A review of research indicates that cognitive restraint is insufficient in accounting for the relationship between restraint and negative affect eating. To explore what mechanism may be responsible for restraint effects, college students in two samples (Total N=378) completed the Three-Factor Eating Questionnaire (TEQ), a restraint scale consisting of three factors: cognitive restraint, disinhibition, and hunger sensitivity. Data analysis revealed a small inverse correlation between cognitive restraint and hunger sensitivity. Hunger sensitivity was strongly correlated with disinhibition, while there was no correlation between cognitive restraint and disinhibition. In a second study, normal weight female college students completed the TEQ. Afterwards, subjects' moods were manipulated using the Velten Mood Induction Procedure and subjects were given candies to eat while completing the Velten procedure. Although interim data on 47 subjects from this 2 (hunger sensitivity level) by 2 (neutral or depressed mood) factorial study appeared to show an interaction, with depressed high hunger sensitivity subjects eating approximately three times as much as other subjects, the level of variability in eating was very high and the interaction was not statistically significant. The results did suggest that hunger sensitivity may be a predictor of negative affect eating. (NRB)