Fancy seeing you here…again: Uncovering individual‐level panel data in repeated cross‐sectional surveys

In: Public administration review: PAR, Band 83, Heft 6, S. 1761-1771

Abstract

AbstractMany theories in Public Administration and Public Management explicitly relate to changes over time in the attitudes, values, perceptions, and/or motivations of public‐sector employees. Examining such theories using (repeated) cross‐sectional datasets may lead to biased inferences and an inability to expose credible causal relationships. As developing individual‐level panel datasets is costly and time‐consuming, this article presents a method to make better use of existing surveys fielded repeatedly among the same respondent pool without individual identifiers. Specifically, it sets out an approach to create a system of unique identifiers using information about respondents' background characteristics available within the original data. The result is a panel dataset that allows tracking (a subset of) individual respondents across time. The article discusses issues of feasibility, credibility as well as ethical considerations. The methodology has further practical value by highlighting data characteristics that can help minimize identifiability of respondents while creating public‐release datasets.