Cohabiting couples’ economic circumstances and union transitions: a re-examination using multiple imputation techniques

This study examines how sample attrition and missing partner data influence studies of cohabitors’ union transitions. We rely on data from both waves of the National Survey of Families and Households (NSFH). Cohabitors with missing partner information or who were lost-to-follow-up have significantly fewer years of schooling and lower yearly earnings than cohabitors with complete data. Multiple imputation techniques are used to construct missing partner information and outcomes. Relying on prospective data from both waves of the NSFH that has not been corrected for selective data loss misrepresents cohabiting couple’s outcomes by: (1) overstating the extent to which cohabiting couples marry; (2) underestimating the amount of time spent in cohabiting unions; and (3) misrepresenting the effects of men’s economic characteristics on union transitions. Our results suggest that theories regarding how economic characteristics shape marriage transitions among singles may not be appropriate for those already living together.