Online Program

Return to main conference page

All Times ET

Program is Subject to Change

Tuesday, June 15
Tue, Jun 15, 2:30 PM - 4:30 PM
TBD
Balancing Response Burden and Data Quality

Evaluating the Impact of Enrollment Nonresponse on an Establishment Panel Survey (308123)

*John Dixon, Bureau of Labor Statistics 
Josh Langeland, U.S. Bureau of Labor Statistics 

Keywords: Survey enrollment, nonresponse bias

The U.S. Bureau of Labor Statistics Current Employment Statistics (CES) is a panel survey that collects payroll information from businesses each month, which is used to produce employment estimates. When an establishment is first sampled, it enters an enrollment phase where an interviewer attempts to make contact with someone in the business to gain cooperation. After enrollment, respondents are re-contacted each month to provide updated data. Previous research has found a relationship between establishment characteristics (e.g., firm size) and CES nonresponse generally. In this paper however, we focus on the enrollment stage only to identify establishment characteristics that might be related to enrollment propensity. Furthermore, using administrative data (from the Quarterly Census of Employment and Wages), we examine the potential bias introduced to employment estimates when establishments fail to enroll.