In 2007, median household incomes of three groups -- married men, married women and unmarried women -- were about 60% higher than those of their counterparts in 1970. But for a fourth group, unmarried men, the rise in real median household income was smaller -- just 16%. (These household income figures are adjusted for household size and for inflation. For more details, see the methodology in Appendix B in the full report.)
Part of the reason for the superior gains of married adults is compositional in nature. Marriage rates have declined for all adults since 1970 and gone down most sharply for the least educated men and women. As a result, those with more education are far more likely than those with less education to be married, a gap that has widened since 1970. Because higher education tends to lead to higher earnings, these compositional changes have bolstered the economic gains from being married for both men and women.
If you are focused on the divergence between the bottom of the income distribution and the upper middle class, then I think these changes in marital patterns are a big part of it--including the fact that less-educated women are less likely to "marry up" than they were in 1970.
On the other hand, I think when people talk about inequality nowadays, they are mostly focused on the difference between, say, the 90th percentile of earners and the 99.9th percentile of earners. Many articulate critics of inequality are somewhere in between those percentiles. If given a choice between policies that compress the income distribution at the top end or policies that compress the income distribution between the bottom and the 90th percentile, which policies would these critics select?