Martijn van Teffelen

Provoked aggression, psychopathy and narcissism 55 3 female lab-assistant provided the participant with negative verbal comments such as “You should try harder” or “This way you will not get a good estimation”. During the task, the experimenter left the room to provide the student-assistant the opportunity to practice her skills in guiding the task. After the questionnaire, participants received standardized bogus feedback indicating that they had only achieved an average score on the test, and that such a score is unusual and unacceptable for a college student. If participants were non-students, the bogus feedback indicated that they had only achieved an average score on the test, and that this score was lower than they predicted. Research showed that that this procedure significantly increases anger (Gilbert & Thompson, 1999; Jäncke, 1996; Lobbestael et al., 2008). Aggression Aggressive behavior was measured using a competitive reaction time task (CRTT) (Warburton & Bushman, 2019). In this task, participants engaged in a game against a computer of whom they thought it was another participant. The goal was to click the mouse button as fast as possible on a rectangle when it changed from yellow to red. The time it took to turn from yellow to red randomly varied between 1000 and 2000 ms. The player that supposedly reacted fastest won the trial. In reality, the game was programmed to let the player win in approximately 50% of trials, in a total of 25 trials (see Appendix A for the pre-set trial specifications of intensity, duration and win/lose). The participant that lost received the unpleasant tone. The duration and loudness of this tone was determined by the other player prior to each trial, with a maximum of 100dB. The registered loudness and duration of tones served as a quantitative index of physical aggression (Giancola & Parrott, 2008). At present, there is no consensus on how to score the CRTT. To avoid ‘cherry picking’, we followed the scoring procedure described by Brugman et al. (2015). First, principal components analysis (PCA) with orthogonal rotation was run to determine the underlying data structure. As the number CRTT trials variables high (i.e., 2 times 25) and sample size was rather small we decided to cluster CRTT trials based on three characteristics: response type (intensity/duration), preceding experience (win/loss) and phase (before/after receiving a first noise blast by the opponent). This resulted in eight items (e.g., trials won before first provocation – intensity) who were entered in the PCA. Based on Scree-plot and Kaiser-criterion (i.e., eigen values greater than one) inspection the PCA resulted in 2 factors: unprovoked aggression (i.e., intensity/duration on trials before receiving a first noise blast from the opponent), and provoked aggression (i.e., intensity/duration on trials after receiving a first noise blast from the opponent). This factor structure converges with earlier findings (Brugman et al., 2015). The unprovoked and provoked aggression factor scores were used in the analysis.