On the Accuracy of Retrospective Attitudinal Data

Abstract

Survey researchers usually assume that retrospective data become less accurate with the passage of time between event and interview. Available research focuses on retrospective behavioral data, although the findings may be assumed to generalize to retrospective attitudinal data. The present study tests the assumption that attitudes in the past are recalled less accurately over time and finds it unwarranted. Data from a national sample survey show that people who recall attitudinal information about a bureaucratic encounter which occurred 10 or more years in the past generally showed the same pattern of relationships among aspects of the encounter as did people who had contacted a government agency only two or three months prior to the interview.