U.S. flag

An official website of the United States government

Skip Header


The Impact of Imputation Procedures on Distributional Characteristics of Low Income Population

Written by:
Working Paper Number SEHSD-WP1987-08 or SIPP-WP-32

Abstract

The analysis of the impact of the imputation procedures on the distributional characteristics of the low income population is concentrated on the population receiving benefits under the Food Stamp Program as reflected in the Survey of Income and Program Participation (SIPP). This is a means-tested in kind transfer program for which benefits are calculated as an explicit function of selected characteristics including income, unit size, presence o f earnings and presence o f elderly or disabled members. The existence of this explicit relationship provides a point of comparison for the outcome of the imputation process.

The analysis first examines the impact of imputation on aggregate characteristics of households receiving benefits and then examines the impact on aggregate characteristics of the subset of the household covered by the benefits (i.e., the food stamp unit). Noting a small impact of the imputation process on these aggregate statistics, the discussion shifts to a closer examination of the population whose benefits or income are imputed. In so doing, we show that the imputation process does not preserve the known relationship between benefits and the determinants of benefit levels.

In addition to analyzing the impact of imputation of benefits or income, we demonstrate that the distributional characteristics of the food stamp population are different when the unit of analysis is changed from the household concept routinely employed by the Census Bureau to the food stamp unit. We further describe some problems encountered in the formation of food stamp units with SIPP.

Page Last Revised - January 11, 2022
Is this page helpful?
Thumbs Up Image Yes Thumbs Down Image No
NO THANKS
255 characters maximum 255 characters maximum reached
Thank you for your feedback.
Comments or suggestions?

Top

Back to Header