Lena Jost; Josef Brüderl; Katrin Auspurg

LMU Munich

Almost all sociological claims are not framed as general laws but rather as conditional claims that are valid in certain contexts but not in others. Moderation analysis is the analytical tool to examine such conditional claims. Thus, moderation analysis contributes to the development of theory and the design of targeted interventions. Given its relevance, moderation analyses should be designed and reported in a clear, meaningful and sound way.

However, recent methodological contributions have raised concerns about the adequacy of current research practices in moderation analyses. Reviews have demonstrated shortcomings in moderation analyses in political science and psychology. Methodological articles from these disciplines, as well as from epidemiology and statistics, have called for more careful consideration of conceptual as well as statistical aspects in moderation analyses.

This article aims to introduce the methodological debate to sociology. First, it synthesizes central claims from the methodological literature on estimand definition, identification, and estimation in moderation analysis. Sociological examples are used to build sociological intuition and to illustrate the challenges and potential solutions. Second, this paper provides a systematic review of moderation analyses recently published in leading sociological journals. This allows us to evaluate the extent to which the methodological recommendations have been implemented in substantive research, as well as to identify areas where further clarification and engagement in methodological discussions are needed.

Gerhard Krug

Institute for Employment Reserach (IAB)

Nonparametric propensity score methods are increasingly being used in social research to avoid misspecification bias in parametric methods such as linear regression. However, these methods can also be vulnerable to misspecification if the propensity score is estimated using parametric methods. Various methodological innovations have been developed to reduce or eliminate misspecification bias in propensity score methods, but they are underused in sociological research. This study conducts a comprehensive Monte Carlo simulation study to evaluate the performance of these innovations compared to that of standard methods. The results show that while some of the more recently developed extensions or alternatives to propensity score methods can substantially reduce misspecification bias, some are biased even in the absence of misspecification. In addition, most methods are subject to bias amplification due to “hidden dual misspecification”, a problem previously overlooked in methodological research. Among the estimators evaluated, entropy balancing was the most successful in both eliminating regular misspecification bias and reducing bias amplification. The covariate balancing propensity score and also augmented inverse probability weighting performed well. This study concludes that these estimators deserve more attention in applied social research.

Marc Hannappel¹; Sabine Zinn²

¹ University of Koblenz; ² Deutsches Institut für Wirtschaftsforschung

Computer simulations remain a marginal method in German sociology, with microsimulations in particular receiving little attention in mainstream methodological literature. In contrast, agent-based simulations have gained visibility — largely due to Hedström’s „Dissecting the Social“, which established their relevance within the framework of Analytical Sociology. Analytical Sociology seeks to explain social phenomena through mechanism-based models rather than universal laws. It emphasizes the actions and interactions of reflexive agents embedded in social contexts, and it critiques certain applications of traditional survey research for failing to capture the dynamics of social interaction and emergence. Agent-based simulations are central to this approach because they can model complex macro-level patterns as the result of simple, rule-based micro-level behaviours. However, whereas ABM often explores theoretical mechanisms through rule-based behaviour, microsimulations typically focus on empirically grounded, probabilistic modelling of individual-level transitions based on observed data. As such, microsimulations are more closely aligned with empirically oriented methodological frameworks. One such framework is Empirical-Analytical Sociology, which emphasizes theory-driven, transparent, and replicable empirical research. Microsimulation fits particularly well within this framework. It allows for the decomposition of complex phenomena into analytically tractable components (model structure), depends on a high-quality empirical database (simulation population and transition parameters), and relies on explicit theoretical assumptions to guide model design.This presentation will therefore not only introduce the theoretical foundations of microsimulation but also demonstrate how it can be applied to analyse interdependencies between biographical events. By bridging empirical data and formal modelling, microsimulations offer a promising method for understanding dynamic social processes.