Jonathan Eggleston, Economist at the U.S. Census Bureau
Friday, April 21st
398 New Cabell Hall
Does Encouraging Record Use for Financial Assets Improve Data Accuracy? Evidence from Administrative Data
Many surveys ask respondents to look at financial records in order to improve data accuracy. However, the assumption that record use reduces measurement error has not been tested yet with a comparison to administrative data. In this project, we compare interest income in the Survey of Income and Program Participation (SIPP) to administrative IRS 1040 tax data. We also examine whether a change in the record use question along with other improvements in the 2014 SIPP Panel are associated with increased record use and improved data accuracy. Our preliminary results show that record use is associated with an approximately 17 percent decrease in measurement error. We also find that the change in the record use question increased the number of respondents who used financial records by about 14 percentage points, but the associated increase in data accuracy was small. Finally, we discuss the implications these results have for the design of record use questions and whether these questions are beneficial in household surveys.