Abstract

Panel surveys measure how social and economic circumstances change over time. However, levels of change are often over-estimated in panel surveys: responses to a question are often not consistent across interviews, even if the respondent’s situation has not changed (e.g. Lynn and Sala 2006). Dependent interviewing (DI) has been pro-posed and used as a tool to reduce spurious change and increase the longitudinal consistency of responses. However simply reminding res-pondents of a previous response or adding an edit check does not necessarily solve problems with the underlying question (see, Jäckle 2008). In addition, researchers frequently voice concern that reminding respondents of answers they have given in previous interviews may encourage acquiescence (i.e., agreement with a statement regardless of its content), which may swap the problem of spurious change for the problem of spurious stability.
One of the aims this study is to evaluate the effects of different as-pects of question wording and format on the behavior of interviewers and respondents, and on the resulting measures of change. We will conduct a second wave with the respondents from the 2011 filter experiment, using their responses to the first wave to do dependent interviewing. The questionnaire was designed to support this follow up study. In addition, the administrative records available at IAB add a unique possibility to this research. We will test for panel effects by repeating some of the filter and follow up questions from the first wave. We will also test additional types of filter questions that we were unable to include in the first wave of the study due to time and budget constraints, such as the complex How Many/Go Again questions that are most relevant for IAB panel studies such as ALWA and PASS.