Hazard perception assessment - How much ecological validity is necessary?

Hazard perception - the driver's ability to identify road hazards - is one of the fewmeasurable aspects of driving competency that can explain the high accident risk of some groups of drivers[1, 2]. Hazard perception tests differ regarding several features. Only a few empirical studies have been conducted to systematically compare different design features of hazard perception tests as e.g. the presentation mode (static vs. dynamic) of the included traffic scenarios[3, 4]. Beside the variation of the stimulus material, hazard perception tests differ regarding the kind of task that has to be fulfilled by the participants. In general, experts outperform novices more clearly with increasing ecological validity ofa task - its similarity to the demands of a certain domain [5]. As an empirical indicator of validity is a clear difference between experts and novices, ecological valid tasks are expected to increases the quality of a test. The aim of the present study was to investigate whether the criterion validity of a hazard perception test can be enhanced by the application of an ecological valid task. 35 learner drivers and 31 experienced drivers were presented 21 animated driving scenarios that did or did not contain a potential hazard. One half of the participants were to react to hazards within the animation (high ecological validity) whereas the others worked on multiple-choice questions provided after the presentation (low ecological validity). The ANOVA revealed a main effect of expertise: experts scored highest. Moreover, a main effect of the task's ecological validity could be found: the multiple-choice task facilitated the test. No interaction between expertise and ecological validity could be found. The results indicate that both taskslead to valid measures. Contrary to our expectations, the ecological validity of the task did not influence the criterion validity. Further research should be conducted e.g. in order to investigate, whether a further increase of ecological validity - e.g. by simulated driving - results in extended expertnovice differences. (C) 2015 The Authors. Published by Elsevier B.V.