The current study investigated the impact of Learned Helplessness (LH) training on normal (N), depressed (D), and bulimic (B) female college students (N=135). Participants were selected through a screening procedure using an eating behavior questionnaire and the Beck Depression Inventory (BDI; Beck, 1971) as the primary instruments. Bulimic subjects met the DSM-III criteria for that disorder, and depressed subjects met a criterion of 20 or above on the BDI. Normal subjects had BDI scores below 10 and endorsed 1 or less of the critical items regarding eating behavior which had been used to select bulimic subjects. A11 three groups were matched for weight, height, and age. Each category of subjects was divided into three groups which were exposed to Learned Helplessness (LH), Contingent Feedback (CF), or No Training Control (NTC) conditions. Dependent measures included performance on an anagram test (latency and errors), performance prediction and performance satisfaction questionnaires, and a Self-Statement Test. Results indicated no difference in anagram performance following LH training relative to CF and NTC conditions for normal and bulimic subjects, although depressed subjects demonstrated longer response latency and more errors as a function of training (LH>CF). NTC subjects did not differ significantly from the other conditions, however. Ratings of performance satisfaction differed as a function of training condition in the expected directions. Bulimic subjects demonstrated an interesting pattern of declining ratings of performance satisfaction across training trials within the CF condition. Also, bulimic subjects were found to make more errors on anagrams of nonfood words relative to food words, although this pattern was not reliably related to training condition. Self-Statement Test items did not produce the expected group x training condition effects. The present findings provide limited support for the experimental hypothesis that ineffective coping in stressful situations, mediated by cognitive self-statements, is a useful model for understanding bulimia. Methodological issues are discussed, particularly issues related to subject classification. The present study was likely confounded by the presence of depression (i.e., high BDI scores) among the bulimic subjects. Future research should systematically‘ address the relationship between bulimia and depression. / Ph. D.