Dependent interviewing: effects on respondent burden and efficiency of data collection

Abstract

Dependent interviewing techniques, where substantive information from previous interviews is fed forward and used in the formulation of questions or to prompt post-response edit checks, are increasingly employed by panel surveys. While there is substantial evidence that dependent interviewing improves the quality of longitudinal data, claims of improved efficiency of data collection and reduced respondent burden are mostly anecdotal. This article uses data from a large experiment to systematically compare the effects of different question designs on efficiency and burden. The comparison highlights the wide variety of design options for dependent interviewing questions and their corresponding effects. In the present setup, efficiency gains were mainly due to reductions in coding costs for occupation and industry questions. The article concludes by identifying the conditions under which dependent interviewing offers the largest scope for efficiency gains and burden reduction.