Academic Journal
View in EDS

Choosing Covariate Balancing Methods for Causal Inference: Practical Insights From a Simulation Study.

Bibliographic Details
Title: Choosing Covariate Balancing Methods for Causal Inference: Practical Insights From a Simulation Study.
Authors: Peyrot E; Université Paris Cité and Université Sorbonne Paris Nord, Inserm, INRAE, Center for Research in Epidemiology and StatisticS (CRESS), Paris, France., Porcher R; Université Paris Cité and Université Sorbonne Paris Nord, Inserm, INRAE, Center for Research in Epidemiology and StatisticS (CRESS), Paris, France.; Centre d'Épidémiologie Clinique, Assistance Publique-Hôpitaux de Paris, Hôtel-Dieu, Paris, France., Petit F; Université Paris Cité and Université Sorbonne Paris Nord, Inserm, INRAE, Center for Research in Epidemiology and StatisticS (CRESS), Paris, France.
Source: Statistics in medicine [Stat Med] 2026 Jul; Vol. 45 (15-17), pp. e70672.
Publication Type: Journal Article
Language: English
Journal Info: Publisher: Wiley Country of Publication: England NLM ID: 8215016 Publication Model: Print Cited Medium: Internet ISSN: 1097-0258 (Electronic) Linking ISSN: 02776715 NLM ISO Abbreviation: Stat Med Subsets: MEDLINE
Imprint Name(s): Original Publication: Chichester ; New York : Wiley, c1982-
MeSH Terms: Observational Studies as Topic*/methods , Causality*, Monte Carlo Method ; Humans ; Computer Simulation ; Propensity Score ; Models, Statistical ; Confounding Factors, Epidemiologic ; Sample Size ; Least-Squares Analysis
Abstract: Background: Weighting methods are widely used for confounding adjustment in observational studies, but their finite-sample behavior depends on implementation choices and empirical overlap. We compare IPTW, just- and over-identified covariate balancing propensity score (CBPS), CBPS by tailored-loss function (CBPS-TLF), energy balancing (EB), and kernel optimal matching (KOM).
Methods: We conducted Monte Carlo simulations across 36 main scenarios varying sample size, treatment prevalence, and a complexity factor increasing confounding and reducing overlap. The main simulation considered a null constant treatment effect, with non-null constant effects examined as sensitivity analyses. Average treatment effects and average treatment effects on the treated were estimated using weighted least squares (WLS) and doubly robust (DR) estimators. Inference followed published recommendations when feasible. An empirical illustration used PROBITsim.
Results: Performance depended on the estimator and scenario complexity. Under WLS, IPTW and CBPS-TLF were more sensitive to complexity, while standard CBPS often behaved similarly to IPTW but with less deterioration in some high-prevalence settings. EB and KOM showed more stable point-estimation patterns across scenarios. DR estimation reduced differences between weighting methods when all confounders were included in the outcome model, although confidence-interval performance remained heterogeneous. PROBITsim results were consistent with simulation patterns.
Conclusions: The study should be read as practical guidance rather than a ranking of methods. It identifies settings where weighting analyses become sensitive to prevalence, overlap, tuning, and variance estimation. Confidence intervals that account for weight construction and tuning remain an important open practical issue.
(© 2026 The Author(s). Statistics in Medicine published by John Wiley & Sons Ltd.)
References: Stat Med. 2014 May 10;33(10):1685-99. (PMID: 24323618)
Stat Sci. 2007;22(4):569-573. (PMID: 18516239)
Stat Methods Med Res. 2019 Jan;28(1):3-19. (PMID: 28573919)
Stat Med. 2021 May 10;40(10):2305-2320. (PMID: 33665870)
NPJ Digit Med. 2023 Nov 27;6(1):221. (PMID: 38012221)
Value Health. 2010 Mar-Apr;13(2):273-7. (PMID: 19912596)
Am J Epidemiol. 2019 Jan 1;188(1):250-257. (PMID: 30189042)
Stat Med. 2015 Dec 10;34(28):3661-79. (PMID: 26238958)
Epidemiology. 2000 Sep;11(5):550-60. (PMID: 10955408)
Stat Med. 2026 Jul;45(15-17):e70672. (PMID: 42420803)
Stat Med. 2020 Dec 30;39(30):4922-4948. (PMID: 32964526)
J Clin Epidemiol. 2023 Oct;162:29-37. (PMID: 37562726)
Am Stat. 2009 May 1;63(2):155-162. (PMID: 22544972)
Am J Epidemiol. 2021 Aug 1;190(8):1659-1670. (PMID: 33615349)
Stat Med. 2004 Oct 15;23(19):2937-60. (PMID: 15351954)
PLoS One. 2011 Mar 31;6(3):e18174. (PMID: 21483818)
Grant Information: ANR-22-CPJ1-0047-01 Agence Nationale de la Recherche; ANR-23-IACL-0008 Agence Nationale de la Recherche; ANR-18-CE36-0010-01 Agence Nationale de la Recherche
Contributed Indexing: Keywords: Monte Carlo simulation; causal inference; inverse probability of treatment weighting; observational study; treatment effect estimation
Entry Date(s): Date Created: 20260709 Date Completed: 20260709 Latest Revision: 20260813
Update Code: 20260814
PubMed Central ID: PMC13346537
DOI: 10.1002/sim.70672
PMID: 42420803
Database: MEDLINE
Description
ISSN:1097-0258
DOI:10.1002/sim.70672