The effect of survey attrition in longitudinal surveys: evidence from Peru, Coke d'Ivoire and Vietnam

The World Bank has been conducting household surveys in several developing countries. In several countries, follow-up surveys have been conducted in which the same dwellings that were visited in an earlier survey are visited again, and if the same people live there and agree to be interviewed, they yield longitudinal information. The surveys lose over time some of their samples (survey attrition). I test whether the same regressions apply to people who are interviewed repeatedly as to those who are interviewed once. I assess the effect of survey attrition in the Living Standards Measurement Surveys of Peru (1991-1994 and 1985/1986-1990), Cote d'Ivoire (1985-1988) and Vietnam (1992/1993-1997/1998). The preponderance of the evidence indicates that survey attrition does not have a major impact on the estimates of equations of schooling attainment, labor force participation, self-employment, wages and fertility. (C) 2002 Elsevier Science B.V. All rights reserved.