Year: 2003
Working paper number: 041
Author: Lee, Una
Unit: SSU

This paper will examines the problem of attrition in panel studies, and the different sets of methodological issues facing researchers in developed and developing countries. In particular, the study explores he hypothesis that attrition is contingent on certain specific community, household, and/or individual level characteristics, and examines options for reducing the impact of attrition through tracking. The study reviews the existing literature on panel data and attrition, examines the question of selectivity and attrition bias, and documents various techniques employed by researchers to reduce attrition.

Publication file: wp41.pdf