Common Method Variance and other sources of bias in road traffic research i
KeywordsRoad crashes, traffic offences, common method variance, social desirability, method bias, meta-analysis, self-report data, road safety.
AbstractReducing the number of fatalities and injuries due to vehicle crashes is a major focus for road safety researchers throughout the world. Research on driver behaviours, attitudes and associated crashes is often conducted using self-report surveys. When the predictor and criterion variables are drawn from a single source (such as a cross-sectional survey) as is generally the case, it has been generally acknowledged that a range of methodological issues associated with this approach have the potential to reduce the efficacy of predictive measures. Common method variance (CMV) is bias that occurs when both the predictors and the predicted variables are gathered and analysed in the same manner from the same source. CMV effects have the potential to distort findings by artificially strengthening or weakening the observed relationships between items of interest. This is of particular concern to road safety researchers, given the distinct possibility that the small associations often observed within road safety studies may in turn be inflated due to CMV biases. There is a significant need to examine the exact nature and extent of CMV effects within self-reported road safety data and to explore how best to reduce this measurement error.The current body of research was conducted with three primary aims in mind:1. Examine the existence of CMV effects.2. Explore the level of influence that CMV effects have on self-report data, including whether these effects inflate or deflate observed relationships between variable.3. Test research methods to control for the influence of CMV effects on selfreport data.Study 1 involved a meta-analysis that quantified the association between crashes and traffic offences in light of potential moderators. In reference to the exploration for CMV effects, this study examined the extent to which the relationship between self-reported traffic offences and self-reported crashes is stronger than that recorded using only official data sources. Studies yielding a correlation between crashes and traffic offences were collated and a meta-analysis of 144 effects drawn from 99 road safety studies conducted. The impact of factors such as age, time period, crash and offence rates, crash severity and data type, sourced from either self-report surveys or archival records, were examined. After weighting for sample size, an average correlation of r = .18 between crashes and traffic offences was observed over the mean time period of 3.2 years. Evidence emerged suggesting the strength of this correlation is decreasing over time. Stronger correlations between crashes and offences were generally found in studies involving younger drivers. In relation to the existence of CMV effects, a within-country analysis found stronger effect sizes in self-reported data after controlling for crash mean ...