Martijn van Teffelen

Chapter 6 130 al., 2015). This led us to decide that participants who did not complete more than 75% (i.e., six out of eight) sessions were excluded from the analysis ( n = 9). Non-normally distributed variables were square root transformed to obtain normality. Simple effects were analyzed using independent samples t -tests. Results Baseline levels of hostile bias, benign bias and state anger revealed no significant differences ( p s > .139) between the conditions, indicating that random allocation was successful. Means and standard deviations of benign bias, hostile bias, and state anger levels were respectively M = 5.0, SD = 1.7, M = 4.8, SD = 2.8, and M = 7.9, SD = 6.7 in the CBM-I condition and M = 6.4, SD = 2.2, M = 3.2, SD = 2.6, and M = 6.7, SD = 7.0 in the AC condition. Next, to test the hypotheses that CBM-I results in a greater increase in benign bias, greater reduction in hostile bias and greater reduction in state anger reactivity, three repeated measures ANOVA models were run. Condition and time were entered as respectively between-subjects and within-subjects factors. First, findings on benign bias revealed that the condition by time interaction ( F (1, 18) = 5.30, p = .034) factor was significant, whereas the time ( F (1, 18) = 3.82, p = .066) and condition ( F (1, 18) = 0.05, p = .829) were not significant. Follow-up t -tests showed that benign bias significantly increased from pre- to post-intervention in the CBM-I condition ( t = 2.93, p = .019), but not in the AC condition ( t = 0.26, p = .803) (between subjects d = 0.51 at post-intervention, 95% CI = -0.43 to 1.46). Second, findings on hostile bias showed that the main effect of time was significant ( F (1, 18) = 9.70, p = .006), whereas there was no significant main condition effect, and no condition by time interaction effect ( p s > .319). Third, results on state anger reactivity (i.e., change in anger from pre- to post-session) demonstrated that the time effect was significant ( F (1, 18) = 10.50, p = .005), whereas the condition and the interaction were not significant ( p s > .316). Study 1 - Discussion In Study 1, we examined the feasibility of an eight-session CBM-I intervention. In line with our expectations, findings revealed that CBM-I altered benign bias in a positive direction compared to AC. Contrary to our expectations, results revealed a non-differential decrease over time for hostile bias and anger reactivity for both conditions. One explanation for this pattern of findings is that no such effects are present. Other explanations are that the current sample is too small to detect such effects, or that hostility levels were not high enough. In short, Study 1 revealed that CBM-I has moderate to good feasibility, at least in terms of improvement in benign bias. Therefore, we tried to replicate these findings in a larger trial, which was the focus of Study 2.