In the United States, obtaining a driver’s license is widely considered a teenage rite of passage. The first step to this milestone usually consists of obtaining a learner’s permit, which allows a young driver to operate a motor vehicle with adult supervision. Data show that crash rates during this period of adult-supervised practice are low, then increase about tenfold when young drivers begin to drive independently (Mayhew, Simpson, & Pak, 2003). Several in-vehicle technologies aim to help parents continue to monitor and instruct their teen drivers after they are driving on their own. Evaluations of these technology-based interventions have reported varying levels of success in reducing risky behaviors associated with abrupt braking, steering, and acceleration manoeuvres as well as speeding and seatbelt nonuse. Simons-Morton, Zhang, Jackson & Albert (2012) report risky driving behaviors characterized by high gravitational-forces predict subsequent near-crash events. One intervention presented young drivers with video feedback about unsafe driving behaviors, including abrupt decelerations, accelerations, and steering manoeuvres. Parents and teens received weekly reports with descriptions of unsafe driving events and the teen’s event frequency compared to other participants, along with a CD containing the videos of triggered events. McGehee et al. evaluated this intervention with two cohorts of newly licensed drivers. The number of unsafe driving events per 1,000 miles driven (event rate) was calculated for preintervention, intervention, and post-intervention phases. The first study enrolled 25 teens from rural Iowa. Compared to the preintervention phase, the average event rate was reduced by 58% in the first two months of intervention and by 76% during the third and fourth months. The second study, involving 36 teens from suburban Minneapolis, produced similar results. Both studies found that the intervention was particularly effective among the initially riskiest drivers. Another evaluation employed a device that recorded sudden braking/acceleration, speeding, and seatbelt nonuse. This study enrolled 85 newly licensed teens in suburban Washington, DC. The protocol assigned drivers to one of four experimental conditions: (1) teens received immediate feedback and parents received prompt feedback online; (2) teens received immediate feedback, but were given a chance to correct behaviors before it was posted online; (3) teens did not receive immediate feedback but parents received prompt feedback online; and (4) a control group that received no feedback. Teens in the study showed significant reductions in instances of speeding more than 10 mph over the limit only in condition 2, where the device gave teens an in-vehicle alert that parents would receive an e-mail report about the speeding incident, but teens believed they could cancel the report by correcting their behaviors. Evaluations of another non-video device were conducted in Israel and in the United Kingdom. The device employed proprietary pattern recognition algorithms to identify risky driving manoeuvres including speeding. When the system identified risky driving, a panel of three LEDs (green, yellow, and red) displayed immediate feedback to the driver, and families were provided with online access to view each driver’s risk index. In Israel, families accessed online information 68.6% of the time. Modelling analyses revealed reduced risk among participants whose parents checked the online information (Prato et al., 2010). In the United Kingdom, overall event frequency (risky manoeuvres per minutes of driving) decreased by 52%. The results of an analysis controlling for driving experience indicated the reduction in event frequency could be attributed to the availability of feedback. These studies provide evidence that both video and non-video feedback can reduce the frequency of young drivers’ unsafe behaviors. Though teens and parents have noted privacy concerns and deterioration of trust as the primary deterrents to installing monitoring devices, parents reported being less likely to consider installing a device that uses video. Interestingly, the majority of teen drivers in the two studies conducted by McGehee et al. reported they did not consider the event recorder an invasion of privacy. Research is needed to determine if the use of video is necessary to achieve the behaviors changes seen in previous evaluation. Achieving comparable levels of behaviors change without the use of video has clear cost implications for parents and designers of similar systems. The current study compared effects of similar interventions, one with video-based feedback and one without, on teens’ driving behaviors. This study collected and analysed data from two cohorts of newly licensed drivers, one rural and one suburban, to address the following research questions: 1. To what extent do two technology-based interventions, one including video feedback and one with non-video feedback, reduce unsafe driving behaviors of newly licensed teen drivers relative to (a) a baseline period, and (b) a control group? 2. Does including video with the intervention produce a significantly different effect than a similar intervention without the video? The final dataset included data from 68 teen drivers. About half the participants were from a rural site while the other half were from a suburban site. A video event recorder (ER) was installed in each teen’s personal vehicle for 20 weeks. Each teen was randomly assigned to one of three conditions: control, video feedback intervention, or non-video feedback intervention. During the 4-week baseline segment, data were collected for all teens, but none received feedback. During the 16-week intervention phase, the participants assigned to the intervention conditions received feedback. The ER continued to collect data for the control group, but provided no feedback. Technicians mounted a DriveCam video ER (model DC3) on the inside of the windshield behind the rearview mirror in each participant’s vehicle. This palm-size device had two video cameras, a three-axis accelerometer, a 12-second video data buffer, an infrared illuminator for lighting the vehicle’s interior at night, and a cellular transmitter. The device continuously buffered audio and video from both inside and outside the vehicle (see Figure 2), but only wrote data to internal memory when the vehicle motion exceeded an acceleration threshold, typically due to abrupt steering or braking. Based on the guidance of the manufacturer and the threshold values used in another naturalistic driving study, researchers set both lateral (side-to-side) and longitudinal (front-to-back) thresholds for this project at ±0.50g. Each video event captured the eight seconds before and four seconds after an event trigger. Event data and videos were encrypted and automatically uploaded daily to DriveCam’s fleet services server via a secure cellular connection. The research team viewed the event videos using password-protected DriveCam accounts online. In addition, custom software downloaded each event (.DCE) file and saved the event metadata in an SQL database on a dedicated computer at the University of Iowa. (Author/publisher)
Samenvatting