[apo]

That's a good study, but I'm not sure it disentangles income from education:

> Moreover, the “divorce gap” between college graduates and those with less education was larger in the NLSY79 cohort than it was for the 1950–1955 birth cohort. In the NLSY79 cohort, the divorce rate for first marriages is nearly 20 percentage points lower for those who have completed their bachelor’s degree compared with those who have completed high school, regardless of whether they have some college or not. The gap is even greater, approaching 30 percentage points, when comparing those with a college degree to those with less than a high school diploma. Just as with first marriages, college graduates were more likely to stay in a second marriage when compared with groups that have less education.

Couples that get married after degrees have waited longer to get married than couples that either drop out of high school or marry after high school. College grads tend to earn more.

So it seems any income difference must account for bachelor's degrees and the extra time before getting married that requires.