Research That Matters (January 17 - 20, 2008)

Purpose: Establishing cross-cultural equivalency of self-efficacy measures is essential to any discussion of an individual's estimate of their capabilities when two or more cultural groups are evident in the sample. Self-efficacy as defined by Albert Bandura is, the self-report of a person's estimate of their capabilities within a specific domain (1977). Schwarzer & Scholz (2002) suggested in their findings that self-efficacy is a universal construct and that it is unidimensional. The purpose of this examination was to determine if Schwarzer & Jerusalem‘s (1995) General Self-Efficacy Scale (GSES) met tests for cross-cultural equivalency. This paper suggests tests for cross-cultural equivalency are not met by the GSES and thus comparison of self-efficacy across cultures may be flawed. Methods: Data was obtained from Schwarzer & Scholz‘s original data set. Two culturally distinct subsamples of adolescent females were selected for analysis. Using LISREL, factor analysis of a sub sample of 15-20 year old females (N=644) found German student (N=422) perceptions of self-efficacy appear to be unidimensional but Japanese perceptions (N=222) do not. Descriptive analyses provided an item-by-item comparison of cross-cultural equivalence across the groups. Analysis shows that all of the cross-cultural equivalency hypotheses tested can be rejected and therefore cross-cultural equivalency is not met. Results: The analysis suggests that the GSES used by Schwarzer & Scholz did not meet the tests of cross-cultural equivalence. The covariance structure and the factor loadings were not equivalent between the samples. The factor pattern was different between the two samples. The factor variances and co-variances were not equivalent between the two samples of adolescent females. Further, the analysis yielded a finding that suggests the GSES is not unidimensional in all cultures. A two-factor model emerged for the Japanese sample of adolescent females Conclusions and Implications: These findings suggest that in research related to self-efficacy the researcher must be fully aware of the distinct differences in how perceptions and beliefs in individual capabilities are both acquired and expressed. In preparing measures of self-efficacy, it may not be enough to have native translations of the measures. It is as important to understand the learning contexts within the culture, as it is to examine the foundations and applicability of social cognitive learning theory in given cultures. As role development and socialization affect both the acquisition and demonstration of knowledge and skills then, it seems intuitive that these would also play a significant role in the expression a person's estimate of their capabilities. The key to measure development then is to pilot the measures with indigenous experts and to obtain pilot data on the target population to conduct cross-cultural equivalency tests prior to large-scale data collection. These findings should be of interest to practitioners who are working with clients representing multiple ethnicities or cultural groups. Researchers examining domain specific self-efficacy and general self-efficacy in multi-cultural groups should also find this analysis useful in examining disparities in the reported self-efficacy among study participants from different cultural groups.