The Reliability of Retrospective Unemployment History Data

Abstract

This paper sets out to examine the reliability of data on men's and women's past experiences of unemployment spells which has been gained by asking individuals to recall these spells and the dates of their occurrence. Surveys seek individuals' history data either to gain information about the initial conditions of respondents, or to examine these histories in their own right. It is much less expensive to collect data about individuals' histories by asking them to recall their experiences than it is to collect the same data from a panel survey over many years. However, important questions arise concerning the reliability and validity of data collected by the recall method. Reviews of the tests carried out on recall data show that there are some important lessons to learn, particularly when individuals are being asked to recall experiences short in duration and low in importance (Dex 1995). Validity is difficult to assess. This paper seeks to extend the body of knowledge about the reliability of recall data on individuals' unemployment histories by providing empirical data on two large-scale surveys which collected similar data on unemployment histories from large numbers of British adults in the early 1990s; Wave 2 (1992) of the British Household Panel Study (BHPS) and the 1994 Family and Working Lives Survey (FWLS).