Homework compliance has been identified as a robust predictor of treatment outcome for depression and, to a lesser extent, anxiety disorders, including social phobia. The current study tested the following hypotheses: (1) homework compliance is positively related to ratings of global improvement, (2) homework compliance is negatively related to symptom reduction, (3) the relation between homework compliance and treatment outcome varies according to the nature of the homework exercise, and (4) expectancy is positively related to early homework compliance, in a clinical sample of individuals with social phobia. Results provided limited support for the relation between compliance and ratings of improvement, but did not support a negative relation between compliance and symptom reduction. Further, the results provided limited support for the hypothesis that compliance with exposure versus non-exposure homework would differ significantly in terms of their relation to treatment outcome, but did not support the relation between compliance and expectancy.