ATE: summary.ATE – R documentation

Pricing

Become an expert in R — Interactive courses, Cheat Sheets, certificates and more!

Get Started for Free

Documentation

ATE

summary.ATE

Summarizing output of study.

Description

summary method for class "ATE"

Usage

## S3 method for class 'ATE'
summary(object, ...)

## S3 method for class 'summary.ATE'
print(x, ...)

Arguments

`object`	An object of class `"ATE"`, usually a result of a call to `ATE`.
`x`	An object of class `"summary.ATE"`, usually a result of a call to `summary.ATE`.
`...`	Further arguments passed to or from methods.

Details

print.summary.ATE prints a simplified output similar to print.summary.lm. The resulting table provides the point estimates, estimated standard errors, 95% Wald confidence intervals, the Z-statistic and the P-values for a Z-test.

Value

The function summary.ATE returns a list with the following components

`Estimate`	A matrix with point estimates along with standard errors, confidence intervals etc. This is the matrix users see with the `print.summary.RIPW` function.
`vcov`	The variance-covariance matrix of the point estimates.
`Conv`	The convergence result of the `object`.
`weights`	The weights for each subject in each treatment arm. These are same as the `weight` component of the `"RIPW" object`.
`call`	The `call` passed on as an argument of the function which is equivalent to `object$call`.

Author(s)

Asad Haris, Gary Chan

Examples

library(ATE)
#binary treatment

set.seed(25)
n <- 200
Z <- matrix(rnorm(4*n),ncol=4,nrow=n)
prop <- 1 / (1 + exp(Z[,1] - 0.5 * Z[,2] + 0.25*Z[,3] + 0.1 * Z[,4]))
treat <- rbinom(n, 1, prop)
Y <- 200 + 10*treat+ (1.5*treat-0.5)*(27.4*Z[,1] + 13.7*Z[,2] +
          13.7*Z[,3] + 13.7*Z[,4]) + rnorm(n)
X <- cbind(exp(Z[,1])/2,Z[,2]/(1+exp(Z[,1])),
          (Z[,1]*Z[,3]/25+0.6)^3,(Z[,2]+Z[,4]+20)^2)

#estimation of average treatment effects (ATE)
fit1<-ATE(Y,treat,X)
summary(fit1)
#plot(fit1)

#estimation of average treatment effects on treated (ATT)
fit2<-ATE(Y,treat,X,ATT=TRUE)
summary(fit2)
#plot(fit2)

#three treatment groups
set.seed(25)
n <- 200
Z <- matrix(rnorm(4*n),ncol=4,nrow=n)
prop1 <- 1 / (1 + exp(1+Z[,1] - 0.5 * Z[,2] + 0.25*Z[,3] + 0.1 * Z[,4]))
prop2 <- 1 / (1 + exp(Z[,1] - 0.5 * Z[,2] + 0.25*Z[,3] + 0.1 * Z[,4]))

U <-runif(n)
treat <- numeric(n)
treat[U>(1-prop2)]=2
treat[U<(1-prop2)& U>(prop2-prop1)]=1

Y <- 210 + 10*treat +(27.4*Z[,1] + 13.7*Z[,2] + 
            13.7*Z[,3] + 13.7*Z[,4]) + rnorm(n)
X <- cbind(exp(Z[,1])/2,Z[,2]/(1+exp(Z[,1])),
            (Z[,1]*Z[,3]/25+0.6)^3,(Z[,2]+Z[,4]+20)^2)

fit3<-ATE(Y,treat,X)
summary(fit3)
#plot(fit3)

ATE

Inference for Average Treatment Effects using Covariate Balancing

v0.2.0

GPL (>= 2)

Authors

Asad Haris <aharis@uw.edu> and Gary Chan <kcgchan@uw.edu>

Initial release

2015-02-16