A Power Analytic Investigation of Research in Adult Education: 1970-1982

The statistical analysis procedures used by adult education researchers have become more sophisticated over the last decade. However, critics have charged that researchers have failed to examine the technical adequacy of their work. In this study, the methodological soundness of adult education research was examined through a power analysis of the statistical tests used in volumes 21-32 of Adult Education. For each of the 1666 statistical tests reported, power estimates were obtained for three levels of hypothesized effect size. The results indicate that the power of statistical tests appearing in the adult education research literature is low when researchers are interested in finding small and medium effects. Many studies are being conducted with little chance of making a correct rejection of the null hypothesis. These findings are consistent with results obtained from similar analyses conducted in related fields, and point to ways in which researchers can increase the power of their statistical tests