[Predicting factors of eating disorders and general psychological symptoms in female college students]

To compare and determine the relative effects of six of the variables (family functioning, self-esteem, obsessive-compulsive symptoms, depressive symptoms, locus of control and Body Mass Index-BMI) related to the etiology of eating disorders in predicting eating disorder symptoms and the general symptom index.

Two hundred and ninety-two female college students completed the Rosenberg Self-Esteem Scale (RSES), the Family Assessment Device (FAD), the Eating Attitude Test (EAT), the Symptom Check List (SCL-90-R), Rotter's Internal-External Locus of Control Scale (RIELCS), and the Beck Depression Inventory (BDI). Multiple regression analysis was employed, and scores of the obsessive-compulsive subscale of SCL-90-R, BMI, RSES, FAD, BDI and RIELCS were used as predictors of EAT total scores. The second multiple regression analysis was employed with the same independent variables as predictors of the General Symptom Index (GSI) of SCL-90-R.

Obsessive-compulsive symptoms were the best predictor of the total score of EAT, and the second was the BMI scores. Depressive symptoms and self-esteem scores were significantly correlated with EAT scores but these variables were not chosen for inclusion in the regression equation. There was a strong relation between scores on BMI and scores on EAT. BMI scores were correlated nonsignificantly with the scores of the other variables. RIELCS and FAD scores were not significantly correlated with EAT scores. Obsessive-compulsive symptoms were also the best predictor of GSI of SCL-90-R and the second best predictor was the RSES. FAD was significantly correlated with SCL-90-R GSI scores, depression and obsessive compulsive scores.

It appears that obsessive-compulsive symptoms were the best predictor of pathologic eating behavior. BMI was the second predictor of EAT scores. Locus of control and family functioning were not correlated with eating pathology. Family functioning scores were significantly correlated with the GSI scores, and depressive and obsessive-compulsive symptoms. Obsessive-compulsive symptoms were also the best predictor of the GSI scores.