Estimation and Replicate Variance Estimation of Deciles for the Survey of Construction

• Activity Number: 665
• Type: Contributed
• Date/Time: Thursday, August 4, 2011 : 10:30 AM to 12:20 PM
• Sponsor: Section on Survey Research Methods
• Author(s): Stephen James Kaputa*+ and Katherine Jenny Thompson
• Companies: U.S. Census Bureau and U.S. Census Bureau
• Keywords: Median ; Modified half-sample ; Interpolation ; Deciles ; Replication ; Variance

Abstract

The U.S. Census Bureau publishes median estimates for several residential housing characteristics through the Survey of Construction (SOC), a national multistage probability sample that uses the modified half-sample (MHS) method of variance estimation. Currently, median estimates are produced using interpolation over data-dependent intervals (bins), after scaling by the 75th percentile. Thompson and Sigman (2000) demonstrated that this procedure has good statistical properties for medians computed from a highly skewed sample. This research explores decile estimation methods, with the constraints of maintaining reliable median estimates and using MHS replication. Using empirical data and simulated populations, we considered three different interpolation methods along with the traditional decile estimation method (no bins). Our simulation uses two finite populations to assess the statistical properties of each method via relative bias of the estimates, relative bias and stability of the variance estimates, and coverage rates. We found that a variant of the current procedure using the 95th percentile as a scaling factor produces decile estimates with the best statistical properties.