The Reinterview Program for the BLS Compensation Surveys

Steven P. Paben


In order to ensure the quality of its data, BLS conducts reinterview programs for various compensation surveys, including the National Compensation Survey (NCS), its predecessor the Occupational Compensation Survey Program (OCSP), and the Employment Cost Index (ECI). This paper is an overview of findings in the reinterview programs over several years. In the reinterview programs, a random subsampling of collected establishments is taken. Formerly, within each subsampled establishment, the reviewer investigated only job data that were suspected of error. Currently, the reviewer investigates randomly selected job data. The paper examines this shift from reviewer-based to sampled-based data selection. Also discussed are:determining the reasons for reporting errors, determining the impact of reporting errors on published estimates, assessing the change in level of reporting errors over time, and the impact of prior knowledge by the original interviewer of the focus of the reinterview on the reinterview results.