A Bounding Approach to Age–Period–Cohort Analysis: A Demonstration Using Public Crime Concerns in Germany

Abstract Objectives In this paper, we present an application of a bounding approach to age–period–cohort (APC) analysis, recently developed by Fosse and Winship (2019), offering a transparent and assumption-aware alternative to commonly used identification strategies. Specifically, we apply the method to evaluate long-term trends in public concern about crime in Germany, addressing whether observed changes reflect genuine period effects. Methods For the demonstration, we use data from the German Socio-Economic Panel (SOEP) covering 1994–2020 focusing on the item measuring the concerns about crime in Germany. Results Using an intuitive and theoretically warranted assumption that crime concerns are unlikely to decrease from late adolescence into adulthood, we identify a declining period trend. However, additional constraints might be needed to narrow the solution space for age and cohort effects. Conclusions A bounding approach to APC analysis is easy to implement. The method allows extracting meaningful conclusions under transparent and testable assumptions, offering a promising alternative to models relying on hidden or arbitrary constraints.