Randall Parker pointed me to a study finding that those least likely to attend college based on family background, abilities, and peer group tend to gain the most from it relative to those more likely to do so based on the same considerations. After chewing on this, he asked a more general question: Is single motherhood dysgenic or eugenic? If women get knocked up by men out of their league for marriage, their children could benefit as a result.
I'm not aware of a way to measure this directly, but from previously sifting through data on cheating in the GSS, I know that more intelligent people are more likely to cheat than less intelligent people are. I wondered, then, if there was a substantial difference in the relationship between intelligence and cheating for men and women.
If smart men and dumb women tend to cheat, it's conceivable that the consequences are eugenic--dumb men end up raising children that are not theirs with their dumb, unfaithful wives, children who are instead the offspring of smarter cuckolds. And smart men end up not only having children with their smart wives but also other children with other women of lesser intelligence as well. Because men are able to procreate multiple times in a short period of time while women are not, smart men spreading their seed around and dumb women taking the seed of men smarter than they are is eugenic. In contrast, if the pizza delivery guy knocks up the rich, desperate housewife, the result is more dysgenic than if her husband would've done the inseminating.
This makes some enormous assumptions about a couple of aspects of cheating for which I have no evidence: 1) That cheaters do not take measures to separate their creeping fornication from procreation, and 2) That cheaters move up or down the intelligence ladder when they cheat, rather than tending to cheat with people similar to themselves (and their spouses). Further, it doesn't really get at Randall's question, since the GSS only queries about cheating among those who cheated while married--respondents who have never been married are excluded.
As it turns out, there isn't a big difference among men and women when it comes to intelligence and infidelity. The correlation between wordsum score and having cheated is .69 (p=.02) for men and .54 (p=.09) for women. Either both men and women are cheating with those similar to themselves, or both are descending down the intelligence ladder when they fool around. If everything assumed above is accurate, the eugenic effect is marginal at the most.
GSS variables used: EVSTRAY(1-2), SEX(1)(2), WORDSUM