Explainable Double ML for Institutional Trust in Policy Evaluation

Chai & Zheng (2023) show double ML (DML) can outperform DiD but suffers from poor interpretability. This research integrates SHAP (SHapley Additive exPlanations) into DML frameworks to trace causal estimate variations to institutional factors (e.g., corruption levels, media freedom). When studies like Goriacko et al. (2024) and Dahabreh & Bibbins-Domingo (2024) report conflicting policy effects (e.g., COVID-19 treatments), our method decomposes discrepancies into data-driven vs. assumption-driven sources. Unlike Lourenço et al.’s (2024) scoping review—which notes underreporting in ML—we provide transparency for policymakers. This bridges Kreif & DiazOrdaz’s (2019) ML-tools with political science’s need for actionable insights.

References:

1. Machine learning in policy evaluation: new tools for causal inference. N. Kreif, Karla DiazOrdaz (2019). Oxford Research Encyclopedia of Economics and Finance.
2. Medication use evaluation of tocilizumab implementation in COVID-19 treatment guidelines: A causal inference approach.. Pavel Goriacko, Ari Moskowitz, Nadia Ferguson, Saira Khalique, Una Hopkins, Nicholas Quinn, M. Sinnett, Eran Y Bellin (2024). American Journal of Health-System Pharmacy.
3. Machine Learning Algorithms to Estimate Propensity Scores in Health Policy Evaluation: A Scoping Review. Luís Lourenço, Luciano Weber, L. Garcia, Vinicius Ramos, João Souza (2024). International Journal of Environmental Research and Public Health.
4. Causal Inference About the Effects of Interventions From Observational Studies in Medical Journals.. Issa J. Dahabreh, Kirsten Bibbins-Domingo (2024). Journal of the American Medical Association (JAMA).
5. The Explaimability of Double Machine Learning Causal Inference in Quasi-Natural Experiments—A Study Based on County Panel Sample Data. Zongxuan Chai, Tingting Zheng (2023). Automated Machine Learning.
6. Machine learning in policy evaluation: new tools for causal inference. N. Kreif, Karla DiazOrdaz (2019). Oxford Research Encyclopedia of Economics and Finance.