Analysing home-ownership of couples: The effect of selecting couples at the time of the survey

Abstract

The analysis of events encountered by couple and family households may suffer from sample selection bias when data are restricted to couples existing at the moment of interview. The paper discusses the effect of sample selection bias on event history analyses of buying a home by comparing analyses performed on a sample of existing couples with analyses of a more complete sample including past as well as current partner relationships. The results show that, although home-buying in relationships that have ended differs clearly from behaviour in existing relationships, sample selection bias is not alarmingly large. © 1996 Kluwer Academic Publishers.