This article is a statistical catastrophe. Sample bias is just the most blatant of many problems.

She might have a different perspective if she'd interviewed a few (dozen) folks who married young and wrong, before they had matured enough to really know themselves or their spouse.