Flipbook

CHAPTER 4 106 Main analyses 2 Table 2 shows that there were moderate to strong significant correlations between most questionnaires at the pre-test. Correlations with the OCI-R checking cognitions subscale were added separately as an additional check of the validity of the CCS and Checklist. To test the study hypotheses, a series of 2 (Time: pre-test vs. post-test) x 3 (Condition: experimental vs. monitor vs. control) mixed ANOVAs 1 were conducted to examine the effect of the safety behavior manipulation on checking cognitions (CCS), obsessive beliefs (OBQ-44 RT), and general anxiety symptoms (BAI). Checking cognitions scale (CCS) There was no main effect of Time, F (1,86) = 0.27, p = .60, and a trend of Condition, F (2,86) = 3.10, p = .05, η p ² = .07. The crucial Time x Condition interaction was significant, F (2,86) = 5.62, p = .005, η p ² = .12, see Figure 1. Planned comparisons showed a pre- to post-test increase in the experimental group, t (29) = 2.40, p = .02, d = 0.28, a trend for a decrease in the monitor group, t (28) = 1.97, p = .06, d = 0.14, and no change in the control group, t (29) = 0.56, p = .58. The change in CCS scores differed between the experimental and monitor group, t (49.11) = 3.08, p = .003, d = 0.80; and the experimental and control group, t (58) = 2.27, p = .03, d = 0.59; but not between the monitor and control group, t (57) = 0.89, p = .38. Hence, consistent with our hypothesis, performing OCD-related checking behavior for one week increased participants’ checking-related cognitions. 2 Exploratively, it was examined whether the checking behavior manipulation would have an effect on general checking behavior in an unrelated task. The visual search task of Toffolo et al. (2013) was administered both at pre- and post-test, because this induced and measured checking behavior. In this task participants had to report whether a target was present or absent in a search field (see Toffolo et al., 2013 for details of the task). A 2 (Time) x 2 (Trial type: target-absent vs. target-present trials) x 3 (Condition) mixed ANOVA was performed to compare groups on pre- to post-test changes in search time (checking behavior) in both target-absent and target-present trials. Although overall, participants searched longer in absent- than present-trials, F (1,82) = 572.94, p < .001, η p ² = .88, and there was an overall decrease in search time from pre- to post-test, F (1,82) = 14.48, p < .001, η p ² = .15, the crucial interaction effects were not significant, p s > .26.