Women living with their male partners are more likely to give birth to boys than are single women. The “marriage effect” (strictly speaking cohabiting effect) is small (51.5% boys in the cohabiting group versus 49.9% in the single group) but statistically significant. The marriage effect combined with an increase in the number of single mothers may be large enough to explain the decline in male births observed in many developed countries.
The cads versus dads theory had predicted the opposite result. Cads should have more sons than dads because big cads, by their very nature, have good male genes that are ideally passed on to little cads.
Earlier Steve noted that high status parents are more likely to have sons which makes evolutionary sense because succesful males are more likely to have children than unsuccesful ones while females are likely to have children regardless, so you want to specialize in boys when times are good and in girls when times are bad. The “marriage effect” could be an expectational consequence of the same logic. If women “know” that their male partners are unlikely to be around they know that they are more likely to have low-status in the future and thus should have girls.
I have not read the original research but it is easy to go wrong. Since girls cause divorce (and presumably also separation) it is critical to measure cohabitation before conception in order to distinguish “cohabiting causes boys” from “girls cause separation.” The research claims to measure cohabitation before “birth/conception” but retrospective bias could easily cause mothers to say that the father left before the birth (or before the child’s sex was known).