Reports of relationship timing: Missing data and couple agreement

Abstract

In studying changes in family formation over time, social researchers and demographers have primarily relied on retrospective and prospective marital and cohabitation histories collected from surveys. With the increasing use of these types of retrospective questions in surveys, social researchers have now begun to pay more attention to the quality of the data and the degree of accuracy with which respondents are able to remember past events. The purpose of this paper is to explore the incidence and demographic and socio-economic correlates of recall error and inaccuracy of reporting of marriage and cohabitation dates. In the first part we investigate the degree of precision with which dates are remembered using both descriptive and multivariate analysis. We then compare married and cohabiting partner's reports about when their relationship started in order to check the consistency of with which both partners date the same event.