Administrative Data and Social Research: Coverage, Comparability, and the Limits of Measurement

Administrative data have become central to the social sciences and to the social indicators used to track populations, allocate resources, and evaluate policy. Generated by states and institutions in the course of administration, they offer scale, continuity, and near-complete population coverage, which reduces concerns about representativeness and allows individuals to be followed over time. In turbulent times, when indicators are used to diagnose crises and justify interventions, what is recorded, how it is recorded, and who or what remains missing become central to understanding governance and the production of social knowledge. Because these data are produced for administrative rather than research purposes, how well they measure social phenomena is an important and increasingly relevant question.

This session focuses on methodological and substantive problems that arise when administrative records are used to describe and explain social outcomes: measurement error, inconsistent definitions across time and space, changing classifications, coarse or one-dimensional categories, outdated records, selective coverage, linkage error, and overinterpretation of statistics shaped by policy rules and reporting practices. These problems limit comparability across cohorts, periods, institutions, and jurisdictions, and may leave some populations absent, misclassified, or only partly recorded.

We welcome empirical, methodological, theoretical, and comparative papers that examine these issues, particularly in research on labour markets, social assistance, healthcare, education, and taxation. Contributions may combine administrative data with surveys, qualitative research, or other sources. The session considers how these data can be used for research across the social sciences, and what they can and cannot show.