Bridging a Survey Redesign Using Multiple Imputation: An Application to the 2014 CPS ASEC

Abstract

Abstract The Current Population Survey Annual Social and Economic Supplement (CPS ASEC) serves as the data source for official income, poverty, and inequality statistics in the United States. In 2014, the CPS ASEC questionnaire was redesigned to improve data quality and to reduce misreporting, item nonresponse, and errors resulting from respondent fatigue. The sample was split into two groups, with nearly 70% receiving the traditional instrument and 30% receiving the redesigned instrument. Due to the relatively small redesign sample, analyses of changes in income and poverty between this and future years may lack sufficient power, especially for subgroups. The traditional sample is treated as if the responses were missing for income sources targeted by the redesign, and multiple imputation is used to generate plausible responses. A flexible imputation technique is used to place individuals into strata along two dimensions: 1) their probability of income recipiency and 2) their expected income conditional on recipiency for each income source. By matching on these two dimensions, this approach combines the ideas of propensity score matching and predictive means matching. In this article, this approach is implemented, the matching models are evaluated using diagnostics, and the results are analyzed.