This study seeks to quantify determinants, and costs, of the labor—force participation of married women. We use demographic and earnings data from the Health and Retirement Study. The earnings data constitute an unusually long panel but have the defect of lacking corresponding reports on work hours. By using a highly structured model and concentrating on the participation margin, we nevertheless feel that we can make substantial progress. Our preliminary regression results imply that married women’s market work disrupts their household consumption slightly less than one half as much as men’s work (relative to complete household retirement). We lay out a course of additional steps that can, we believe, clarify these results even more precisely in the near future.