I am working on estimating the ATT for a multiply imputed data set using Full Propensity scores Estimation. I got the total ATT, but cannot find the function to estimate the ATT for Sex (male, female), Language spoken (English, Spanish) to compare the performance for each group. The code is below:
Full Propensity scores Estimation
imp.data$ps <- imp.data$match.weight <- rep(0, nrow(imp.data))
for (i in unique(imp.data$.imp)) {
in.imp <- imp.data$.imp == i
imp.data$ps[in.imp] <- glm(Y ~ race + Ethnicity + Sex, data = imp.data[in.imp,], family="binomial")$fitted.values
m.out <- matchit(Y ~ race + fedEthnicity + Sex, data = imp.data[in.imp,], method = "full", ratio = 1, replace = TRUE)
I am working on estimating the ATT for a multiply imputed data set using Full Propensity scores Estimation. I got the total ATT, but cannot find the function to estimate the ATT for Sex (male, female), Language spoken (English, Spanish) to compare the performance for each group. The code is below:
Full Propensity scores Estimation
imp.data$ps <- imp.data$match.weight <- rep(0, nrow(imp.data)) for (i in unique(imp.data$.imp)) { in.imp <- imp.data$.imp == i imp.data$ps[in.imp] <- glm(Y ~ race + Ethnicity + Sex, data = imp.data[in.imp,], family="binomial")$fitted.values m.out <- matchit(Y ~ race + fedEthnicity + Sex, data = imp.data[in.imp,], method = "full", ratio = 1, replace = TRUE)
imp.data$match.weight[in.imp] <- m.out$weights
}
summary(m.out)
Define the multiply imputed data
m.data <- with(match.data(m.out), split(imp.data, imp.data$.imp)) View(m.data) mi.out <- to_zelig_mi(m.data$
1, m.data$2, m.data$3, m.data$4, m.data$5, m.data$6, m.data$7, m.data$8, m.data$9, m.data$10, m.data$11, m.data$12, m.data$13, m.data$14, m.data$15, m.data$16, m.data$17, m.data$18, m.data$19, m.data$20, m.data$21, m.data$22, m.data$23, m.data$24, m.data$25, m.data$26, m.data$27, m.data$28, m.data$29, m.data$30, m.data$31, m.data$32, m.data$33, m.data$34, m.data$35, m.data$36, m.data$37, m.data$38, m.data$39, m.data$40, m.data$41, m.data$42, m.data$43, m.data$44, m.data$45, m.data$46, m.data$47, m.data$48, m.data$49, m.data$50, m.data$51, m.data$52, m.data$53, m.data$54, m.data$55, m.data$56, m.data$57, m.data$58, m.data$59, m.data$60, m.data$61, m.data$62, m.data$63, m.data$64, m.data$65, m.data$66, m.data$67, m.data$68, m.data$69, m.data$70, m.data$71, m.data$72, m.data$73, m.data$74, m.data$75, m.data$76, m.data$77, m.data$78, m.data$79, m.data$80, m.data$81, m.data$82, m.data$83, m.data$84, m.data$85, m.data$86, m.data$87, m.data$88, m.data$89, m.data$90, m.data$91, m.data$92, m.data$93, m.data$94, m.data$95, m.data$96, m.data$97, m.data$98, m.data$99, m.data$100)ATT estimation
z.att0.mi <- zelig(SCORES~ Y, data = mi.out, model = "ls") ATT(z.att0.mi, treatment = "Y") qi0 <- get_qi(z.att0.mi, qi = "ATT", xvalue = "TE") combine_coef_se(z.att0.mi)
Appreciate your help!