Towards an Explanation of Overeating Patterns Among Normal Weight College Women: Development and Validation of a Structural Equation Model
Although research describing relationships between psychosocial factors and various eating patterns is growing, a model which explains the mechanisms through which these factors may operate is lacking. A model to explain overeating patterns among normal weight college females was developed and tested. The model contained the following variables: global adjustment, eating and weight cognitions, emotional eating, and self-efficacy. Three hundred ninety-one participants completed the following self-report indices: the Questionnaire on Eating and Weight Patterns-Revised, the Student Adaptation College Questionnaire, the Weight Efficacy Life-Style Questionnaire, the Center for Epidemiological Studies on Depression, the State-Trait Anxiety Inventory, the State-Trait Anger Expression Inventory, the Emotional Eating Scale, the COPE, the Dutch Eating Behaviors Questionnaire - Restraint Scale, and a self-reported frequency of current eating patterns. Forty participants were excluded based on responses suggestive of obesity (BMI>27.3), severe dietary restraint, or bulimia nervosa, resulting in a final sample of 351. Correlational matrices, factor analysis and structural equation modeling with LISREL 8.B were progressively used to develop the best measurement model and assess the goodness of fit of the proposed structural model. The model provided an excellent fit to the data (GFI=.95; AGFI = .92; RMSEA = .048) and explained as large amount of the observed variance in overeating patterns among normal weight college females (R² = .78). An alternative model, which included dietary restraint as a predictor variable was also tested and compared to the proposed structural model. On all indices of model fit and model parsimony, the proposed model without dietary restraint appeared superior. Moreover, dietary restraint was not a significant direct contributor to the explanation of overeating patterns among normal weight college females. In the final structural model, all variables had a significant direct effect on eating patterns (p < .01). Further examination revealed a large total effect of adjustment as well as a strong direct influence of emotional eating on overeating patterns (direct effect =.52, p <.001). Because emotional eating captures the extent to which negative emotions produce an urge to eat, treatment and prevention programs should specifically target acquisition and practice of alternative coping strategies for dealing with negative emotions.