Conference Paper ISA - RC33 Research Committee on Logic and Methodology Conference
Claims versus evidence. An evaluation of dependent interviewing
01 Sep 2008
Dependent interviewing (DI) is a standardised questioning method particular to longitudinal surveys that utilises data gathered in previous interviews of a respondent to formulate question text. Although widely used, the performance of DI has rarely been evaluated. DI is expected to improve data quality, reduce response burden and facilitate the role of interviewers. However, evidence for each of these claims is inconclusive if not non-existent. This paper has four aims. First, we evaluate the impact of DI on interviewers and their survey role, such as the ease of administering DI questions, how DI affects interview flow, awkwardness of question wording, DI affect on interview lengths, etc.. Next, we analyse the impact of DI on respondents including their concerns about confidentiality and reaction to DI questions. We explore the impact of DI on interview dynamics and in particular on the rapport between interviewers and respondents. We will conclude by evaluating the implication of DI for longitudinal studies. To address our objectives, we examine a diverse set of qualitative and quantitative data gathered in a pilot for Wave 16 of the British Household Panel Study.