Imputation of Income, Poverty, and Medicaid Status in the Ohio Family Health Survey

• Activity Number: 216
• Type: Contributed
• Date/Time: Monday, August 3, 2009 : 2:00 PM to 3:50 PM
• Sponsor: Section on Survey Research Methods
• Author(s): Ronaldo Iachan*+ and Bo Lu and Thomas Duffy
• Companies: Macro International, Inc. and The Ohio State University and Macro International, Inc.
• Keywords: Imputation ; Income ; Regression ; Hot deck

Abstract

The Ohio Family Health Survey (FHS) is a telephone survey of the health and health insurance status of adults and children in Ohio. The FHS prescribed confidence intervals for estimates of insurance status for several population subgroups: rural regions, ethnic minorities, families in poverty, etc. This paper presents and assesses the imputation methods developed for the survey with special attention to the imputation of income and Medicaid status, two inter-related variables with a substantial amount of missing data. Both Medicaid status and income were imputed using multivariate regression models for ordinal and binary outcome variables (Proc MI). The predictors included the following variables: Gender, Age, Education, Race/ethnicity, Tenure (Own vs. Rent), Employment status (full time versus not), Insurance status, Household size, and Marital status. We developed 5 models.