The effect of income question design in health surveys on family income, poverty and eligibility estimates

Michael Davern, Holly Rodin, Timothy J. Beebe, Kathleen Thiede Call

Research output: Contribution to journalArticle

27 Scopus citations

Abstract

Objective. To compare systematic differences between an "omnibus" income measure that asks for total family income amounts with a central survey item and an aggregated income measure that sums specific amounts of income obtained from multiple income sources from everyone within a family. Data Source. The 2001 Current Population Survey-Demographic Supplement (CPS-DS). Data were collected from 78,000 households from February through April 2001. Study Design. First, we compare the omnibus family income to the aggregated family income amounts for each family. Second, we use the various aggregated family income sources to predict whether there is a mismatch between the omnibus and aggregated family income amounts. Finally, we assign a new aggregated amount of income that is restricted to be within the range of the omnibus amount to observe differences in poverty rates. Data Collection. Data were extracted from University of Michigan's ICPSR website. Principal Findings. There is a great deal of variation between the omnibus family income measure and the aggregated family income measure, with the omnibus amount generally being lower than the aggregated. As a result, the percent of people estimated to be in poverty is higher using the omnibus income item. Conclusions. Health surveys generally rely on an omnibus income measure and analysts should be aware that the income estimates derived from it are limited with respect to poverty determination, and the related concept of eligibility estimation. Analysts of health surveys should also consider matching respondents or multiple imputation to improve the usability of the data.

Original languageEnglish (US)
Pages (from-to)1534-1552
Number of pages19
JournalHealth Services Research
Volume40
Issue number5 I
DOIs
StatePublished - Oct 1 2005

    Fingerprint

Keywords

  • Eligibility
  • Imputation
  • Income measurement
  • Poverty
  • Survey design

ASJC Scopus subject areas

  • Health Policy

Cite this