A methodological note on modeling the effects of race: the case of psychological distress

Psychological distress is an important indicator of the mental well‐being of the population. Findings regarding racial differences in distress are inconclusive but may represent an important pathway through which disparities exist across a number of physical health outcomes. We used data from the 1994 Minority Health Survey, a nationally representative multiracial/ethnic sample of adults in US households, to examine racial/ethnic differences in psychological distress (n = 3623). Our primary study aim was to examine differences between additive and multiplicative models in assessing the influence of income and gender on the race/distress relationship. We hypothesized that additive models do not sufficiently account for potential interactions of race with income and gender, and may therefore mask important differences in distress between racial groups. The results suggest that our hypotheses were supported. After adjusting for income, there were no statistically significant differences in distress levels between racial groups. However, significant differences emerge when multiplicative models are used demonstrating the complexities of the intersection of race, income and gender in predicting psychological distress. Black men and women of higher income status represent a particularly vulnerable group, whereas Hispanic men are especially hardy. We discuss the implications of our findings for future work on racial health disparities. Copyright © 2008 John Wiley & Sons, Ltd.