Comparison of Trained and Untrained Novice Drivers’ Gaze Behavior in Risky and Non-Risky Scenarios

PC-based training programs have been developed that have been shown to improve novice drivers’ hazard anticipation skills. Such programs give novice drivers information about particular driving situations (scenarios) where hidden threats could appear. We wanted to know whether this improvement in trained novice drivers’ scanning skills was simply because the trained drivers were scanning more in general or, instead, were scanning more specifically in the scenarios in which potential threats could appear. In order to evaluate this question, we trained 11 novice drivers using a PC-based program and then compared their hazard anticipation performance on a driving simulator with the hazard anticipation performance of 11 untrained novice drivers. The drivers’ eye movements were recorded for the duration of the drives. The glances of the drivers to the right (the correct response in most of the risky scenarios) were analyzed for each of the relevant risky scenarios and for stretches of non-risky situations. The trained drivers did look to the right 6.5% more in the non-risky situations than did the untrained drivers, although the difference was far from significant. However, the trained drivers looked to the right 32.7% more in the risky scenarios than in the non-risky situations, indicating they were discriminating quite well between the two situations. The untrained drivers also showed a smaller, but significant, discrimination between the risky scenarios and non-risky situations, as they looked to the right 18.9% more in the risky scenarios than in the non-risky stretches