Abstract:
This research aimed 1) to develop a hazard perception test for use while driving in a Thai traffic context; 2) to compare the difference of percentage correction and response times as related to experience and hazard types; 3) to compare the difference of eye movement in experience and hazard types; 4) to examine the correlations among response time, percent correction, and eye movement. The research involved a sample of Thai drivers, males and females, aged 20-45 years old, divided into three groups of 25 people each who were either non-experienced, novice, or experienced. The research instruments included video clips categorized by three hazard types: behavior prediction hazard, environment prediction hazard, and a dividing and focusing attention hazard. The EyeLink 1000 Plus, SR Research Experiment Builder, and EyeLink Data Viewer program were used to create a test program and to record data. Data were analyzed by using descriptive statistics ANOVA, Two-way ANOVA, Two-way MANOVA, and Pearson correlation coefficients. The results revealed that 1) the hazard perception test in Thai traffic context involving a total of 30 videos clips was both reliable and valid. 2) Experienced drivers had a higher percent correct (p<.01, np 2 = .28) and lower response time (p<.01, np 2 = .37) than both the novice and non-experienced drivers. It was found that the Environment prediction hazard, when compared to the other two hazards, had more percent correct (p<.01, np 2 = .41) and faster response time (p<.01, np 2 = .35). 3) Eye movement variables (first fixation time, dwell time, fixation count, first saccade amplitude and first saccade angle) did not differ by driving experience, but did differ by hazard type (p<.01, np 2 = .69). 4) Eye movement had a low correlation with either the percent correction or the response time. These findings may help to screen driver hazard perception, and serve as a guideline for driving safety training.