This study compares the effects of a number of variables previously found to be important predictors of well-being such as income, health, and social interaction on one dimension of well-being, happiness, at pre- and postretirement for husbands and wives. In contrast to some of the research that has compared men and women, the present research found that husbands and wives have different predictors of happiness. A central finding of this research is that including spouse's characteristics (for example, health and happiness) increases the ability to predict the happiness of the respondent, suggesting that a dyadic approach has utility.