On testing factorial invariance: A reply to J.C.F. de Winter

Since its publication, the Driver Behavior Questionnaire (DBQ) has been used for comparing subgroups of respondents on the constructs formed through factor analyzing the questionnaire items. However, not enough attention has been paid to ascertaining that the instrument actually measures the same constructs in the same way in all respondent groups. I recently published an article (Mattsson, 2012) that aimed to do this for the Finnish 28-item version of the DBQ using the stage-wise factorial invariance approach in the Exploratory Structural Equation Modeling (ESEM) context. de Winter (2013) commented on the publication, arguing that the results were artifacts due to measurement error that too many factors were extracted and that too strict criteria for invariance were applied. In this contribution, I reply to each criticism and suggest methodological approaches for ensuring the measurement invariance of self-report instruments such as the DBQ.     

驾驶员风险驾驶行为及影响因素分析

为探究中国驾驶员风险驾驶行为产生的原因及其影响因素,利用修正的曼彻斯特DBQ问卷对349名中国驾驶员进行了问卷调查。经过探索性因素分析(EFA)得到了4因子结构模型,分别命名为认知错误、违规行为、无意失误和记忆力流失,并利用验证性因素分析(CFA)对该模型进行了验证。研究了性别与驾驶行为的关系,结果表明男性驾驶员更容易发生违规行为,而女性驾驶员发生无意失误行为的频率较男性驾驶员偏高。通过变量间的相关性分析,研究了驾驶员的统计学信息、4因子以及交通事故之间的关系,构建了基于Logistic回归的交通事故预测模型,研究表明违规行为和年龄是影响交通事故的重要参数。     

Aberrant driving behaviors: A study of drivers in Beijing

The addition of massive numbers of new drivers with varied driving experience to roads in China suggests it is important to understand the nature of aberrant driving behaviors for this new set of drivers. A paper-based and an Internet survey were administered. Factor analysis produced a five-factor structure for each survey. The distinction between violations and errors indicated in previous studies was confirmed. The violations included emotional violations, risky violations and self-willed violations, and the errors included inexperience errors and distraction errors. In contrast to previous work, age was not found to be a good predictor of violations though driving experience was. Contrary to expectations, non-automotive (bicycle) roadway experience or level of driving training failed to predict poor driving behavior. On-road experience is the key to risk for China's drivers. Good agreement between the paper-based and Internet surveys indicate online surveys to be a feasible way to conduct research of driving behavior at low cost.     

Dimensions of driving anger and their relationships with aberrant driving

The purpose of this study was to investigate the relationship between driving anger and aberrant driving behaviours. An internet-based questionnaire survey was administered to a sample of Chinese drivers, with driving anger measured by a 14-item short Driving Anger Scale (DAS) and the aberrant driving behaviours measured by a 23-item Driver Behaviour Questionnaire (DBQ). The results of Confirmatory Factor Analysis demonstrated that the three-factor model (hostile gesture, arrival-blocking and safety-blocking) of the DAS fitted the driving anger data well. The Exploratory Factor Analysis on DBQ data differentiated four types of aberrant driving, viz. emotional violation, error, deliberate violation and maintaining progress violation. For the anger–aberration relation, it was found that only “arrival-blocking” anger was a significant positive predictor for all four types of aberrant driving behaviours. The “safety-blocking” anger revealed a negative impact on deliberate violations, a finding different from previously established positive anger–aberration relation. These results suggest that drivers with different patterns of driving anger would show different behavioural tendencies and as a result intervention strategies may be differentially effective for drivers of different profiles.     

The Manchester Driver Behaviour Questionnaire as a predictor of road traffic accidents

The Driver Behaviour Questionnaire (DBQ) has mainly been used as a predictor of self-reported road traffic accidents. The associations between crashes and the violation and error factors of the DBQ, however, may be spuriously high due to reporting bias. In the present study, the DBQ was tested as a predictor of self-reported and recorded accidents in four samples of private and professional drivers. The findings show that the DBQ scale only predicts self-reported accidents, not recorded crashes, despite the higher validity of company data and the higher means of the recorded data across these samples. The results can be explained by a common method variance bias. In a review of the DBQ research, the use of the instrument was found to be heterogeneous concerning the number of items, scales used and factor analytic methods applied. Thus, the DBQ may not be as homogeneous and as successful in predicting accidents as is often claimed.     

