ABSTRACT
The study consisted of two samples of young drivers who completed a questionnaire, containing questions on driving while being in an improper state, driving with high speed, violations and accidents. McKenna defined differential accident involvement as the issue of whether it is possible to predict crash involvement from psychological tests, whereas accident proneness was seen as statistical modelling, that is, distribution fitting. The use of several safety outcome variables in parallel is fairly common in traffic safety research. However, it should be stressed again that there can really only be one safety, and that the effect sizes for different dependent variables should therefore be in the same relative association with the predictors. Even fairly simple logical/mathematical errors occur in traffic safety research, when the dependent variable is to be calculated, even when the necessary information for countering them is present, as is evident when the authors themselves have included it in their report.
