Research in the field of social representations and of norm conditionality has enabled the construction of specific tools such as the conditional script questionnaire (CSQ) and the use of various methods of analysis. The first aim of this study is to show the differences of conditionality between male and female drivers toward pedestrians. To test this hypothesis, a version of the CSQ has been fitted to the pedestrian. We show that conditionality toward the pedestrian is more significant for men in certain specific situations and we highlight that women are more aware of the vulnerability of pedestrians. The second stage is aimed at improving knowledge between conditionality and risk taking, by using Bayes’ theorem for the first time. It is demonstrated that a Bayesian network based on the CSQ can be built in order to model the perception of hazardous situations.