The effect you describe could be sufficiently studied by comparing married partners to unmarried long term partners.

Not the ridiculous idea of people who thought they were married but were apparently not due to some technicality or bureaucratic error which they were not aware of.