This study reports reliability and validity of the Dutch Reactive-Proactive Aggression Questionnaire (RPQ). In total, 845 participants completed the RPQ along with other measures of aggression. Groups consisted of non-offender participants, criminal offenders, youngsters (age 6–18), and adults (age above 18). Test–retest stability in a subsample of 324 childhood arrestees was good (all ICC's > 0.41). A confirmatory factor analysis supported the same two-factor structure as in the original RPQ. Convergent validity was adequate (all r's > 0.16). Moreover, results demonstrated that the proactive and reactive subscales were differentially related to measurements of callousness and impulsiveness, respectively. Criterion validity was shown in that non-offender subjects demonstrated significantly lower RPQ scores than offender samples. Finally, construct validity was demonstrated in that violent offenders showed higher aggression scores than non-violent offenders. Proactive aggression showed different developmental trajectories within non-offender versus criminal samples, indicating that this form of aggression may be more pathological.