Propensity Score Analysis in Observational Studies with Confounders and Missing Data

Author(s)	Ms. Margaret Batocael, Dr. Bernadette Tubo
Country	Philippines
Abstract	This study explores the use of propensity score matching to reduce bias in estimating treatment effects from observational data. Specifically, it evaluates the performance of logistic regression and machine learning-based methods for propensity score estimation under conditions involving missing data and complex confounding structures. Simulation studies were conducted using both complete and imputed datasets across varying levels of missingness, unmeasured confounding, and nonlinearity in the true propensity score. Logistic regression (LR), generalized boosting models (GBM), and Bayesian additive regression trees (BART) were compared based on estimation accuracy and covariate balance. Performance was assessed using root mean square error (RMSE) mean absolute error (MAE), R-squared, absolute standardized mean differences (ASMD), and Kolmogorov–Smirnov (KS) statistics. The results highlight trade-offs in model robustness, particularly between predictive accuracy and covariate balance, offering practical insights for selecting appropriate propensity score models in complex observational settings.
Keywords	Propensity Score, Observational Study, Missing Data, Unmeasured Confounding
Field	Mathematics > Statistics
Published In	Volume 7, Issue 2, March-April 2025
Published On	2025-04-30
DOI	https://doi.org/10.36948/ijfmr.2025.v07i02.43209
Short DOI	https://doi.org/g9g75s

About IJFMR Fees & Payment Current Issue Publication Archive	Submit Research Paper Track Submission Status Publication Guidelines Publication Ethics Peer Review & Plagiarism	Join as a Reviewer Editors & Reviewers Reviewer Referral Program Get Reviewer Membership Certi.	Website/Journal Policies Usage Policy Content Policies Privacy Policy

Contact Us		+91-9687-828-838	editor@ijfmr.com

International Journal For Multidisciplinary Research