I'm trying to do Inverse probability treatment weighting (IPWT) in R using two packages, ipw with the ipwpoint function to weghitit with the weghiti function. By passing the same parameters and data I obtain different results (e.g., different weights and consequently different standardized mean difference (SDM) values for the covariate). I want to understand why and on which these two methods are different.
weight_1 <- ipwpoint(exposure = treat, family = "binomial", link = "logit", numerator = ~ 1, denominator = ~ age_baseline + sex + hist_diabetes, data = cohort)
temp_weightit <- weightit(treat ~ age_baseline + sex + hist_diabetes, data = cohort, estimand = "ATE", method = "glm", stabilize = TRUE)