Data from a large epidemiologic survey were examined to determine the relationship of religious practice (worship service attendance), spiritual and religious self-perception, and importance (salience) to depressive symptoms. Data were obtained from 70,884 respondents older than 15 years from the Canadian National Population Health Survey (Wave II, 1996-1997). Logistic regression was used to examine the relationship of the religious/spiritual variables to depressive symptoms while controlling for demographic, social, and health variables. More frequent worship service attendees had significantly fewer depressive symptoms. In contrast, those who stated spiritual values or faith were important or perceived themselves to be spiritual/religious had higher levels of depressive symptoms, even after controlling for potential mediating and confounding factors. It is evident that spirituality/religion has an important effect on depressive symptoms, but this study underscores the complexity of this relationship. Longitudinal studies are needed to help elucidate mechanisms and the order and direction of effects.