An Application of Calibration Approach in Weight Trimming for Stratum Jumpers

• Activity Number: 568
• Type: Topic Contributed
• Date/Time: Wednesday, August 4, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Section on Survey Research Methods
• Author(s): Yan K. Liu*+ and Phillip S. Kott and Lance L. Harris
• Companies: IRS and RTI International and IRS
• Keywords: Calibration ; Outlier ; Statistics of Income ; Stratum Jumper ; Weight Trimming

Abstract

The Statistics of Income division of the IRS started a panel sample of individual returns in tax year 1999. This panel sample is also used for cross-sectional estimations, with a small supplemental sample added to it each year. The base-year panel sample was a stratified sample, where stratum boundaries were formed using the return income. Because the income distribution is highly skewed, base weights vary dramatically. This poses a particular problem for returns whose income grows so dramatically that its associated base weight is no longer appropriate for cross-sectional estimations. In addition to stratum jumpers, high-income returns from both new filers and surviving filers in out-years are not well represented by weights based on selection probabilities. In this paper, we consider a calibration approach to adjusting return weights, including trimming for influential stratum jumpers.