Martijn van Teffelen

Interpretation bias modification for hostility 129 6 each statement on a 4-point Likert scale ranging from 0 (not likely at all) to 3 (very likely). Internal consistency of the SIP-AEQ is good and convergent and discriminant validity are adequate (Coccaro et al., 2009). State anger. State anger was measured using the seven items of the anger subscale of the Profile Of Mood States (POMS) (McNair et al., 1992). Participants were asked to indicate the extent to which items (i.e., “bad-tempered”, “annoyed”, “rebellious”, “furious”, “grouchy”, “angry”, and “on edge”) reflect their current mood state on a 5-point Likert scale ranging from 0 (not at all) to 4 (very much). Internal consistency of the POMS is good and validity is adequate (Wicherts & Vorst, 2004). Procedure The Ethical Review Committee Psychology and Neuroscience at Maastricht University provided ethical approval to carry out the study (ECP- 170_09_11_2014). The study took place between February and July 2016. When participants arrived at the lab, the study’s procedure was explained and written informed consent was obtained. In our information letter we explained participants that we were studying how we can reduce feelings of hostility with a computer training. We explained that our training aimed to change the thought processes that are involved in hostility. Next, participants were told that the experiment involved two experimental conditions; an active condition and a placebo condition to which participants were randomly allocated. An independent technician from another department at Maastricht University carried out permutated block (i.e., blocks of 4 and 6) randomization (stratified by gender) using https:// www.randomizer.org/. The participants were blind to the condition. Participants completed bias and anger measures prior to engaging in the first session of their allocated intervention. The next seven sessions and post-intervention assessment were carried out at home. During the last session, bias and anger measures were again completed. The first and last session took approximately one hour, and the other sessions took about 20 minutes. At the end participants were fully debriefed, thanked and reimbursed with €35 for their participation. Statistical analyses SPSS version 24 was used for all analyses. First, means and standard deviations were computed to examine the baseline characteristics of bias and anger measurements. Second, independent t -tests or Man-Whitney U -tests were run to examine baseline differences. Third, to test the hypothesis that CBM-I leads to greater reductions in benign and hostile bias and anger three repeated measures ANOVA models were run. In each model, condition, time, and condition by time interaction were entered as predictors. Analytic practices in the current multi- session CBM-I literature on how to deal with missing sessions are divergent. That is, some authors choose to exclude cases with missing data (Lang et al., 2012), while others decide to include cases that at least underwent at least 75% of the intervention (Brettschneider et