462 – SPEED: Survey Research Methods
Donor Selection for MCBS Income and Asset Imputation
Kanru Xia
NORC at the University of Chicago
Meimeizi Zhu
NORC at the University of Chicago
Edward Mulrow
NORC at the University of Chicago
Income and asset (IA) survey questions are known to have a high non-response rate. To mitigate this, surveys often allow respondents to report a value-range in lieu of an exact value. Still, most population based surveys implement item non-response imputation based on an appropriate respondent donor pool to enhance IA data quality. Prior research of elderly populations has suggested that those who provide value-range responses to asset questions have higher asset values than exact-value responders. Moreover, asset value-range respondents are more representative of item non-responders. We will examine whether IA exact-value respondents differ from value-ranges respondents in IA-related key demographic variables in the Medicare Current Beneficiary Survey (MCBS), a continuous, multipurpose survey of a nationally representative sample of the Medicare population that implemented a new IA questionnaire in 2015. In addition, we will compare IA item non-response imputation using a donor pool of all respondents versus only value-range respondents and see if restricting the donor pool would increase the imputed IA means. Outcomes will provide guidance for future MCBS IA imputation.