Social cognitive theory suggests that self-constructs are better predictors when the specificity levels of self-constructs (as predictors) and the criterion variable closely correspond. In this study it was examined whether self-constructs measured at an intermediate level of specificity better correspond with the criterion variable (course grades: an intermediate level of specificity) than general self-constructs or specific self-constructs. Self-constructs in this study included self-efficacy and self-concept at varying degrees of specificity: general self-efficacy, academic self-efficacy, specific self-efficacy, academic self-concept, and specific self-concept. College students (N ϭ 230) in general education courses voluntarily completed the three self-efficacy measures and the two self-concept measures. Results of correlation and simultaneous multiple regression analyses indicated that the closer the level of specificity of self-efficacy and selfconcept, the stronger the relationship between the two constructs. Both academic self-concept and specific self-concept were significant predictors of term grades. In addition, academic selfconcept was also a significant predictor, whereas neither general self-efficacy nor academic self-efficacy was significant. Measurement issues and implications of the findings are discussed.

This study reports on a summary of 23 separate exploratory factor analytic validation studies of the Bem Sex Role Inventory (BSRI) conducted during the 25 years since its publication. The ERIC and PsycLit databases were used to locate those validation studies. Number of items factor analyzed, sample characteristics, extraction and rotation methods used, names of factors, and factor structures reported in each of 23 studies were tabulated and analyzed. Two to four factors were retained in most of the studies, with masculine factors showing more factorially complex tendencies. Findings from this study suggest that masculinity and femininity have not been adequately operationalized in the BSRI and that the true structure of masculinity/femininity may be more complex than it appears when measured by the BSRI, which suggests that the BSRI may not capture the complex nature of masculinity/femininity. Implications of the results and the potential usefulness of the BSRI are discussed.

The short form of the Bem Sex Role Inventory (BSRI) contains half as many items as the long form and yet has often demonstrated better reliability and validity. This study uses exploratory and confirmatory factor analytic methods to examine the structure of the short form of the BSRI. A structure noted elsewhere also emerged here, consisting of two masculine factors and a single feminine factor. The three-factor model was found to be invariant across gender groups and also across two divergent samples, the first sample of college students and the second sample of accountants. As expected, women were found to score higher on the feminine factor. On a masculine factor that seemed to represent social control, men scored significantly higher than women did. However, no differences were found between men and women on a second masculine factor that seemed to represent a more internal, self-control dimension.

The primary purpose of this study was to examine sex role group differences (androgynous, masculine, feminine, and undifferentiated) in three different levels of self-efficacy (general, academic, and course specific). A sample of 215 undergraduate participants completed the three measures of self-efficacy and the sex role measure. Results showed that there was a multivariate significance among the sex role groups in the three levels of self-efficacy and that both masculine and androgynous groups had significantly higher general and academic efficacy means than the undifferentiated or feminine groups. The major findings in this study are largely consistent with findings of previous researchers, supporting the theory that masculinity may be a key construct in the differentiation of individuals in self-efficacy, which has been shown to be an important predictor of achievement.

The purpose of the present study was to examine the internal structure of scores from Bandura’s Multidimensional Scales of Perceived Self-Efficacy (MSPSE). Data were obtained from college freshmen, a population previously untested with MSPSE. Results show that scores from nearly all items aligned almost exclusively with theoretically and empirically derived dimensions. Some differences in empirically derived dimensions occurred as a result of extraction procedures, although the differences observed still provided theoretically interpretable constructs. These results suggest that Bandura’s scales, with slight revisions, show promise for research on self-efficacy with undergraduate college students.