Despite the strong and persistent influence of Gary Becker’s marriage model, the model does not completely explain the observed correlation between married women’s labor market participation and overall divorce rates. In this paper we show how a simple sociologically inspired extension of the model realigns the model’s predictions with the observed trends. The extension builds on Becker’s own claim that partners match on preference for partner specialization, and, as a novelty, on additional sociological theory claiming that preference coordination tend to happen subconsciously. When we incorporate this aspect into Becker’s model, the model provides predictions of divorce rates and causes that fit more closely with empirical observations.