reduce crashes, is based on the results from two recent studies. The first study was conducted by the National Highway Traffic Safety Administration. They evaluated a tactical hazard anticipation-training program using 2,500 California 16-, 17- and 18-year-old drivers as the experimental group and an equivalent number of 16-, 17- and 18-year-old drivers as the placebo. Training was delivered immediately before solo licensure. The crashes of the trained males over the year following training were reduced significantly, by 23%. The crashes of the trained females did not change significantly. Three things are notable. First, the treatment group was exposed to training for only 17 minutes on average. Second, only one of six skills was targeted for training (tactical hazard anticipation). Third, training was administered only once. Yet, despite this, as noted there was a 23% decrease in crashes for the male teen drivers. A similar decrease in crashes was reported in altogether different study. Unfortunately, the reduction in crashes in this latter could not be separated out by gender. With the above in mind, it would appear that a training program that addressed all six skills across the three different crash types could not only lead to a decrease in crash rates of males at least as large as the decrease in the crash rates of males in the NHTSA study, but could also lead to such reductions in the crash rates for females. The expectation that it would do such for females is based on the evidence that the crashes for females are less severe on average than for males (e.g., females are more likely to be involved in rear-end crashes) and are more likely to involve failures to maintain attention (e.g., novice teen female drivers are more than twice as likely to talk on the cell phone as novice teen male drivers). The training program evaluated in California by the National Highway Traffic Safety Administration did not address these two areas of concern for female teen drivers. It included few if any scenarios from the less severe crash types and did not target attention. In summary, seven questions were addressed in this study: (1) whether an omnibus training program could be developed targeting all six skills across all three types of crash scenarios; (2) whether the training program could be developed using open source software that was easily downloadable from the internet, that could readily be viewed on a variety of devices, and that was freely available; (3) whether the training program could be administered in one two-hour session without reducing the effectiveness of the training of the individual skills, skills which if taught one at a time using current programs targeting only single skills would take at least 4.5 hours; (4) whether the effect of training would be present across all skills and crash types immediately after training; (5) whether the effect of training would persist three to six months after the first administration; (6) whether an additional training session could improve the skills more; and (7) whether the training would impact males and females differently immediately after training and three to six months later. To begin, in the current study a PC-based training program, ACCEL (Accelerated Curriculum to Create Effective Learning), was developed in PowerPoint with VBA (Visual Basic for Applications) embedded in it to provide added functionality. The training program can be downloaded from the internet onto any device that supports PowerPoint including PCs, laptops, tablets and smartphones. The program specifically targets for training the six skills that the literature suggests are most likely to lead to crash reductions in the three riskiest crash types. To be clear, these six skills are strategic hazard anticipation, tactical hazard anticipation, strategic attention maintenance, tactical attention mitigation, strategic attention maintenance, and tactical attention maintenance. An example can make most clear what we mean by the different skills. Consider what we call the truck midblock crosswalk scenario. A plan view (bird’s eye or top down view) of this scenario is given in Figure 1. Imagine you are in a car driving toward the crosswalk starting at the red arrow. There are cars (light blue) parked on both sides of the road with one travel lane and one parking lane in each direction. There is a SUV (dark blue) parked right in front of the marked, midblock crosswalk. A pedestrian sign is located both upstream of the crosswalk and immediately at the crosswalk. A pedestrian could potentially emerge from behind the light blue SUV parked in front of the crosswalk and will be referred to as the latent threat. The threat is a latent one because we do not know whether it will or will not emerge. Using this scenario as an example, the six skills for this scenario are: 1) Strategic hazard anticipation: upstream of the crosswalk (at the location of the red arrow), observing the pedestrian sign and recognizing that this means there may be latent hazards ahead (possibly pedestrians in a crosswalk who are initially obscured from view); 2) Tactical hazard anticipation: on immediate approach to the crosswalk (within one to two seconds of traveling over the crosswalk), recognizing that a hazard could emerge from the area obscured by the parked SUV, and looking toward the area from where the pedestrian could emerge; 3) Strategic hazard mitigation: upstream of the crosswalk, having recognized that there may be hidden pedestrians ahead (the outcome of a successful strategic anticipation of the hazard), searching for an area downstream where the latent hazards are likely to materialize (e.g., a crosswalk); 4) Tactical hazard mitigation: on immediate approach to the crosswalk, maneuvering the vehicle to avoid as best as possible a latent hazard that might emerge (e.g., slowing down and moving slightly to the left when passing the parked SUV and reaching the crosswalk); 5) Strategic attention maintenance: upstream of and on immediate approach to the crosswalk, not looking away from the forward roadway when a latent hazard is anticipated ahead; 6) Tactical attention maintenance: at any point other than near a crosswalk (or other latent hazard), not glancing away from the forward roadway for more than two seconds. It is important that ACCEL target not only the skills that are most in need of development, but also the crash types in which teens are more likely to be at risk. With this in mind, the six skills were trained using the scenarios from the three most risky crash types: intersection, rear-end and run-off-road crashes. Six different intersection, rear-end and run-off-road scenarios were constructed for 18 unique scenarios. We trained all six skills within a given crash type before moving on to the next crash type. The skills were always trained in the order listed above. After identifying the skills one needs to train and the scenarios in which those skills should be trained, the next question one needs to address in the design of a training program is what type of training method should be used. We used an active method (3M — mistakes, mentoring, mastery), which Romoser and Fisher (2009) found to be more successful than a passive method. The active training