Sample Attrition in the Presence of Population Attrition Seik Kim Department of Economics University of Washington seikkim@u.washington.edu http://faculty.washington.edu/seikkim/ November 18, 2012 Abstract This paper develops a method that accounts for non-ignorable sample attrition in the presence of population attrition for use with a non-representative panel sample. When there is population attrition, refreshment samples are not representative of the rst period population. Therefore, the existing sample attrition-correcting method developed by Hirano, Imbens, Ridder, and Rubin (2001) and Bhattacharya (2008) cannot be applied. This paper shows that the problem can be resolved by generating a counterfactual, but representative cross-section prior to applying their procedure. The proposed method is used to obtain attrition-correcting weights for the native and immigrant panel samples in the Current Population Survey. Keywords: Immigration, Population Attrition, Sample Attrition JEL Classication Codes: C23, C81, J61 0 I am grateful to my advisors, Joseph Altonji, Yuichi Kitamura, Fabian Lange, and Mark Rosenzweig. I have also beneted from helpful comments made by Donald Andrews, Debopam Bhattacharya, Keisuke Hirano, Taisuke Otsu, Peter Phillips, and seminar participants at Purdue University, Seoul National University, University of Carlifornia-Davis, University of Washington, Vanderbilt University, and Yale University. 1