The role of pearl’s causal framework in empirical research

Abstract

This paper underscores the necessity of formulating precise research questions that clarify causal relationships rather than simply identifying correlations and highlights the perils of relying solely on regression analysis in tackling complex causal inquiries without causal diagrams or structural causal models. It introduces Judea Pearl's causal epistemology, including causal graphs, structural causal models, and do-calculus as vital tools for estimating causal effects. It extends to the challenges of confounding and collider effects, the application of do-calculus with basic examples from Law & Economics and the advancements in causal discovery methods through constraint-based algorithms. The paper also offers a brief roadmap on best practices for identification and estimation.