The aim of this paper is to investigate the feasibility of a four-factor solution of a Swedish version of the Driver Behaviour Questionnaire (DBQ-SWE). Previous studies on the British DBQ, as well as the DBQ-SWE, have used an exploratory factor analysis (EFA) technique. The present study is the first using confirmatory factor analysis (CFA). In EFA one explores data to tease out eventual links between the variables included and obtain a smaller number of factors. In contrast, in CFA a hypothesised model is imposed on the data in terms of limited parameters. The latter technique allows testing of how well the model fits data obtained on a single sample or data from several samples. Responses on the DBQ-SWE were obtained from more than 6,000 Swedish male and female drivers. Several individual samples were formed representing different age spans. The DBQ-SWE model was shown to be an approximate fit relative to four of the individual study samples. In addition, the model was an approximate fit relative to the gender of the respondents in these four samples. The model did not perform well with respect to data from relatively older drivers. The discussion includes methodological issues connected to DBQ-related investigations. (Author/publisher) For the covering abstract see ITRD E116881.
Samenvatting