Dependent interviewing in establishment surveys

Abstract

Dependent interviewing is a method in which respondents are presented with their previously reported data during data collection to aid response formation, reduce response burden and improve data quality. This method has been widely studied in longitudinal household surveys and found to be effective at reducing response error. Although some establishment surveys have adopted its use, its effect on response burden and data quality has not been empirically studied with such rigor. Responding to an establishment survey often entails reporting of proprietary information, use of records, self-administered forms, input from multiple individuals, and changes in respondents over time. These factors may impact the effectiveness of dependent interviewing on response burden and data quality. Given the differences in the response process between establishment and household survey respondents, more research is needed. The U.S. Energy Information Administration (EIA) is attempting to fill some of those gaps in the research. EIA uses dependent interviewing in several surveys but has never evaluated its use. Recently, EIA conducted usability testing and respondent debriefings with respondents to examine the effect of dependent interviewing on response burden and data quality. We present findings from this research and discuss the implications for establishment surveys and future research.