Most current models in health psychology assume that stress adversely affects physical health. We re-examined this assumption by reviewing extensive data from the literature and from six samples of our own, in which we collected measures of personality, health and fitness, stress, and current emotional functioning. Results indicate that self-report health measures reflect a pervasive mood disposition of negative affectivity (NA); self-report stress scales also contain a substantial NA component. However, although NA is correlated with health compliant scales, it is not strongly or consistently related to actual, long-term health status, and thus will act as a general nuisance factor in health research. Because self-report measures of stress and health both contain a significant NA component, correlations between such measures likely overestimate the true association between stress and health. Results demonstrate the importance of including different types of health measures in health psychology research.