BACKGROUND: Two competing theoretical models to explain academic performance were proposed. The interference model stresses the detrimental effect of task-irrelevant thoughts during the test-taking situation whereas the deficit model suggests Study Habits and domain-specific skills as main predictors of test performance. AIMS: The study compares the two models by determining the relative contribution of Test Anxiety, Study Habits, and Maths Skill to performance in a statistics exam. SAMPLE: Sixty-six undergraduate students who were enrolled in the first semester of two parallel introductory statistic courses participated in the study. METHOD: Hierarchical regression analyses were performed on the performance in the final statistics exam. The unique variance attributable to Test Anxiety, Study Habits, and Maths Skill was calculated. RESULTS: Both Maths Skill and Test Anxiety added unique variance in explaining performance, whereas Study Habits did not. Although Maths Skill emerged as relatively more important than Test Anxiety, a purely deficit-based account nevertheless appears untenable because interfering effects of Test Anxiety during the examination also contributed an important portion of variance. CONCLUSIONS: It is recommended that cognitive-attentional accounts stressing test anxiety be supplemented by a deficit formulation, and that multimodal counselling address both Test Anxiety and skill deficits. COMMENT: Methodological problems in investigating the causal relationship between skill deficits, anxiety, and performance are discussed.