Attention Allocation and Incidental Recognition of Emotional Information in Dysphoria

For secondary analyses, we also calculated the percent correct within each emotion category (i.e., the number of correct yes and no responses given for dysphoric, aversive, positive, and neutral stimuli / total responses × 100). A 4 (valence: dysphoric, aversive, positive, neutral) × 2 (dysphoria group: dysphoric, non-dysphoric) repeated measures ANOVA examined whether dysphoria groups differed in percentage of words from each stimuli category correctly recognized. The valence main effect, F(3, 156) = 17.32, P < .001, partial η² = .25, was significant. LSD pairwise comparisons indicated that percent of correct responses was lowest for dysphoric words compared to aversive, neutral and positive words (Ps < .01). The interaction between word valence and dysphoria group was non-significant, F(3, 156) = .91, P = .44, partial η² = .02. Further, the main effect of dysphoria group was also non-significant, F(1, 52) = .01, P = .94, partial η² = .00. Thus, percentage of correctly identified word stimuli did not differ for dysphoria groups (see Table 1).

An indirect effect is the product of the direct effects that comprise the indirect effect pathway. Although exact significance tests are available for indirect effects involving up to three variables, Cohen and Cohen (1983) suggested that a reasonable proxy for exact significance tests for indirect effects involving four or more variables is whether each of the constituent direct effects are significant. If each direct effect is significant, as in our case, Cohen and Cohen (1983) suggested that the indirect effect is also likely significant. However, these effects tend to be small, given that the indirect effect size is a product of multiple direct effects.