Imputation of Family Income and Maximal Utilization of Auxiliary Data: A Case Study of the 2012 Ohio Medicaid Assessment Survey (OMAS)

• Activity Number: 499
• Type: Contributed
• Date/Time: Wednesday, August 7, 2013 : 8:30 AM to 10:20 AM
• Sponsor: Survey Research Methods Section
• Author(s): Jamie Ridenhour*+ and Marcus Berzofsky and Caroline Blanton and G. Lance Couzens and Timothy Sahr and Bo Lu and Amy Ferketich
• Companies: RTI International and RTI International and RTI International and RTI International and Ohio Colleges of Medicine, Government Resource Center, Ohio State University and The Ohio State University and Ohio State University
• Keywords: dual-frame ; telephone survey ; imputation

Abstract

Imputation of survey data is widely employed, and practitioners may draw from a large body of methods for doing so. The methods chosen must satisfy the requirements of the data users while leveraging auxiliary data and accounting for the complexities of the collection instrument. Individual, household, and family income are measures commonly collected in surveys of households, and while these data often carry relatively high importance when compared to other variables, they are often also subject to high levels of nonresponse, and estimates of income can be quite sensitive to misspecification of the imputation model. This paper addresses the imputation of family income using percentile-constrained inverse-CDF, regression, and hot-deck techniques for the 2012 Ohio Medicaid Assessment Survey (OMAS). OMAS is a telephone survey of 22,929 households with a primary objective of estimating the number of Medicaid-eligible and uninsured persons in the State of Ohio. As Medicaid eligibility is determined by income, its imputation can impact policy decisions. The techniques are presented in the context of the larger, post-collection data activities, including imputation of related variables.