Abstract
Hostility is thought by some researchers to be a correlate of depressive symptoms and by others to be a risk factor for depression. Previous studies of longitudinal trends in depression and hostility suggest that hostility confers risk for depression, but each such study used demographically restricted samples. In contrast, this study seeks to examine longitudinal trends in depression and hostility using a large sample of Americans exhibiting clinically relevant levels of depressive symptomatology. Data from the second and third waves of the National Survey of Families and Households (NSFH) were analyzed using structural equation modeling to explore longitudinal relationships between hostility and depression among American adults exhibiting elevated depressive symptomatology during at least one wave of interest (N = 2044). For the current sample, the hostility and depression measures used in the NSFH were found to exhibit measurement invariance both across the second and third waves of measurement and also across male and female groups of participants. Correlations between hostility and depression were significant and positive at each wave of measurement in all specified models, supporting the idea that hostility represents an intraepisodic feature of depression for many. Hostility levels at one time point, however, did not predict depressive symptoms at a later time point, failing to support the notion that hostility confers risk for the later development of depression. The present findings have implications for understanding heterogeneity in clinical presentations of depression and diagnostic criteria for depression. The current findings are also discussed with consideration of the cognitive theory of depression posited by Beck as well as the cognitive specificity perspective.