An Evaluation of BLS Noise Research for the Quarterly Census of Employment and Wages
Ye Yang, Ali Mushtaq, Santanu Pramanik, Fritz J. Scheuren, David Hiles, Michael Buso, Shail J. Butani, East-West Highway · 2011
The Quarterly Census of Employment and Wages (QCEW) program of the Bureau of Labor Statistics (BLS) publishes tabulations of monthly employment, quarterly wages, and number of establishments by industry and geography. In accordance with BLS policy, data provided to the Bureau in confidence should be used only for statistical purposes. In particular, the publication of data collected from BLS surveys should not lead to the identification of cooperating respondents. The BLS has been concerned about the current cell suppression method used with the QCEW. To address such concerns, BLS has conducted research about the random noise method as an alternative method. This paper provides an assessment of the BLS research to date. The paper begins with a review of the current cell suppression method in Section 2, focusing on the major disadvantages of the current method. Section 3 reviews the random noise model, its current application to QCEW at BLS, and the results to date. Section 4 provides an assessment of the properties of the BLS noise model under certain distributional assumptions of the noise factors. Finally, Section 5 provides some concluding remarks.