This paper investigates the hypothesis that individuals are less willing to marry when there are more potential partners to search amongst, and thus when others are also less prone to marry. To do this, it develops a reduced-form method [a variation on Manski's (1993) model] that allows identification of such spillovers in two-sided matching markets. Estimates from this method are internally consistent, unbiased, robust to different definitions of the marriage market, and large enough to warrant attention. Additional evidence suggests that the effect works via the proposed search mechanism.