Holder： LI FAN(Duke University)
Time：2025-04-18 15:00-16:00
Location：Conference Room 220, New Public Health Building, Peking University Health Science Center
Abstract:
This paper develops theoretically justified and assumption-lean analytical formulas for sample size and power calculation in the propensity score analysis of causal inference using observational data with treatment at a single time. By analyzing the variance of the inverse probability weighting estimator of the average treatmenteffect (ATE), we decompose power calculations into three components: propensity score distribution, potential outcome distribution, and their correlation. We show that, in addition to the standard inputs in the power calculation of randomized trials, two parameters, which quantify the strength of confounder-treatment and the confounder-outcome association, respectively, are necessary to determine the minimal sample size of an observational study. To quantify the confounder-treatment association, we proposeto use the Bhattacharyya coefficient, which measures the covariateoverlap and, together with the treatment proportion, leads to auniquely identifiable and easily computable propensity score distribution.To quantify the confounder-outcome association, we propose a sensitivity parameter bounded by the R-squared statisticof the regression of the outcome on covariates. The proposed method is applicable to both continuous and binary outcomes. We provide simulated and real examples to illustrate the proposed method.We develop an associated R package PSpower.
About the Speaker:
Fan Li is a professor in the Department of Statistical Science, with asecondary appointment in Biostatistics and Bioinformatics at Duke University. Her primary research interest is statistical methods forcausal inference and health data science. She also works on Bayesian analysis and missing data. She was the editor for Social Science, Biostatistics and Policy of the Annals of Applied Statistics.AE for JASA and AOS. She is an elected fellow of the American Statistical Association (ASA) and the Institute of Mathematical Statistics (lMS).
Your participation is warmly welcomed!
欢迎扫码关注北大统计科学中心公众号，了解更多讲座信息!
