An analysis of fatal crash data up to 2013 showed the fatal crash rate per mile driven is nearly twice as high for 16- and 17-year-olds as it is for 18- and 19-year-olds (Insurance Institute for Highway Safety, n.d.).This suggests the presence of one or more deficiencies during the first months of licensure that lead to an increased crash risk for newly licensed teen drivers. Analyses of police crash reports by McKnight and McKnight (2003) showed that the main failures leading to crashes among drivers 16 to 19 years old included inadequate visual scanning (ahead, to the sides and to the rear–43.6%), poor attention maintenance (23.0%), and inappropriate speed management (20.8%). Research has also shown that this group has poor anticipation of the existence and position of unexpected hazards (Pradhan et al., 2005) and does not control the speed, acceleration, and position of their vehicles well for hazard avoidance (Fisher et al., 2002; Sagberg & Bjørnskau, 2006). One way to respond to these teen crash problems is to train new teen drivers to counter the deficiencies in their driving skills. One such training program that addresses hazard anticipation and poor scanning issues is the Risk Awareness and Perception Training (RAPT) program (e.g., Fisher et al., 2010). Researchers have conducted multiple studies of the impacts of various versions of the RAPT program approach using novice drivers in simulators (Pollatsek, Narayanaan, Pradhan, & Fisher, 2006) and in the field (Pradhan, Pollatsek, Knodler, & Fisher, 2009). These studies found that RAPT increased the likelihood that newly licensed drivers would anticipate hazards correctly as measured by the position of their gaze as they drove. The current research effort had two phases. Phase 1 objectives included: • Updating the packaging and delivery of the RAPT training so the resulting RAPT product could be confidently delivered to a wide audience. • Training a large sample of newly licensed teen drivers using the updated RAPT program and collecting information from a similar group of newly licensed teens who did not receive the training. Phase 2 objectives included: • Constructing a database that contained participant demographics, RAPT training data, and one year of post-training driver record data (crashes and traffic violations). • Analyzing the data to answer the following research questions: o During the first year of driving, did drivers who completed the RAPT program have fewer crashes or traffic violations on their driving records than the comparison group? o Is there a differential effect of RAPT training by sex as measured by number of crashes; types of crashes; and violations? o Does individual performance on the pre-test and post-test included in the RAPT program correlate with crashes and citations? Updating the program. In order to enhance the flow and appearance of the training and increase the stability of its operation, the entire training package was re-programmed and streamlined with higher quality graphics. Site Selection. The researchers selected California as the site for this research because the California Department of Motor Vehicles (DMV) had an excellent and interested research department and an extensive number of licensing offices in which RAPT could be given to newly licensed young drivers. DMV selected six offices, three in the Los Angeles (“South”) area and three in the San Francisco Bay (“North”) area, with a high-flow of young drivers applying for first-time driver licenses. Participants. Participants included 5,251 drivers 16 to 18 who had just passed their onroad driving exams for provisional or unrestricted license (first licenses) at one of the six selected California DMV field offices. Training Procedure. A DMV technician acted as proctor and asked the participant to sit at a computer and read the first screen that explained the study. Once the participant consented and typed in his/her driver’s license number, the RAPT or comparison (pre-test only–no training) program began. Proctors ran either the RAPT or comparison program on alternate weeks such that all participants in an office for a given calendar week completed either the RAPT or comparison program. Data. Driver license number information and scores from measures generated within the RAPT and comparison programs were obtained from the desktop computers for all participants. License and RAPT score information from all offices were combined into a single file that was then merged with crash and violation data obtained from DMV’s Driver Record Master (DRM) files for a period of 12 months post-licensure. Crash data were obtained without knowledge of participant treatment group assignment. Property damage only (PDO) and injury crashes contained in the DRM were obtained from the California Highway Patrol (CHP) and through self-reports to the State (i.e., by a crash-involved driver and/or an insurance company). There were no fatal crashes and too few injury crashes to analyze independently. The DRM provided only a total count of reportable violations and, therefore, researchers could not investigate individual violation types. Analysis. The project involved several analyses including tests of group equivalency, analysis of RAPT pre-test/post-test scores, crash analyses (frequency and time to first crash), and analyses of traffic violations. Analyses focused on differences between the RAPT and comparison groups and by other covariates of interest. The study achieved its process objectives, and the participant assignment process employed produced a dataset without meaningful bias between those who did and did not receive RAPT training. This dataset supported comparative analyses of the effects of RAPT on the crash and violation records of trained and similar untrained drivers during the initial year of their unsupervised driving careers. The crash results indicated that RAPT training was associated with a statistically significant crash rate reduction of 23.7% for males. Interestingly, females exposed to RAPT showed a higher crash rate relative to females who did not complete the training, but this difference was not statistically significant. A differential effect on the sexes is not inconsistent with previous research focused on the teen driver. Although the precise causes of the differential response to RAPT by sex are currently unclear, factors dealing with both traffic safety and computer-based learning could have played a role. The analyses of time to first crash showed no effect of RAPT on time to first crash. Overall, the results of this study provide perhaps the first encouraging evidence that brief, computer-based training interventions can have a positive influence on driving safety for newly licensed teen drivers even if only for males. Given the size of the sample included in this study, the single State venue, the fact that the RAPT implementation studied used only a limited number of selected scenarios, and the experimental limitations of the study, the reader must exercise caution when generalizing these findings. Nevertheless, the potential importance of the existence of some positive crash-based results for guiding future research and development with respect to the driver training process cannot be overlooked. Further research is needed to clarify the uncertainties arising from this study, particularly related to the lack of effectiveness, and even potential detrimental effects, of RAPT on female crash rates. Once a better understanding of the effects of RAPT is achieved, it would be useful to assess how best to employ hazard perception training using a program such as RAPT in the driver training process. (Author/publisher)
Samenvatting