The purpose of the current study was to investigate three aspects of the construct validity of attributional style assessment instruments. The first purpose was to determine the independence of stability and globality. The second was to determine if controllability was a dimension of attributional style. The third purpose was to determine if inventories that use real or hypothetical events measure attributional style equally well. One hundred fifty-nine female, and one hudred fifty-five male subjects, completed four questionnaires that assessed attributional style. Results provided some support for the general construct of attributional style. All four factors were found, which demonstrates the support for the four factor model. However, the two factor model may be the best overall method. No method factor (real or hypothetical stimulus event) solution was obtained. Possible sex differences are discussed. The findings are discussed in terms of attribution theory. Suggestions for further research are presented.