Using results from 2 large cardiovascular studies, the authors examined the utility of treating psychological response styles as confounds (e.g., factors undermining relationships with other self-report variables) versus distinct personality traits in the prediction of cardiovascular health. Study 1 consisted of a 3-year prospective study of ambulatory blood pressure levels in healthy adults (N = 125). Study 2 comprised a 12-week drug treatment program for ischemic heart disease patients (N = 95). Participants completed measures of psychological factors and self-deception and impression management in each study. Results consistently favored using response styles as direct predictors. Self-deception scores predicted elevated 3-year diastolic and systolic blood pressure changes in Study 1 and poorer treatment outcomes in Study 2. Statistically controlling for response style effects within the psychological factors generally did not improve predictions. These findings argue against the conceptualization of response styles as stylistic confounds.