The other problem with the data is how were the divorces counted? Since there are so few states where same sex marriage is legal, and those marriages being considered legal are fairly recent, then the number of same sex couples really shouldn't do much to change the ratio of divorced men to divorced women. Also, if the survey only counted being divorced as a single tally and not counting someone who was divorced more than once as multiple tallies, then really, each state should pretty much have the same divorce rate men vs. women. So, what accounts for such wide swings? Yes, it could be that multiple divorces count more than once, but more likely, the survey results are flawed because it counts the divorces in the state where the respondents live in as opposed to counting the divorces in the state where they occurred. Which basically means that the divorce rate of men and women in each state is inaccurate, and therefore no real conclusions can be reached comparing the states' economic factors and divorce rate.
Ian