The first aim of the present study was to examine whether the items on the Alcohol Expectancy Questionnaire Tension-Reduction Subscale (TR-AEQ) measure a common underlying factor. The second and third aims, respectively, were to assess the factorial invariance of the TR-AEQ across gender and drinking frequency. Subjects were 503 social-drinking college undergraduates. Confirmatory factor analyses indicated that a common-factor model with seven pairs of correlated disturbances fit the data well. Equivalence testing yielded factorial invariance across gender and partial factorial invariance across drinking frequency. These data support the factorial validity of the TR-AEQ and have implications for psychometric assessment and the evaluation and identification of problem drinking behavior.