Many studies have established that married people fare better than their never-married counterparts in terms of psychological well-being. It is still unclear, however, whether this advantage is due primarily to beneficial effects of marriage or to the selection of psychologically healthier individuals into marriage. This study employs data on young adults from both waves of the National Survey of Families and Households to test hypotheses based on both selection (N= 878) and relationship (N= 722) effects. Further, we differentiate union formation into cohabitation and marriage with and without prior cohabitation. Results indicate no evidence of selection of less depressed persons into either marriage or cohabitation, but a negative effect of entry into marriage on depression, particularly when marriage was not preceded by cohabitation.