Consistency of Retrospective Cohabitation and Marital Histories

Charles Q. Strohm, University of California, Los Angeles

Retrospective data is one of the main sources of information on non-marital cohabitation, but the quality of these data has not been thoroughly evaluated. In this paper, I investigate the consistency of retrospective data on cohabitation and marriage using the 2000, 2003, and 2005 General Household Surveys (GHS), cross-sectional, probability samples of the British adults aged 16-59. I evaluate the extent of recall error by comparing rates of entry into cohabitation and marriage for the same age and period in each dataset. I examine differentials in recall error by socio-demographic factors and characteristics of the union such as duration. Finally, I investigate the implications of any recall error for substantive conclusions about the association between pre-marital cohabitation and subsequent marital disruption. Preliminary results show that cohabitations in the teenage and young adult years, as well as an individual’s first cohabitation are more likely to be underreported.