Longitudinal Survey Weight Calibration: Theory and Methods

• Activity Number: 477
• Type: Contributed
• Date/Time: Wednesday, August 1, 2012 : 8:30 AM to 10:20 AM
• Sponsor: Section on Survey Research Methods
• Author(s): Siyu Qing*+ and Michael D. Larsen and Mary A Foulkes and Beilei Zhou
• Companies: The George Washington University and The George Washington University and The George Washington University and The George Washington University
• Keywords: Panel survey ; Calibration estimation ; Variance estimation ; Sample survey ; Cross-sectional study

Abstract

The National Science Foundation's Survey of Doctorate Recipients is conducted every two or three years and collects detailed information on individuals receiving PhDs in science and engineering in the U.S. and some others with PhDs from abroad in these areas. Survey weights adjust for oversampling and nonresponse on a cross sectional basis. A significant portion of the sample (e.g., 60% on 3 or more surveys from 1993-2006) appears in multiple survey years and can be linked across time. No longitudinal weight exists that would enable estimation of statistical models or comparison of finite population characteristics using data from multiple survey waves together. This paper develops theory and methods for calibration estimation to construct such a longitudinal weight. Methods for producing positive weights and for variance estimation are described. Issues including the use of multivariate calibration targets are discussed.