ABSTRACT
The improvement of the validity of the driving test can be seen as a possibility to reduce the novice drivers` high accident risk. One empirical indicator of test validity is a clear difference between expert and novice test performance. Expertise research in general has revealed that experts outperform novices more clearly with increasing domain specificity of the given task–it's similarity to the characteristic demands of a defined domain–(Glaser & Chi, 1988), and that expert knowledge is better connected and thus more flexible than that of novices in a certain domain (Spiro, Feltovich, Jacobson, & Coulson, 1991). Expertise research in the domain of road safety tries to identify quantifiable aspects of driving skill that can explain the increased accident liability of novice drivers (Horswill & McKenna, 2004). Hazard perception which is the driver's ability to identify emerging hazards and potential dangerous situations has been identified as one such relevant skill (McKenna & Crick, 1994). The aim of the present study was to investigate whether the domain specificity of hazard perception tasks can be enhanced by the implementation of animated instead of static pictures of traffic scenarios and whether the ambiguity of the presented hazards has any impact on this effect.
According to expertise research we assumed, that experts outperform novices more clearly if dynamic presentations are involved, because they are more specific 430for the driving domain than static pictures. This effect could be even more obvious if the given tasks require the use of driving specific knowledge if rather implicit than explicit hazards are presented.
In an experimental study 63 learner drivers (novices) and 70 individuals with some driving experience (experts) were to react on potential hazards within 31 different driving scenarios presented on a computer screen. Whereas previous studies into hazard perception assessment included either videos (e.g. Sagberg & Bjornskau, 2006) or static pictures (e.g. Biermann et al., 2008) of traffic situations, we compared animations and pictures of exactly the same traffic scenarios. We used a 2 by 2 by 2 factorial mixed design with repeated measures on the factor hazard ambiguity (explicit vs. implicit hazard) and the two between subjects factors expertise (experts vs. novices) and presentation mode (animated vs. static).
The results indicated that experts in general identified more breaking hints (ηp 2 = .20) and by tendency were faster in reacting than novices (ηp 2 = .03). This confirms that the reaction task represents a valid measurement of driving expertise. There was no simple interaction between the factors expertise and presentation mode but a three way interaction (ηp 2 = .08). The interaction indicated that animation does not generally improve the assessment of hazard perception but for those scenarios in which the use of knowledge about traffic rules is required (implicit hazards). This reveals that the decision to choose animations in order to assess hazard perception should depend on content aspects of a task. This allows the conclusion that the most specific and valid reaction time task for the driving domain includes the animated presentation of traffic scenarios with implicit hazards.
