In contrast, the current study based assessment of PD traits on the the interview methodology of the SCID II. Alternatively, the lack of correlation between borderline and narcissism traits on the one hand and aggression types on the other hand in the current study is based on a statistical approach where the effects of all other PD traits were controlled for. It might be the case that narcissism traits are indeed associated with increased levels of aggression but that this relationship is in fact driven by comorbid levels of other PD traits, like antisocial PD. This is supported by the facts that at a zerocorrelation level, borderline and narcissism traits were significantly related to both raw reactive and proactive aggression. Other studies of these PDs and aggression only examined one of these two PD (traits) in isolation. Berman et al. (1998) were an exception to the latter generalization and did control for comorbid PD pathology, but they also failed to find aggression to be uniquely related to borderline or narcissistic PD.