The relationships between self-reported general health, employment, and attitudes toward the employment of married women have been analyzed for a representative sample of married, middle-aged women in the United States. The cross-sectional data indicate that women who were in the labor force had better health than women who were out of the labor force. In addition, women whose labor force status was compatible with their attitudes toward employment tended to have better health than women for whom there was a discrepancy between labor force status and attitudes. Analysis of the longitudinal data indicate that several causal mechanisms contributed tothe relationships observed in the crosssectional data. For the women with favorable attitudes toward employment, it appears that being a housewife had more detrimental effects on health than being employed. In contrast, for the women with unfavorable or neutral attitudes toward employment, it appears that employment status did not affect health. For healthy women, being employed may have contributed to more favorable attitudes toward employment. Healthy women were more likely than unhealthy women to stay in the labor force. Thus, it appears that there are multiple causal relationships linking employment status, attitudes toward employment and women's health.