Mixed Design ANOVA Analysis of Bulimia Methods The objective of this paper is to assess the effect of

Mixed Design ANOVA Analysis of Bulimia

Methods

The objective of this paper is to assess the effect of perfectionism and perceived weight concern (and possibly the interaction of these two factors) on the change in the disordered eating score in time, as measured by the Eating Disorder Examination - Questionnaire (EDE-Q), measured before and after the treatment. A mixed-design three-way ANOVA is performed to assess the research hypothesis. The within-subject factor is time, and the between-subject factors are perfectionism and perceived weight. A group of 160 subjects is randomly assigned to one of the four treatment cells formed by the 2x2 combinations of the factor levels: (1) An active control counseling group with no emphasis on perfectionism or weight concern, (2) a perfectionism intervention group, (3) a weight concern treatment group, and (4) a combined treatment for perfectionism and weight concern. The scores for these 160 subjects were obtained before and after the intervention.

Results

A mixed-design three-way ANOVA was conducted with the aid of the SPSS software package. Table 1 in the Appendix shows that Time is a significant factor ( F(1, 156) = 365.930, p < .001, partial eta ² = 0.701 ), and in fact there is a significant overall decrease in the EDE-Q scores after the intervention, when controlling for the between-subjects factors. This change in EDE-Q scores after the intervention is also affected by the levels of the two between-subjects factors, as the interaction terms Time*perf_focus (F (1, 156) = 63.20, p < .001, partial eta ² = 0.288 ), Time*wc_focus (F(1, 156) = 17.07, p < .001, partial eta ² = 0.098 ), and Time*perf_focus*wc_focus (F(1, 156) = 14.259, p < .001, partial eta ² = 0.084 ), are all significant (Consistent results with these results are obtained through the within-subjects effects ANOVA table, in Table 2, in the Appendix).