PSpower: Sample Size and Power for Propensity Score Weighted Estimators
Computes sample size and power for causal inference studies that
use propensity score (PS) weighting. Supports continuous, binary, and
time-to-event (survival) outcomes under four estimands: average treatment
effect (ATE), average treatment effect on the treated (ATT), average
treatment effect on the controls (ATC), and average treatment effect on
the overlap population (ATO). For continuous and binary outcomes, the
asymptotic variance of the Hajek inverse probability weighting estimator
is derived under a logit-normal propensity score model, approximated by a
Beta distribution matched through the Bhattacharyya overlap coefficient.
For survival outcomes, the asymptotic variance of the propensity-score-
weighted partial likelihood estimator is used for randomized trials and
observational studies. The Schoenfeld formula is also available for
randomized trial settings.
Documentation:
Downloads:
Linking:
Please use the canonical form
https://CRAN.R-project.org/package=PSpower
to link to this page.