The hazard prediction test: A comparison of free-response and multiple-choice formats

Abstract

Hazard perception skill is often related to lower crash risk, and the hazard perception test has been widely employed to measure this ability in drivers. An increasingly popular test-variant is the hazard prediction test: driving videos are occluded immediately prior to a hazard and participants are asked to predict how the situation will develop. Early versions of this test asked participants to provide a free-response answer which was subsequently coded. Later versions, however, have used a multiple-choice format where participants are provided with four options presented on screen. While the benefits of a multiple-choice format are obvious in terms of providing immediate feedback without relying on subjective coding, it is unclear whether this change in format affects the discriminative validity of the test. For the current study, a free-response test and a multiple-choice test were created using the same video clips. The free-response test (experiment 1) was found to successfully discriminate between novice and experienced drivers, with the latter predicting more hazards correctly. The answers provided by participants in Experiment 1 were then used to generate the options for a multiple-choice test (experiment 2). This second test was also found to discriminate between novice and experienced drivers, and a comparison between the two tests failed to reveal an advantage for one over the other. Despite this, correlations between prediction accuracy and both years of post-license driving, and annual mileage, were only significant for the multiple-choice test. The results suggest that the multiple-choice format is not only time- and cost-efficient, but is ostensibly as good as the free-response test in discriminating between driver groups.