program included three modules for each combination of skill and scenario: a mistakes module (putting drivers in an unfamiliar setting where they can make errors), a mentoring module (providing the drivers with immediate feedback and explaining how to avoid such errors in the future), and a mastery module (allowing drivers to correct their mistakes). This is often referred to as error training. First Training Session: ACCEL and Placebo: Fifty participants, all novice drivers between the ages of 16 and 18 with fewer than six months of solo driving experience were brought to the lab and exposed to ACCEL. ACCEL took about two hours to complete. Another 25 novice drivers in the same age range were exposed to a placebo program. The placebo training consisted of videos explaining the importance of vehicle maintenance and how drivers should properly maintain their vehicle (e.g., checking tire pressure, etc.). This PC-based program took about 70 minutes to complete. First Simulator Evaluation: Is ACCEL Effective in the Short Term?: One hundred participants were evaluated on the driving simulator in the initial evaluation. In particular, immediately after training, eye movements were recorded and vehicle measures were collected from the above total of 75 novice drivers (16 to 18 years old with less than 6 months’ experience), of which 50 were ACCEL-trained and 25 were placebo-trained, and a total of 25 experienced drivers (28 to 55 years old with at least 10 years’ experience), all untrained. ACCEL training was found to improve the performance of novice drivers in six out of the six of the trained skills when compared to placebo-trained teens. The improvement was significant in five of the six skills: tactical and strategic hazard anticipation, strategic hazard mitigation, and tactical and strategic attention maintenance. It was marginally significant in the last of the six skills: tactical hazard mitigation. Importantly, in this regard (tactical hazard mitigation) the difference in the speeds of the ACCEL-trained novice drivers and the experienced drivers did not differ from one another in the area of the latent hazard (in fact, the ACCEL-trained novice drivers were actually traveling slower than the experienced drivers). The results are consistent with the hypothesis that combined skill training can be delivered effectively in a relatively short amount of time. With respect to gender, the performance of ACCEL-trained female drivers was better than the performance of the placebo-trained female drivers. Moreover, the effect of training did not vary across genders. These finding are of interest because they suggest not only that ACCEL could have an effect on the crashes of female novice teen drivers, something that was not found in the California study, but that this effect could be the same order of magnitude for both females and males. Second Training Session: ACCEL: All 75 novice drivers were asked to return for a second evaluation after a period of three to six months. Twenty-six eventually did, 10 placebo-trained and 16 ACCEL-trained. Of the 16 ACCEL-trained drivers, half were trained a second time, either using their own PC at home or using the PC in the lab. Thus, the ACCEL group was now divided into two groups, those who received training only once (ACCEL-1) and those who received training twice (ACCEL- 2). Second Simulator Evaluation: Is ACCEL Effective in the Longer Term: Twenty-five participants were evaluated on the driving simulator a second time between three and six months after the initial training. The eye data from one of the placebo-trained participants was corrupted and so that participant could not be used in the analysis. We asked three primary questions. First, we wanted to know whether the effects of training endured over time. In five of the six skills, the ACCEL-1 group did better than the placebo group. The one skill in which the ACCEL-1 group fared more poorly is tactical attention maintenance. None of these differences were significant, presumably because of the small sample size. If a second training session is added, the participants in the ACCEL-2 group did better than the placebo group in all six skills. Three of the six differences between the ACCEL-2 group and the placebo group were significant. Second, we wanted to know whether a second training session enhanced the performance of participants who had been so exposed compared to those participants who had been exposed to ACCEL only once. In four of the five behaviors that were indexed by glances, the ACCEL-2 group performed better than the ACCEL-1 group: strategic hazard anticipation, tactical hazard anticipation, strategic hazard mitigation and tactical attention maintenance. The one glance indexed skill in which the ACCEL-1 group performed better than the ACCEL-2 group was strategic attention maintenance. In the skill measured by vehicle behavior, tactical hazard mitigation, the participants in the ACCEL-2 group drove on average faster than the participants in the ACCEL-1 group before the latent hazard, in the immediate vicinity of the latent hazard, and after the latent hazard. We are not sure why this is the case. Third, we wanted to know whether the training would be as effective for the female novice drivers as for the male novice drivers. There were not enough data to answer this question. However, we could still look at whether trained female drivers performed better than untrained female drivers. Among female drivers in the ACCEL-2 group, they performed better than female drivers in the placebo group did on five of the six skills (exclusive of tactical attention maintenance). Among female drivers in the ACCEL-1 group, they performed better than female drivers in all six of the skills did. In summary, as noted at the outset, seven questions were addressed in this study. First, we asked whether an omnibus training program could be developed that was targeted for training all six skills across all three types of crash scenarios. The answer is yes. Second, we asked whether the training program could be developed using open source software that was easily downloadable from the internet to a wide variety of devices, that could readily be viewed on these devices, and that was freely available. The answer is yes. Third, we asked whether the training program could be administered in one two-hour session without reducing the effectiveness of the training of the individual skills, which if taught one at a time using current programs targeting only single skills would take at least 4.5 hours. The answer is yes. Fourth, we asked whether the effect of training would be present across all skills and crash types immediately after training. The answer is yes. Fifth, we asked whether the effect of training would persist three to six months after the first administration. The answer is yes. Sixth, we asked whether an additional training session could improve the skills even more. The answer is yes. Finally, we asked whether the training would affect males and females differently immediately after training and three to six months later. The answer is no immediately after training. However, attrition was too high to answer the question definitively six months after training. (Author/publisher)
Samenvatting