Causal Identification in Crime Research: Lessons From Structural Causal Models

This paper addresses identification challenges in social sciences using observational data through tailored examples from crime and criminal justice research. The paper highlights how structural causal models and directed acyclic graphs make the assumed causal structure explicit and help diagnose potential pitfalls such as collider bias and the use of inadmissible instruments or covariate sets. By distinguishing identification from estimation, the paper emphasizes the importance of specifying a causal model that makes identifying assumptions explicit and supports valid inference under those assumptions. The case studies demonstrate how structural causal models can sharpen identification strategies and improve the credibility of causal claims in applied crime research.