Martijn van Teffelen

Provoked aggression, psychopathy and narcissism 59 3 Table 2 Study means and standard deviations and difference testing Social exclusion ( n = 42) Insult ( n = 44) Diff. test Bayes’ factor Mean ( SD ) Mean ( SD ) t ( p ) / Z ( p ) BF 10 Unprovoked aggression -.02 (1.11) .04 (.92) -.38 (.702) .24 Provoked aggression a .02 (1.02) .08 (1.07) -.27 (.787) .24 Negative affect Baseline b 14.43 (5.59) 12.57 (2.65) -1.54 (.124) 1.24 Post-manipulation b 15.45 (6.54) 13.64 (4.14) -1.01 (.311) .64 Change b 1.02 (3.74) 1.07 (3.32) .336 (.737) .23 Threat perception 39.92 (18.84) 46.67 (24.54) -1.43 (.157) .55 Psychopathy 297.86 (28.60) 298.59 (31.58) -.11 (.910) .23 Narcissism 148.31 (22.32) 151.53 (23.62) -.65 (.519) .27 Note. BF 10 = Bayes factor. a n=7 cases were excluded who did not believe the aggression manipulation. b These variables were not normally distributed. First, a Wilcoxon signed-rank test was run to test the prediction that negative affect increases after provocation. The test was run on baseline and post-provocation scores averaged across condition and showed that negative affect increased from pre- ( M = 13.48, SD = 4.42) to post- ( M = 14.52, SD = 5.49) provocation, Z = 2.09, p = .037, d = .46. These results suggest that our provocation procedures increased negative affect. Then, Bayesian statistical analysis was performed to examine differences or equality between social exclusion and insult, in terms of (un)provoked aggression, change in negative affect, and threat perception (Hypothesis 1). The null hypothesis (i.e., no difference) was tested against the alternative hypothesis that there was a difference in terms of these variables. In the analyses with provoked aggression as (dependent) variable, seven people were suspicious after they first received a noise blast from the opponent and thus removed from all further analyses. Bayes’ factors (see Table 2) indicate moderate evidence in favor of the null hypothesis. More precisely, these findings indicate that for unprovoked aggression, provoked aggression, negative affective change, and threat perception, the data was respectively 1/ BF 10 = 4.16, 4.15, 4.35, and 1.81 times more likely to have occurred under the null-hypothesis than under the alternative hypothesis. In other words, there was no difference between social exclusion and insult in terms of unprovoked and provoked aggression, negative affective change, and threat perception.