Martijn van Teffelen

Chapter 4 88 Third, in the model with state anger (POMS) as the dependent variable (H3) the random effect of therapist on state anger was not significant ( p = .480). Findings showed that state anger increased from baseline (M0) ( M = 2.51, SE = .65) to post-intervention (M3) ( M = 2.28, SE = .93), ( b = 0.55, B = 2.28, t = 2.45, p = .016). Interventions did not differ significantly in terms of change in state anger from baseline (M0) to post-intervention (M3) ( p ’s > .081). Taken together, findings on the immediate efficacy of I-CR revealed that from baseline (M0) to post-intervention (M3) I-CR led to larger reductions in hostile beliefs compared to CR and AC, while both I-CR and CR resulted in a larger decrease in aggressive tendencies and anger compared to AC. Also, all conditions showed an increase in state anger from baseline (M0) to post-intervention (M3). Hostility traits To test the hypothesis that both I-CR and CR lead to a larger reduction in hostility traits at one-week follow-up compared to AC (H4), mixed regression models were run with SIP-AEQ, VDT, STAXI-2T, AQ-H and FOA scores as the dependent variables and time, intervention and intervention by time as independent variables. Therapist was entered as a random effect. Results revealed that the random effect of therapist was not significant in any of the models ( p ’s > .459). The time effect was significant for all dependent variables ( p ’s < .018), except for VDT ( p = .880). SIP-AEQ ( b = -0.45, B = -1.80, SE = .44, t = -4.12, p < .001), AQ-H ( b = -0.67, B = -2.70, SE = .44, t = -6.14, p <.001) and FOA ( b = -0.30, B = -2.07, SE = .77, t = -2.68, p = .009) scores decreased significantly from baseline to follow-up. None of the interaction effects between intervention and time were significant ( p ’s > .088). In sum, findings on the intervention efficacy on hostility traits demonstrated that all hostility-trait-variables scores but the VDT decreased in all experimental conditions. Sustained intervention efficacy at one-week follow-up To test the hypothesis that the intervention efficacy sustained over time we analyzed fixed effects on hostile beliefs and aggressive tendencies at one-week follow-up (T2a) and after an emotional stressor (T2b). These fixed effects are presented in Table 3 (M5 and M6). First, results show that at 1-week follow-up (T2a) hostile belief was lower in both the I-CR ( b = -0.42, t = -3.02, p = .003) and CR conditions ( b = -0.31, t = -2.23, p = .029) compared to AC, but the significant post-intervention (M3) difference between I-CR and CR (when CR was used as a reference category) did not sustain ( B = -4.58, SE = 5.02, t = -.91, p = .364). Also, findings revealed that at one-week follow-up aggressive tendencies were significantly lower in the I-CR condition ( b = -0.28 , t = -2.00, p = .048 ), but not in the CR condition, compared to AC. The difference between I-CR and CR (when CR was selected as reference category) was not significant, ( b = -3.99, SE = 3.49, t = -1.15, p = .256).