Gender Differences in Depressive Symptoms: Statistical Considerations

Publisher Website
Google Scholar
Abstract
Large community studies of depressive symptomatology provide mixed evidence concerning gender differences. The present paper investigates the effect of the high skewness typically present in distributions of depressive symptom scores on findings of a gender difference in depression. Because of this skewness, a few extreme scores among women (or in any subgroup) can produce a significant between-groups difference in untransformed scores even when the two groups' distributions are otherwise similar. Data from a community survey ( n = 1316) are consistent with this hypothesis. The data also suggest that gender increases depression scores indirectly, rather than directly, for the few women scoring in the extreme upper tail of the distribution. It is concluded that the gender difference in depressive symptoms lacks robustness, and that when it does occur, the effect of gender on depression can be understood as indirect, i.e., as mediated by gender differences in rate of employment, job status, education, and income.

This publication has 25 references indexed in Scilit: