Attrition in Longitudinal Household Survey Data: Some Tests for Three Developing-Country Samples

42 Pages Posted: 20 Apr 2016

See all articles by Harold Alderman

Harold Alderman

International Food Policy Research Institute (IFPRI)

Jere Behrman

University of Pennsylvania - Department of Economics

Hans-Peter Kohler

University of Pennsylvania - Department of Sociology

John A. Maluccio

affiliation not provided to SSRN

Susan Cotts Watkins

affiliation not provided to SSRN

Date Written: September 2000

Abstract

Results from this study of the extent and implications of attrition for three longitudinal household surveys from Bolivia, Kenya, and South Africa suggest that multivariate estimates of behavioral relations may not be biased because of high attrition. This suggests that demographers and other social scientists can proceed with collecting longitudinal data to control for unobserved fixed factors and to capture dynamic relationships.

For capturing dynamic demographic relationships, longitudinal household data can have considerable advantages over more widely used cross-sectional data. But because the collection of longitudinal data may be difficult and expensive, analysts must assess the magnitudes of the problems specific to longitudinal but not to cross-sectional data.

One problem that concerns many analysts is that sample attrition may make the interpretation of estimates problematic. Such attrition may be especially severe where there is considerable migration between rural and urban areas. And attrition is likely to be selective on such characteristics as schooling, so high attrition is likely to bias estimates.

Alderman, Behrman, Kohler, Maluccio, and Watkins consider the extent and implications of attrition for three longitudinal household surveys from Bolivia, Kenya, and South Africa that report very high annual attrition rates between survey rounds.

Their estimates indicate that:

· The means for a number of critical outcome and family background variables differ significantly between those who are lost to follow-up and those who are re-interviewed.

· A number of family background variables are significant predictors of attrition.

· Nevertheless, the coefficient estimates for standard family background variables in regressions and probit equations for the majority of outcome variables in all three data sets are not significantly affected by attrition.

So attrition is apparently not a general problem for obtaining consistent estimates of the coefficients of interest for most of these outcomes. These results, which are very similar to those for industrial countries, suggest that multivariate estimates of behavioral relations may not be biased because of attrition. This would support the collection of longitudinal data.

This paper - a product of Rural Development, Development Research Group - is part of a larger effort in the group to evaluate the impact of Bank-funded investments. The study was funded by the Bank`s Research Support Budget under the research project Evaluation of the Impact of Investments in Early Child Development (RPO 682-34). The authors may be contacted at halderman@worldbank.org, jbehrman@econ.sas.upenn.edu, kohler@demogr.mpd.de, or swatkins@pop.upenn.edu.

Suggested Citation

Alderman, Harold and Behrman, Jere R. and Kohler, Hans-Peter and Maluccio, John A. and Cotts Watkins, Susan, Attrition in Longitudinal Household Survey Data: Some Tests for Three Developing-Country Samples (September 2000). Available at SSRN: https://ssrn.com/abstract=632518

Harold Alderman (Contact Author)

International Food Policy Research Institute (IFPRI) ( email )

1201 Eye St, NW,
Washington, DC 20005
United States

Jere R. Behrman

University of Pennsylvania - Department of Economics ( email )

Ronald O. Perelman Center for Political Science
133 South 36th Street
Philadelphia, PA 19104-6297
United States
215-898-7704 (Phone)
215-573-2057 (Fax)

Hans-Peter Kohler

University of Pennsylvania - Department of Sociology ( email )

3718 Locust Walk
Philadelphia, PA 19104-6297
United States
215-898-7686 (Phone)
215-898-2124 (Fax)

John A. Maluccio

affiliation not provided to SSRN

No Address Available

Susan Cotts Watkins

affiliation not provided to SSRN