This paper investigates the relationship between the probability of divorce and marriage specific investments. As these investments in terms of childcare and household activities are likely to increase the marital surplus, they are consequently likely to decrease the risk of divorce. All such activities, however, are characterized by gender role bias through, for example, social norms. In periods in which married women enjoy greater outside options (e.g., by increasing their labor force participation), it is expected that households in which the husband takes on typically female chores are less likely to dissolve, while couples in which the wife takes on typically male chores are more likely to divorce. The paper tests this hypothesis using data from the National Longitudinal Survey (NLS) of Mature Women, the NLS Young Women, and the NLSY79. The prediction is strongly supported by the data with respect to older cohorts while it loses empirical relevance when tested on younger individuals. Furthermore, asymmetric effects between genders gain importance over time. Finally, an explanation for the relationship between divorce and marital investments is offered in terms of increasing intra-household time consumption complementarities. To this end, data from the American Time Use Surveys from 1965 to 2005 are studied to illustrate how time spent together by partners in the same household has become increasingly crucial in the American family.