Martijn van Teffelen

Provoked aggression, psychopathy and narcissism 53 3 METHODS Participants Participants were recruited through the use of flyers on university campus. Inclusion criteria were age between 18 and 60 and male gender. Advanced level psychology students were excluded due to expected familiarity with the study’s procedures and knowledge of the study’s topic. Women were excluded to prevent floor effects because psychopathic and narcissistic traits are more prevalent in men (Coid & Yang, 2011; Grijalva et al., 2015). The main hypothesis was tested using Bayesian statistical analysis. Therefore, traditional power analysis was not possible. In Bayesian statistical analysis, “one uses an entire distribution of parameters instead of a single point value for the effect size” (Kruschke, 2013). Specifically, precision planning in Bayesian statistical analysis uses posterior distributions of previous data. As such distribution is absent at present, a default prior distribution was chosen. For more information on this, see Kruschke and Liddell (2018). In total, 94 people entered the study. To give an indication of the relative precision of the present data set in frequentist statistical terms, a group size of N 1 = N 2 = 47 would have enabled the detection of a between subjects difference in CRTT response of .36 with 80% power, using a standard deviation of SD = .63 (Bobadilla et al., 2013) and α = .05. Prior to conducting data analysis, we excluded people who did not believe the provocation procedures ( n = 8). The final sample thus consisted of N = 86 people with a mean age of 28 years, of who the majority was student (76%). Sample characteristics are shown in Table 1. Analyses showed that the groups did not significantly differ in age, nationality, education level and current work situation. Experimental Provocations Social Exclusion: Negative Social Belonging Feedback In this condition, participants were told that we were interested in testing different aspects of social interaction and that they were to interact with a second participant, who was in fact a female confederate. In this interaction task, the participant and the confederate were set in separate rooms and received the instruction that they were to engage in interaction by speaking to each other through a microphone. During the task one of them would take on the role of ‘speaker’ while the other would take on the role of ‘evaluator’. We told the participants that these roles were assigned to them randomly, whereas in reality participants always received the role of speaker. When the task started, the evaluator (i.e., the confederate) read aloud questions from the Relationship Closeness Induction Task (Sedikides et al., 1999). The sequence of questions started out with general questions (e.g., “What is your first name?” and “How old are you?”) and gradually became more personal (e.g., “Describe the last time you felt lonely” and “If you could change one thing about yourself, what would that be?”). The speaker (i.e., the participant) answered these questions. Each minute the evaluator then supposedly rated on a scale from one to seven