"... of the divorces studied, apparently regardless of demographic, there seems to be a statistically significant ..." Most generally, correlation is not causa

"... of the divorces studied, apparently regardless of demographic, there seems to be a statistically significant ..." Most generally, correlation is not causation. Maybe people who are more attracted to each other sexually form more asymmetric marriages or maybe people with stronger sex drives are more likely to follow gender stereotypes. Just taking a sample from a self-selected population doesn't even indicate anything about causation. What we need is a study where the researchers randomly assign men and women into marriages with partners chosen by the researchers where they will divide up roles according to gender stereotypes or divide up responsibilities evenly. Then the researchers will return a decade later to see how they're doing.