Optimization Targeting of the Annual Refiling Survey

• Activity Number: 517
• Type: Contributed
• Date/Time: Wednesday, August 6, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Government Statistics Section
• Author(s): Marek W. Kaminski*+
• Companies: Bureau of Labor Statistics
• Keywords: NAICS ; Quarterly Census of Employment and Wages ; Census ; Annual Refilling Survey ; Survey Optimization Techniques

Abstract

The Annual Refilling Survey (ARS) is a survey of US business establishments that is part of the Quarterly Census of Employment and Wages (QCEW) program. The purpose of the ARS survey is to ensure the correct North American Industrial Classification System (NAICS) is assigned to those establishments listed on the QCEW longitudinal data base. A complete census of the entire population of establishments takes 3 years. On an annual basis 1/3 of all QCEW establishments are surveyed, with the sample chosen from those establishments which did not participate in the prior 2 years. One measure of the quality of the refiling survey is the estimated percent of establishments misclassified by NAICS. We present an alternative survey technique based on operations research methodology which can potentially reduce the number of ARS surveyed establishments by up to 30% while significantly improving the quality of the survey with respect to the accuracy of industrial classification.