Martijn van Teffelen

Interpretation bias modification for hostility 135 6 Trait measures . To measure different aspects of trait hostility participants completed the hostility scale of the Personality Inventory for DSM-5 (PID-5H), the AQ-H (measuring cognitive aspects of trait hostility), FOA (measuring self-reported trait aggression) and the voodoo doll task (VDT) (DeWall et al., 2013). The 10-item PID-5H measures overall trait hostility and was scored on a 4-point Likert scale ranging from 0 (very false or often false) to 3 (very true or often true). The VDT measures behavioral aggression. During the VDT participants are presented with an ambiguous vignette on a computer (e.g., “You are carrying a heavy load of groceries up to a check-out line at the grocery store and just as you are about to enter in line, someone cuts in front of you. You end up dropping some things on the floor”) (Tremblay & Belchevski, 2004). Participants were then allowed to insert up to 51 pins into a voodoo doll that represents the other person from the vignette. The PID-5H, AQ-H and the VDT have demonstrated good reliability – in the present study α ’s range from .81 to .93- and adequate validity (DeWall et al., 2013; Krueger et al., 2012). Working alliance. To measure potential carry-over effects of CBM-I on working alliance in subsequent psychotherapy we administered the Working Alliance Inventory (WAI), client version (Horvath & Greenberg, 1989), which measures the quality of the therapeutic working alliance between patient and therapist. The 36-item WAI was scored on a 5-point Likert scale ranging from 1 (never) to 5 (always). The WAI shows adequate reliability – in the present study α = .93- and adequate criterion validity (Horvath & Greenberg, 1989). Procedure The Ethical Review Committee Psychology and Neuroscience at Maastricht University provided ethical approval to carry out the study (ERCPN_ 170_09_11_2014_A6). This study was preregistered at https://osf.io/r46jn. The study was registered semi-retrospectively, after seven participants were recruited. Recruitment took place from September 3 rd , 2018 until May 11 th , 2020. The procedure was similar to Study 1, except for four notable differences. First, the first session from this study was carried out at home instead of in our lab. Second, Study 2 was conducted as a double-blind trial, instead of a single-blind study, implying that both patients and experimenter were blind for allocated condition. Third, we administered interpretation bias, state, and trait measures instead of solely interpretation bias and state affect measures. All instruments were measured pre- and post-intervention. Fourth, we monitored if participants would enter therapy after our intervention. When this occurred, we offered one additional ‘booster’ CBM-I or AC intervention session to account for variability in time between the end of the interventions and the start of therapy. The booster sessions consisted of 33 additional scenario training and 38 additional WSAP trials. Five weeks after start of therapy, we administered the WAI once.