It is known that families whose first child is a daughter are some 5% more likely to divorce than those whose first child is a son. As this Inside Science article notes, conventional wisdom has held that somehow daughters increase the likelihood of divorce, or that sons decrease it. The causal reasoning comes from the child preceding the divorce.
The article poses an interesting alternative hypothesis, namely that perhaps preexisting marital stress is more likely to cause male embryos to abort spontaneously. In other words, marriages already in trouble are more likely to produce female children, as female embryos are more likely to carry to term under suboptimal maternal conditions.
My mental image is of the causal arrow spinning 180 degrees to point in the opposite direction. Hat tip to RealClearScience for the link.
