The aim of this study was to test relationships between self-reported driving anger, how people express their anger while driving, traffic violations and accidents in a sample of 229 young drivers. This study administered the French adaptation of the Driving Anger Scale, the French adaptation of the extended version of the violations scale from the Driver Behavior Questionnaire (Delhomme & Villieux, 2005), and our translation of the Driving Anger Expression Inventory : DAX (Deffenbacher, Lynch, Oetting, & Swaim, 2002). Factor analysis of the Driving Anger Expression Inventory items yielded a three-factor solution, with 11 items contrary to the 49 items and four subscales found within the original US sample. The « Personal physical aggressive expression » factor was abandoned because our French sample reported that they were not sufficiently able to express their anger in the manner described by this factor. The « Verbal aggressive expression » factor was relabelled as « Nonverbal aggressive expression ». Confirmatory factor analysis showed that the three-factor solution obtains a better goodness-of-fit to the data. Cronbach α reliabilities for DAX factors ranged from .64 to .83. The two aggressive forms of expression ( « Use the vehicle to express anger » and « Nonverbal aggressive expression » ) correlated positively with each other (r = .36), but were correlated negatively with the « Adaptive/constructive expression » factor (from – .28 to .32). Aggressive forms of anger expression correlated positively with driving anger and traffic violations, whereas the « Adaptive/constructive expression » factor correlated negatively with these variables. Compared to low-anger drivers, high-anger drivers reported being engaged in more traffic violations and accidents. These drivers reported that they expressed their anger more frequently in the manner described in the aggressive forms. On the other hand, low- anger drivers reported more frequently expressing their anger in adaptive and constructive ways. Globally, our results replicated earlier findings obtained in the American context by Deffenbacher et al. (2002) and showed that factors of the DAX. French translation are useful predictors of self-reported violations and complement established measures, such as the DAS. Implications for driver education and interventions were examined.