This study examines the structural validity of scores from the Bem Sex Role Inventory using a maximum likelihood confirmatory factor analysis (CFA). Six hundred and sixty-five graduate and undergraduate students participate in the study. A seven first-order factor model almost identical to the model reported in a previous CFA study is used as the baseline comparison model. The data for testing these models are obtained from an exploratory sample randomly selected from the whole sample. A hierarchical factor structure model with seven first-order factors (compassionate, interpersonal affect, shy, dominant, decisive, athletic, and self-sufficient) and two second-order factors (masculinity and femininity) fit the data quite well. The fit indices based on the validation sample collectively indicate a very good fit. The results of this study are notably consistent with the hierarchical factor models suggested in two previous CFA studies. (Contains 2 tables and 1 figure.)