The tendency for a woman to ovulate two or more eggs at once, which may results in a heterozygotic multiple births is the way I understand it at least to some part genetic.
At least since the introduction of birth control on a large scale it is reasonable to think that women who give birth to twins the first time they give birth, over the course of their life on average give birth to more children than women who only give birth to one child the first time they give birth.
This is speculation on my part but should be easily checked given you live in a country where data of this kind is collected.
Now for twins there are probably some reduced likelyhood to reach adulthood without any health problems related to the fact that he is a twin (such as weighing less at birth). This might reduce the number of children people who are born as twins will have on average. Given good enough population statistics it should be possible to quantify this effect.
Do you know if there have been any scientific investigations of this subject?
I mean that from population statistics produced in many countries it should be an easy task to investigate to what extent women with a genetic tendency to have twins give birth to more healthy offspring during their lifetime than other women and that those genes therefore might be seen as "sweeping", becoming more common.