Three groups were compared on a number of cognitive measures: 27 women who fulfilled operationalized DSM-III criteria for bulimia, 27 women who regularly binge-ate but did not fulfill the operationalized criteria for bulimia, and 27 control women who did not binge eat. Subjects were compared on measures of: 1) endorsement of rigid, perfectionistic beliefs; 2) confidence concerning their ability to argue effectively against such beliefs; 3) independent ratings of the effectiveness of their arguments; and 4) self-evaluations of the effectiveness of their arguments. These measures were assessed for two types of rigid, perfectionistic beliefs: those which concerned issues of eating, weight, and appearance (idiosyncratic beliefs) and those which concerned other personal issues (common beliefs).

Bulimic women were found to endorse rigid, perfectionistic beliefs to a significantly greater extent than either comparison control group, giving especially high endorsements to beliefs concerning issues of eating, weight and appearance. Contrary to hypotheses, no differences were found between groups on their confidence ratings, ability to generate rational arguments, or evaluations of the effectiveness of their arguments for either idiosyncratic or common beliefs. However, a main effect of order of belief packet presentation was found which indicated that it was harder for women across groups to generate rational/effective arguments against the idiosyncratic belief statements, if they were somewhat fatigued by having had already argued against a number of common beliefs. This finding was interpreted as suggesting that factors, such as fatigue, which may impede concentration and affect the ability to think rationally about issues such as those described by the idiosyncratic belief statements.

Overall these results lend partial support for cognitive-behavioral assumptions of bulimia and its correlates. The results support the assumption that bulimic women maintain a number of rigid, perfectionistic beliefs, especially concerning issues of eating, weight and appearance. Moreover, the results suggest that the assumption that bulimic individuals lack active cognitive coping strategies or are less able to respond to rigid, perfectionistic thoughts with strategies such as rational self-statements, needs to be refined. Implications of these findings for the assessment and treatment of bulimia were discussed.