Martijn van Teffelen

Chapter 3 64 dampened narcissism-displayed aggression relationship in case people judge a threat ‘not ambiguous’ (Martinez et al., 2008). Hence, narcissistic traits may predispose to enter a ‘ready- to-attack’ or brooding state when confronted with a perceived unambiguous threat, lashing out when a new threat emerges. Importantly, it has been debated that it is problematic to ‘statistically correct’ psychopathic and narcissistic traits for each other given the high linear inter-relationship. For example, it was shown that narcissism no longer correlated to self- reported aggressiveness after statistically correcting for psychopathy (Vize et al., 2018). The present findings show the opposite. That is, the data revealed a positive association between aggression and narcissism, but only when psychopathic traits were controlled for. Closer inspection of multicollinearity diagnostics revealed that psychopathic traits did not inflate the effect of narcissism on aggression in the present dataset. LIMITATIONS Several limitations impacted the present findings. First, the sample consisted of (healthy) male, predominantly student participants recruited in the general population, troubling the generalization of these findings across gender and to the clinical range of psychopathy and narcissism. Second, our explorative analysis on the relationship between our personality measures and negative affect, threat perception, and aggressive behavior was statistically underpowered. Therefore, these findings require replication. Third, the main aim of the present study was to determine the relative impact of two established provocation procedures. The design therefore did not include a control condition. Although each experimental procedure reliably invoked anger and aggression inclinations in previous work (Buckley et al., 2004; Bushman & Baumeister, 1998; Gilbert & Thompson, 1999; Lobbestael et al., 2008), from this work alone we cannot demonstrate that social exclusion and insult produce more aggression than a control condition. Lastly, the present study examined aggressive responding to an unknown person (i.e., displaced aggression). It would be interesting to observe whether the provoked aggression effect changes or differentiates when people are allowed to respond aggressively to their insulter/rejecter. RESEARCH IMPLICATIONS The present study was primarily designed to directly compare two provocation procedures. Consequently, the present study did not include enough participants to extensively examine trait by condition interaction effects as testing these effects would greatly inflate the chance of type I and II errors. This leaves a number of important questions open to be answered. First, future studies that compare provocation conditions should ideally include a control condition,