Emanuel Robinson

Threshold Effects of Speed-Monitoring Devices on the Speeding Behavior of Drivers

This paper summarizes an evaluation of a speed-monitoring system that provides speed warning feedback to drivers enrolled in a voluntary program. The field study aimed to determine the effects of immediate feedback on drivers, especially chronic speeders. Drivers with at least three speeding violations in the past 3 years were recruited through the Maryland Motor Vehicle Administration. After a 2-week baseline period, the alert system was activated, and driving behavior was monitored for any changes for approximately 4 weeks. Subjects were monitored (silently) for a follow-up phase of 2 weeks. During the treatment phase, drivers received alerts when their speeds exceeded the posted speed limit by more than 8 mph. The findings are encouraging and suggest that verbal alerts are successful in producing short-term changes in driving behavior. Overall, the average proportion of speeding above the alert threshold declined significantly during the treatment phase, an indication that the alerts did have a deterring effect on speeding behavior. Once the alerts were silenced, there was evidence suggesting a sustained change in driving behavior for some participants. Although the proportion of speeding above the threshold was higher during the 2-week follow-up period than during the treatment phase, for some participants the follow-up speeds were lower than those recorded during the baseline phase. Although speeding was reduced during the treatment phase at speeds over the feedback threshold, much of this speeding appeared to have shifted down to just below the threshold. This finding raises an important question regarding optimal levels for setting thresholds for feedback on speeding behavior.

Effects of Multiple-Device Integration Strategies on Driver Comprehension of Crash Warnings

This experiment investigated the extent to which driver comprehension of imminent crash warnings is affected by the degree of integration when there are multiple Connected Vehicle products in the vehicle. Two displays were placed in an experimental vehicle: one representing original equipment and one representing a portable, nomadic device. Participants were randomly assigned to one of five device integration conditions. Participants drove on a test track and were occasionally presented with either a nonurgent message or an urgent crash warning. Participants recognized warnings most quickly when only one display was active in the car; when both displays were active, response times generally improved when messages and warnings were integrated into a single physical location. Warning recognition times were longer when a warning closely followed a non-urgent message on the other display than when a warning followed a non-urgent message on the same display. When messages and warnings from both source displays were integrated into a single display, this effect was not observed, suggesting that the separate display locations are responsible for the increased warning recognition time rather than the different formats of the messages from each source device.