Collecting Previously Reported Data: Testing Telephone Interviewing Techniques in the Occupational Employment Statistics Survey (303554)

Keywords: time series, questionnaire design, previously reported data, telephone, establishment survey, proactive interviewing, reactive interviewing, field-testing

The Occupational Employment Statistics (OES) survey is undergoing methodology changes so that the survey can transition from point-in-time to time-series survey. With this transition and the increase in the number of overlap units in each sample, there are questions about the best method of collecting new data from establishments that have previously submitted data to OES. For larger establishments, OES typically receives payroll records, which are considered to be true snapshots of employment and wages. However, twenty percent of the data for smaller establishments is collected over the telephone, which presents new collection challenges. Should OES consider a reactive collection strategy, where data are collected and then changes questioned during the interview? Or would a proactive strategy, where previously reported data are presented to the respondent prior to response, be more effective? How do we collect accurate data while avoiding satisficing? How do we ensure that the data collected are accurately reflective of changes in establishment employment and wages?

To answer these questions, the Bureau of Labor Statistics designed and conducted a three-group field test. The objectives of the test was to determine the best collection strategy and develop interview scripts. Each group received a specific collection strategy (control, reactive, or proactive) followed by a debriefing. The results were then compared against an electronic data file of their payroll records in order to measure accuracy. This study will describe the field test, discuss the debriefing results, and evaluate the accuracy comparisons.