What are the best methods for estimating causal effects when experiments aren't possible? This review examines techniques for drawing causal inferences from observational data, a common challenge in social sciences where random assignment is often infeasible. The chapter explores the counterfactual framework, widely accepted for modeling causal effects. It then reviews both traditional and modern estimators applicable to cross-sectional data and introduces estimators that leverage longitudinal data's additional information. It covers the estimation of causal effects, instrumental variables, matching estimators, and propensity score methods. This review focuses on methods accessible to quantitatively oriented sociologists, offering a valuable resource for researchers seeking to estimate causal relationships in complex social phenomena. Understanding these methods is crucial for sound sociological research and effective social policy.
As a contribution to the Annual Review of Sociology, this paper aligns perfectly with the journal's mission to provide comprehensive overviews of significant topics in the field. By synthesizing the extensive literature on causal inference from observational data, the review equips sociologists with the methodological tools necessary to address complex research questions. This authoritative summary enhances the journal's value as a key resource for sociological scholars.