Stress and sleep remain significant predictors of health after controlling for negative affect.

Prior research has shown that the predictive power of a stress-health model can be improved by the addition of sleep as a predictor variable. However, negative affect (NA) may act as a confounding variable in stress-health relationships: a source of bias that potentially inflates the observed relationship between stress and health. The current study examines whether stress and sleep remain significant predictors of health after controlling for trait NA. An online survey containing well-established measures of stress, sleep, health, and NA was completed by 460 undergraduates. Hierarchical regression analyses revealed that stress and sleep remained significant predictors of self-reported health after controlling for the effects of NA. Our findings support the notion that stress-health models can be improved by the integration of sleep as an additional predictor variable and demonstrate that the stress/sleep-health relationships are not fully explained by dispositional NA. From a theoretical perspective, the inclusion of NA measures allows researchers to tease apart the unique contributions of predictor variables. This knowledge can subsequently be put to the test with greater confidence in the design of interventions that seek to improve health through stress reduction and improved sleep.