Weisberg, H. (1987). "The Demographics of a New Voting Gap: Marital Differences in American Voting," Public Opinion Quarterly 51(3): 335-343.
This piece has its genesis in an odd cross-tabulation from the 1984 presidential election: Reagan won 63% of the vote of married people, and only 51% of the vote of unmarried people. Given that Reagan nationwide won 59% of the vote, these differences are startling and gave the author pause. The author justifies his interest by pointing out that while the "marriage gap" is smaller than the partisan break based on race, income, and religion, it recovers its salience because the number of unmarried people exceeds the number of black, poor, or Jewish. The cleavage retains its salience even when cross-tabulated by gender, with Republicans performing better among married women than unmarried men. It also has salience for turnout. In explaining the growth of the unmarried voter population, the author advances: the extension of the franchise to 18- to 20-year-olds, the increasing incidence of divorce, and in increase in unmarried cohabitation. The author then uses a path analysis to disaggregate the question of whether demographics drives the marriage gap, finding that the collinearity between racial marriage rates and racial voting behavior renders the marriage gap "partially spurious". Once the author throws out African-Americans by noting that the marriage gap does not appear in their voting behavior, he narrows down to his real interest: the marriage gap in white voters. After controlling for family income, the marriage gap disappears, leading the author to conclude that the marriage gap is driven mostly by the better income prospects for a two-income family.
My Take: This appears to be the opposite conclusion reached by Gerson in this piece. Part of the explanation can be derived from the different methods: Weisberg does nothing but crunch the numbers on national surveys, while Gerson conducted interviews. However, that doesn't seem like enough; Weisberg is literally saying that working married women vote Republican, an outcome that Gerson claims is unlikely. In short, Weisberg's data doesn't tell a coherent story with Gerson's theory, which means that we need either a) more data or b) different theories.