Causal Estimands for Policy Evaluation and Beyond

2025/04/24 by Борис Соколов · 1 voice
Decision Sciences · #Evaluation and Performance Assessment

paper · pdf · doi:10.31235/osf.io/4vtpk_v1

openalex publication_date 2025/04/24 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/14

Abstract

Modern causal inference aims to answer substantive “what if” questions by presenting them as formally defined counterfactual target quantities, or estimands, that can be identified and estimated from data. Yet applied studies often under-specify these target quantities, leaving unclear what causal question is being answered and to what populations or settings the resulting conclusions can be generalized. I argue that this pattern stems partly from the tendency of the methodological literature to foreground identification, estimation, and software implementation while giving less systematic attention to the substantive meaning, scope, and limits of the estimands themselves. To address this issue, I review estimands used in popular research designs to operationalize causal inquiries within the Rubin Causal Model framework. I first introduce the most common average treatment effects, including ATE, ATT, and ATC. I then describe extensions of these estimands, including local and conditional treatment effects; causal interactions and mediation; effects for non-continuous outcomes and for multi-valued and continuous treatments; and longitudinal treatment effects. For each estimand, I provide a substantive explanation, along with examples of research questions it can address. I also outline the key assumptions necessary for identifying the main estimands covered in the paper.

Citations

• Causal Inference Using Potential Outcomes
• Parachute use to prevent death and major trauma related to gravitational challenge: systematic review of randomised controlled trials
• The Dangers of Extreme Counterfactuals
• Estimating causal effects of treatments in randomized and nonrandomized studies.
• The Balance‐Sample Size Frontier in Matching Methods for Causal Inference
• Parachute use to prevent death and major trauma when jumping from aircraft: randomized controlled trial
• Potential Outcome and Directed Acyclic Graph Approaches to Causality: Relevance for Empirical Practice in Economics
• Conjoint Survey Experiments
• Matching Methods for Confounder Adjustment: An Addition to the Epidemiologist’s Toolbox
• What Is Your Estimand? Defining the Target Quantity Connects Statistical Evidence to Theory
• A Practical Guide to Counterfactual Estimators for Causal Inference with Time‐Series Cross‐Sectional Data
• How much should we trust staggered difference-in-differences estimates?
• Recent Developments in Causal Inference and Machine Learning
• What’s trending in difference-in-differences? A synthesis of the recent econometrics literature
• Navigating Potential Pitfalls in Difference-in-Differences Designs: Reconciling Conflicting Findings on Mass Shootings’ Effect on Electoral Outcomes
• Comparing Experimental and Nonexperimental Methods: What Lessons Have We Learned Four Decades After LaLonde (1986)?
• Causal Inference for Qualitative Outcomes

Discussions

• Causal Estimands for Policy Evaluation and Beyond: https://doi.org/10.31235/osf.io/4vtpk_v1 [bsky, 0 points, 0 comments]