Thirty-day self-reported risky driving behaviors of ADHD and non-ADHD drivers

The present study aims to compare differences in reported risky driving behaviors of drivers – males and females – having and not having Attention Deficit Hyperactivity Disorder (ADHD), by using a checklist of driving behaviors based on the Driving Behavior Questionnaire (DBQ). Unlike the studies which employ the DBQ by asking the subjects to fill the questionnaire once, in this present study, the participants were asked to report their behaviors on a daily basis for 30 consequent days. The checklist included two factors of risky driving behavior: Violation and Faults. Thirty-eight drivers – 10 males and 9 females with ADHD, and 9 males and 10 females without ADHD (N-ADHD) as control groups – participated in the study. The results showed that the mean of the unsafe behaviors of ADHD was higher, i.e., less safe driving, compared to that of N-ADHD. However, a statistically significant effect was found only between male ADHD and male N-ADHD for the Faults. In order to check the effect of the length of the study, the 30 days duration of the research was divided into three consecutive periods. The reported driving habits of the female ADHD showed safer behaviors than those of the males. Unlike the findings of N-ADHD of both genders, which showed a tendency towards safer driving reports in the three periods, both genders of the ADHD showed higher rates of Faults, i.e., a decrease in safety driving reports, in the three periods. The findings suggest that ADHD drivers differ from the N-ADHD drivers in making driving mistakes, i.e., Faults, due to their lack of sustained attention, but not in making Violations. However, some of the results in the present study were not very strong. Possible explanations for this as well as methodological considerations are discussed, and further research is suggested.     

Some methodological deficiencies in studies on traffic accident predictors

The paper discusses some methodological problems in (psychological) research on traffic accident predictors and reviews a convenience sample of the literature. Three methodological aspects are identified as being important: reliability of accident predictors, time period for accidents used as dependent variable, and culpability for accidents. Papers are scrutinized and most are found to be wanting in these aspects. Traffic researchers do not adhere to, or hardly even discuss, these basic methodological problems. It is concluded that the current research into (psychological) accident predictors is fraught with methodological deficiencies. Why most studies seem to be deficient in these aspects is not clear, as several researchers have pointed out these problems.     

Validation of the driver behaviour questionnaire using behavioural data from an instrumented vehicle and high-fidelity driving simulator

Data from two previously published studies were used to examine the correlations between scores on the violation, error and lapse sub-scales of the driver behaviour questionnaire, and observed driving speed. One dataset utilised data from an instrumented vehicle, which recorded driver speed on bends on a rural road. The other utilised data from a driving simulator study. Generally in both datasets the DBQ violation subscale was associated with objectively-measured speed, while the error and lapse sub-scales were not. These findings are consistent with the idea that the DBQ is a valid measure of observed behaviour in real driving (its original intended use) and also in simulated driving. The fact that associations were the same in real and simulated driving lends further support to the relative validity of driving simulation. The need for larger and more focused studies examining the role of different motivations in different driving situations is discussed.     

Measurement invariance of the Driver Behavior Questionnaire across samples of young drivers from Finland and Ireland

This article investigates the factor structure of the 27-item Driver Behavior Questionnaire (DBQ) in two samples of young drivers (18–25 years of age); one from Finland and the other from Ireland. We compare the two-, three-, and four-factor solutions using Confirmatory Factor Analysis (CFA) and show that the four-factor model (with the latent variables rule violations, aggressive violations, slips and lapses) fits the data from the two countries best. Next, we compare the fit of this model across samples by the means of a measurement invariance analysis in the CFA framework. The analysis shows that the four-factor model fails to fit both samples equally well. This is mainly because the socially-oriented latent variables (rule violations and aggressive violations) are different in nature in the two samples. The cognitively-oriented latent variables (slips and lapses) are, however, similar across countries and the mean values of slips can be compared using latent variable models. However, the common practice of calculating sum scores to represent the four latent DBQ variables and comparing them across subgroups of respondents is unfounded, at least when comparing young respondents from Finland and Ireland.     

Investigating the factorial invariance of the 28-item DBQ across genders and age groups: An Exploratory Structural Equation Modeling Study

The Driver Behaviour Questionnaire (DBQ) is perhaps the most widely used questionnaire instrument in traffic psychology with 174 studies published by late 2010. The instrument was developed based on a plausible cognitive ergonomic theory (the Generic Error Modeling System, GEMS), but the factor structure obtained in the original study (Reason et al., 1990) did not mirror the theory's conceptual structure. This led to abandoning GEMS and adopting the obtained factor structure as a starting point for further DBQ research. This article argues that (1) certain choices in the original study, concerning statistical methodology and the wording of individual question items, may have contributed to the ways the obtained factor structure deviated from the underlying theory and (2) the analysis methods often used in DBQ studies, principal components (PC) analysis and maximum likelihood (ML) factor analysis, are not optimal choices for the non-normally distributed categorical data that is obtained using the instrument. This is because ML produces biased results when used with this type of data, while PC is by definition unable to uncover latent factors as it summarizes all variation in the measured variables. (3) Even though DBQ factor scores have been routinely compared in subgroups of men and women and respondents of different ages, DBQ's factorial invariance in these groups has not been rigorously tested. These concerns are addressed in this article by framing the results of certain previous DBQ studies as a structural equation model (SEM) and an Exploratory Structural Equation Model (ESEM) and testing measurement model fit in subgroups of respondents. The SEM analyses indicate that the model does not fit data from the whole sample of respondents as it stands, while the ESEM analyses show that a modification of the model does. However, the ESEM analyses indicate the DBQ measures different underlying latent variables in the different subgroups. Based on the analyses and a review of recent advances in attention and memory research, an update to the theory underlying the DBQ is suggested.