Word Association Task. As three versions of this task were employed in this study that differed according to sets of traits paired with each category (Table 5) this enabled an analysis to determine whether responses to three sets of a given trait (e.g., 3 sets of 12 male traits) elicited the same response latency on average. Accordingly, a repeated measures 3 x 9 ANOVA was conducted, with task version (A, B, C) the between subjects factor, and pairing condition (Table 4) the repeated measure.
Prior to conducting this analysis the extra assumptions for a two-way repeated measures ANOVA were evaluated. As the homogeneity of variance assumption was not met, a more stringent criterion of alpha was set (╬▒=0.025) (Tabachnick & Fidell, 1996). The sphericity assumption was also violated, and so results correcting for this factor were obtained.
The main effect of task version (A, B, C) was not significant. Nor was there a significant interaction between task version and pairing condition. This indicated that as task versions varied according to the set of traits paired with the each category label (refer to Table 5), similar responses were elicited to each of the three sets of the one type of trait. As such, results from each version were averaged together for subsequent analyses.
Sentence Completion Task. As there were two versions of this task, an independent samples t-test was performed to determine whether the tendency to explain stereotypic behaviours differed between versions.
In conducting this test, it was evident that the assumption of homogeneity of variance between groups was satisfactorily met. The result of this analysis indicated that while the tendency to explain congruent behaviours did not differ between versions, the tendency to explain incongruent behaviours did (t(38) = -3.084, p < 0.005). This difference may have been due to incidental differences in the need to explain a behaviour (e.g., went to a motor show may have elicited explanations irrespective of its pairing, and let the waiter ignore her/him for fifteen minutes may rarely have elicited an explanatory completion). It was further noted that there was no difference in the tendency to explain stereotypically male as opposed to stereotypically female behaviours.