There is perhaps a survivorship bias in those studies, since divorce is so easy now, only people who are happy in marriage remain. i.e. reported happiness in marriage increases while everyone unhappy simply divorces and so the numbers would get swayed that way. Divorce rates have rocketed and women are very likely to leave marriages these days in many countries. So you end up with a tiny happy cohort that has no bearing on the wider population. So you should produce your sources to see how this is accounted for, ideally.