. . . There are some who criticize the study which is understandable, but the amount of work the researchers poured into it is definitely commendable.
Sure, the ideal study would be to interview 800,000 pairs of divorced couple, or to follow-up a million couples over 3 decades, then perform a factor analysis and a multiple regression to find out what factors contribute to divorce, but we all know that that is unrealistic.
We need to be reasonable when determining what constitutes "scientific rigor". The Bradford Hill criteria (
https://en.wikipedia.org/wiki/Bradford_Hill_criteria) is a good guide and although the "strength" of this study is lacking, it makes up for its short comings in other criterias such as "plausibility", "coherence", and perhaps "biological (psychological in this case) gradient.
Therefore, I disagree with those who said that this study "proves nothing". I simply say it proves something, but we need more study to determine what that something is. Until then, this serves as a good guide.
What this study cannot answer however, is "does crossdressing increases the risk of divorce"? Crossdressing is not one of the mentioned horsemen, although it stands to reason that it could be turned into "contempt". More study is needed.