The inverse relationship between test anxiety and test performance is commonly explained by test-anxious students' tendency to worry about a test and the consequences of failing. However, other cognitive facets of test anxiety have been identified that could account for this link, including interference by test-irrelevant thoughts and lack of confidence. In this study, we compare different facets of test anxiety in predicting test performance. Seven hundred thirty university students filled out the German Test Anxiety Inventory after completing a battery of standardized tests assessing general intelligence and mathematical competencies. Multiple regressions revealed that "interference" and "lack of confidence" but not "worry" or "arousal" explained unique variance in students' test performance. No evidence was found for a curvilinear relationship between arousal and performance. The present results call for revisiting the role of worries in explaining the test anxiety-performance link and can help educators to identify students who are especially at risk of underperforming on tests.