Measuring errors and violations on the road: a bifactor modeling approach to the Driver Behavior Questionnaire.

Area of Science:

• Psychology
• Transportation Safety
• Road User Behavior

Background:

• The Driver Behavior Questionnaire (DBQ) is a validated self-report measure for assessing driving behaviors.
• Existing research highlights the need to clarify the relationships between DBQ sub-components (violations, errors) and crash involvement.
• The efficacy of a condensed DBQ version requires investigation.

Purpose of the Study:

• To examine the factor structure of the Driver Behavior Questionnaire (DBQ) using a bifactor model.
• To determine the predictive validity of DBQ components and a general factor for crash involvement in novice drivers.
• To assess the effectiveness of a reduced-item DBQ.

Main Methods:

• Utilized a bifactor modeling approach on data from the UK Cohort II longitudinal study of 12,012 novice drivers.
• Collected DBQ data at multiple time points (0.5, 1, 2, and 3 years post-test).
• Compared bifactor models with correlated and second-order factor structures; tested a 12-item reduced model.

Main Results:

• A bifactor model with a general factor and specific factors (ordinary violations, aggressive violations, slips, errors) provided the best fit.
• A 12-item DBQ model successfully replicated the full model's structure and yielded highly correlated factor scores.
• The general factor and ordinary violations independently predicted crash involvement at 6 months post-independent driving.

Conclusions:

• The bifactor model effectively captures the structure of the Driver Behavior Questionnaire (DBQ).
• A condensed 12-item DBQ is a viable and effective tool for assessing driving behavior.
• Both general driving behavior tendencies and specific ordinary violations are crucial predictors of novice driver crash risk.