Martijn van Teffelen

Provoked aggression, psychopathy and narcissism 61 3 Table 4 Multiple linear regression of aggression, negative affect change and threat perception on psychopathic and narcissistic traits B SE β p DV: unprovoked aggression R 2 = .01, F (2, 83), p = .748 Psychopathic traits .00 .00 .09 .478 Narcissistic traits -.00 .01 -.02 .885 DV: provoked aggression a R 2 = .07, F (2, 76), p = .078 Psychopathic traits -.01 .01 -.21 .116 Narcissistic traits .01 .01 .31* .026 DV: negative affect change R 2 = .04, F (2, 83), p = .158 Psychopathic traits -.03 .02 -.23 .070 Narcissistic traits .01 .02 .05 .671 DV: threat perception R 2 = .07, F (2, 83), p = .057 Psychopathic traits .06 .09 .07 .553 Narcissistic traits .20 .12 .21 .095 Note. * significant at p <.05; ** significant at p <.01. a 7 extra case excluded as they did not believe the aggression manipulation; when cases are included psychopathic traits significantly predict provoked aggression. DV = dependent variable. Following this, a multiple regression model was run to test the hypothesis that psychopathic traits are negatively associated with negative affective responding in response to threat and, in turn are negatively associated with aggressive responding (i.e., mediation model) (Hypothesis 2). The results showed that psychopathic traits negatively related to negative affective change from pre- to post-provocation (i.e., t = -2.37, p = .020) and that negative affective change positively related to provoked aggression (i.e., t = 2.68, p = .009). The direct relationship between psychopathic traits and provoked aggression was not significant ( p = .738). The results suggest that decreased negative affective change suppressed aggressive responding in people with psychopathic traits. Last, a multiple regression model was run to test the hypothesis that narcissistic traits are positively associated with perceived threat, especially under conditions of agentic threat and, in turn, are positively associated with aggressive responding (i.e., moderated mediation). (Hypothesis 3). The results showed that there was no meaningful interaction between narcissistic traits and condition in terms of perceived threat ( p = .092). Closer inspection of simple effects revealed a positive relationship between narcissistic traits and perceived threat in the insult condition (i.e., t = 2.86, p = .005), whereas there was no significant relationship between narcissistic traits and perceived threat in the social exclusion condition (i.e., t = .31, p = .756). In turn, perceived threat related negatively to unprovoked aggression (i.e., t = -2.07,