Abstract:
|
Most of the empirical research in economics and other social sciences rely on the use of household surveys, and are subject to how well do these surveys measure the topic we want to investigate. We study the causes of measurement error in program receipt in three household surveys that are widely used for empirical research in the USA: the American Community Survey, the Annual Social and Economic Survey of the Current Population Survey, and the Survey of Income and Program Participation. We link these surveys to administrative records for Food Stamps, Public Assistance, and Public Housing payments from the State of New York, to use as validation data. Our results show that error in the report of receipt is related to different household demographics, imputation of missing data, proxy interviewing, salience of welfare to the recipient household, recall period of the question, and stigma associated to government transfers. This paper contributes to the literature by being one of the first to study causes of measurement error in program receipt using different periods of recent data, as well as different variables across different surveys.
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.