Let Me Tell You What You Told Me: Dependent Interviewing in Establishment Surveys

Heather Ridolfo and Jennifer Edgar

Abstract

In panel surveys, respondents are often asked to report the same information several times. In these cases, survey designers are faced with the choice of whether or not to provide respondents with their previously reported data (PRD). On one hand, PRD may reduce respondent burden and improve data quality by giving respondents an anchor on which to base their current reports. On the other hand, PRD may lead to respondent satisficing and lower quality data, by giving respondents the opportunity to simply confirm their previous answers. This is a particular problem as panel surveys are trying to measure change from one wave to another. Previous research on household survey respondents suggests that the impact of PRD can have differential effects, depending on the stability, saliency and complexity of the question topics. Currently there is not much known about the impact of PRD in establishment surveys. This study seeks to understand, through cognitive interviews and behavior coding, how establishment respondents incorporate PRD when formulating their survey responses.