Abstract:
|
In panel surveys, respondents are asked to report the same information several times, e.g. for a different reference period or to update their previous answers. Survey designers must decide whether or not to provide respondents with their previously reported data (PRD). Research on household surveys suggests that the impact of giving prior information can have differential effects, depending on the stability, saliency and complexity of the question topics. However, there is not much known about the impact of PRD in establishment surveys. This study seeks to provide information to allow survey designers to make better decisions about the use of PRD in their survey. Building on prior research, we explore the impact of PRD on data quality in two ways. First, we compare data from the Occupational Employment Survey (OES) with Quarterly Census of Employment and Wages (QCEW) data that were used as PRD during the OES interviews. Second, we code interviews for the Agricultural Labor Survey to classify respondent reactions when PRD is used. This research will determine the impact of PRD on the response process and data quality, allowing survey designers to make informed decisions.
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.