The Use of Imputation and Commercial Data to Improve the Efficiency of Income Stratified Sampling of Households with Young Children

• Activity Number: 227 - Non-Traditional Approaches for Sampling Rare Populations
• Type: Topic Contributed
• Date/Time: Monday, July 31, 2017 : 2:00 PM to 3:50 PM
• Sponsor: Survey Research Methods Section
• Author(s): Paul Burton* and Sunghee Lee and Trivellore Raghunathan
• Companies: University of Michigan and University of Michigan and University of Michigan School of Public Health

Abstract

A "cold call" or random screening to identify and sample a specific, rare, and/or a difficult to find population is prohibitively expensive. Thus, the survey organizations need to develop new screening/sampling strategies in order to keep field efficiency high and to minimize the associated field costs. This paper reports on a strategy that uses data from multiple sources (existing survey data, commercial data and census information), imputation and modeling strategies to predict probability of a household satisfying the eligibility criteria: (1) Households with a 3 to 10 year old and (2) income strata defined as low, medium and high based on the tertiles of the household income. A two-stage sampling strategy was developed to oversample in 3:2:1 ratio from the low, medium and high income ?strata of ?households and with children between the ages of 3 and 10. A pilot study was conducted to evaluate this strategy to estimate the actual eligibility rate. It is estimated that nearly 70% efficiency can be gained using this strategy (29% eligibility) when compared to random screening (20% eligibility).