A quarter of a century of the DBQ: some supplementary notes on its validity with regard to accidents

Abstract

This article synthesises the latest information on the relationship between the Driver Behaviour Questionnaire (DBQ) and accidents. We show by means of computer simulation that correlations with accidents are necessarily small because accidents are rare events. An updated meta-analysis on the zero-order correlations between the DBQ and self-reported accidents yielded an overall r of .13 (fixed-effect and random-effects models) for violations (57,480 participants; 67 samples) and .09 (fixed-effect and random-effects models) for errors (66,028 participants; 56 samples). An analysis of a previously published DBQ dataset (975 participants) showed that by aggregating across four measurement occasions, the correlation coefficient with self-reported accidents increased from .14 to .24 for violations and from .11 to .19 for errors. Our meta-analysis also showed that DBQ violations (r = .24; 6353 participants; 20 samples) but not DBQ errors (r =  − .08; 1086 participants; 16 samples) correlated with recorded vehicle speed.

Practitioner Summary: The DBQ is probably the most widely used self-report questionnaire in driver behaviour research. This study shows that DBQ violations and errors correlate moderately with self-reported traffic accidents.

Keywords:

• Driver Behaviour Questionnaire
• errors
• violations
• self-reported accidents
• crashes
• meta-analysis

Notes

^1. Özkan and Lajunen (2005a) found that the number of self-reported accidents across one's lifetime correlated positively with the respondent's age, which is self-evident because the longer you have been a driver, the greater the chance of ever having been involved in an accident. The paradoxical consequence of using lifetime accidents is that the most dangerous drivers (i.e. young drivers, who have a high violations score) have the least lifetime accidents. Hence, we concur with Af Wåhlberg (2003) that using accident counts across varying time periods is ‘possibly not at all in agreement with what the researchers using this method have intended’ (481).

^2. A fixed-effect model assumes that all studies estimate the same ‘true’ effect size. Accordingly, the effect sizes of the individual studies are weighted by their inverse variances. A random-effects model assumes that the true effect size differs between studies and aims to estimate the average of the distribution of these effects. As a result, the weights assigned to the individual studies by a random-effects model are more homogeneous than the weights assigned by a fixed-effect model. Furthermore, the 95% confidence interval of the summary effect is broader in a random-effects model than in a fixed-effect model.