Combining information processing, personality, and demographic variables to predict motor vehicle crash involvement

Abstract

The present study examined whether an empirical combination of three commonly but independently used predictor categories of crash involvement would result in incremental prediction over and above that provided by individual variables or single predictor categories. A total of 339 participants completed demographic, information processing, and personality measures and a self-report measure of motor vehicle crash involvement. The results of a series of hierarchical linear and logistic regression analyses suggested that, in contrast to expectations based on the existing motor vehicle crash literature, the specified empirical combinations of predictor categories tested in the present study contributed little in terms of improving prediction of crash involvement. Specific personality variables accounted for the majority of explained variance in the criteria, with the demographic and information processing categories contributing little to overall prediction. Consistent with prior research (Arthur & Doverspike, in press; Arthur & Graziano, 1996), conscientiousness was negatively and significantly associated with both at-fault and total crashes. Limitations of the present study and suggestions for future research are discussed.