Abstract
Retrospective surveys are important for documenting vital events in historical time periods. The assumption used in these analyses is that the information reported at the time of the interview portrays a representative picture of past events. Using the Census Bureau's 1990 panel of the Survey of Income and Program Participation, this paper examines the potential problems encountered in deriving retrospective statistics when immigrants — who married or had children outside of the U.S. — constitute a significant proportion of survey respondents. The inclusion of these “immigrated” vital events in the survey resulted in (1) underestimating the mean age at first marriage for all Hispanics in the survey by 0.4 years, and (2) overestimating the premarital first birth ratio for Asians by 50 percent. Given likely future continued immigration from Latin America and Asia, these results suggest that future surveys should include migration histories in order to enable the survey to produce accurate retrospective statistics for these groups.
Get full access to this article
View all access options for this article.