Variance Estimation and Selecting Good Estimands for Causal Effect Queries

Anna K Raichev, Rina Dechter, Jin Tian, Alexander Ihler
Proceedings of the 42nd Conference on Uncertainty in Artificial Intelligence, PMLR 337:5643-5660, 2026.

Abstract

This paper investigates the statistical efficiency of algebraic expressions ("estimands") used to answer causal queries. We first examine structural rules, developing a partial dominance relationship for front-door estimands that extends recent results on back-door estimands. In many models, however, structural rules alone may be insufficient to select an estimand. For such cases, we propose empirical techniques for estimating and comparing the variance of different estimands using both the graph and observational data: 1) a bootstrap-based method, and 2) a computationally simpler yet practically effective method for discrete models which estimates the {Fisher} information matrices. We illustrate both methods’ effectiveness on a variety of causal diagrams and estimand forms.

Cite this Paper

BibTeX
@InProceedings{pmlr-v337-raichev26a, title = {Variance Estimation and Selecting Good Estimands for Causal Effect Queries}, author = {Raichev, Anna K and Dechter, Rina and Tian, Jin and Ihler, Alexander}, booktitle = {Proceedings of the 42nd Conference on Uncertainty in Artificial Intelligence}, pages = {5643--5660}, year = {2026}, editor = {Perković, Emilija and Malinsky, Daniel}, volume = {337}, series = {Proceedings of Machine Learning Research}, month = {17--21 Aug}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v337/main/assets/raichev26a/raichev26a.pdf}, url = {https://proceedings.mlr.press/v337/raichev26a.html}, abstract = {This paper investigates the statistical efficiency of algebraic expressions ("estimands") used to answer causal queries. We first examine structural rules, developing a partial dominance relationship for front-door estimands that extends recent results on back-door estimands. In many models, however, structural rules alone may be insufficient to select an estimand. For such cases, we propose empirical techniques for estimating and comparing the variance of different estimands using both the graph and observational data: 1) a bootstrap-based method, and 2) a computationally simpler yet practically effective method for discrete models which estimates the {Fisher} information matrices. We illustrate both methods’ effectiveness on a variety of causal diagrams and estimand forms.} }
Endnote
%0 Conference Paper %T Variance Estimation and Selecting Good Estimands for Causal Effect Queries %A Anna K Raichev %A Rina Dechter %A Jin Tian %A Alexander Ihler %B Proceedings of the 42nd Conference on Uncertainty in Artificial Intelligence %C Proceedings of Machine Learning Research %D 2026 %E Emilija Perković %E Daniel Malinsky %F pmlr-v337-raichev26a %I PMLR %P 5643--5660 %U https://proceedings.mlr.press/v337/raichev26a.html %V 337 %X This paper investigates the statistical efficiency of algebraic expressions ("estimands") used to answer causal queries. We first examine structural rules, developing a partial dominance relationship for front-door estimands that extends recent results on back-door estimands. In many models, however, structural rules alone may be insufficient to select an estimand. For such cases, we propose empirical techniques for estimating and comparing the variance of different estimands using both the graph and observational data: 1) a bootstrap-based method, and 2) a computationally simpler yet practically effective method for discrete models which estimates the {Fisher} information matrices. We illustrate both methods’ effectiveness on a variety of causal diagrams and estimand forms.
APA
Raichev, A.K., Dechter, R., Tian, J. & Ihler, A.. (2026). Variance Estimation and Selecting Good Estimands for Causal Effect Queries. Proceedings of the 42nd Conference on Uncertainty in Artificial Intelligence, in Proceedings of Machine Learning Research 337:5643-5660 Available from https://proceedings.mlr.press/v337/raichev26a.html.

Related Material