Attrition bias in labor economics research using matched CPS files

Article type: Research Article

Authors: Neumark, David^{a; *} | Kawaguchi, Daiji^b

Affiliations: [a] Public Policy Institute of California and NBER, 500 Washington St., Suite 800, San Francisco, CA, 94111, USA. Tel.: +1 415 291 4476; Fax: +1 415 291 4428; E-mail: neumark@ppic.org | [b] Graduate School of Humanities and Social Sciences, University of Tsukuba, Tsukuba, Ibaraki 305-8573, Japan. E-mail: kawaguch@sk.tsukuba.ac.jp

Correspondence: [*] Corresponding author

Abstract: Short panel data sets constructed by matching individuals across monthly files of the Current Population Survey (CPS) have been used to study a wide range of questions in labor economics. But because the CPS does not follow movers, these panels exhibit significant attrition, which may lead to bias in longitudinal estimates. The Survey of Income and Program Participation (SIPP) uses essentially the same sampling frame and design as the CPS, but makes substantial efforts to follow movers. We therefore use the SIPP to construct "data-based" rather than "model-based" corrections for bias from selective attrition. The approach is applied to two questions that have been studied with CPS data – union wage differentials and the male marriage wage premium. The evidence suggests that in many applications the advantages of using matched CPS panels to obtain longitudinal estimates are likely to far outweigh the disadvantages from attrition biases, although we should allow for the possibility that attrition bias leads the longitudinal estimates to be understated.

DOI: 10.3233/JEM-2004-0236

Journal: Journal of Economic and Social Measurement, vol. 29, no. 4, pp. 445-472, 2004