A woman's labor market participation and risk of divorce are argued to be important explanatory factors for the gender gap in political preferences. We utilize a Norwegian data set which allows a rigid test of these arguments because it includes information on vote choice, preferences regarding child and elder care spending, and extensive information on the relationship with the current partner. We find a gender gap in political preferences, but no evidence that it can be explained by women's risk of divorce, while the impact of labor market participation is not robust across specifications. To some extent, the gender gap in voting is driven by unmarried women voting left.