transport_ittate: Transported intent-to-treat average treatment effect...

In kararudolph/transport: Transported Causal Inference With Targeted Learning

Description

More details to be added.

Usage

transport_ittate(a, z, y, site, w, weights = NULL, superlearner_lib = NULL,
  aamodel, asitemodel, s_awz_model, s_aw_model, aoutmodel, aq2model)

Arguments

a	needs to be named a and have values 0/1
z	matrix, each column be an individual z covariate and can be either binary or continuous (or a mix of both types) this code calculates the ratio dga1s0/dga1s1 with a trick in Zheng(2012)
y	needs to be named y and have values 0/1
site	The site, value 0 for the site where the outcome data is not used and value 1 for the site where the outcome data is used.
w	in a dataframe named w and with names w1:wx
weights	vector with length = nrow(data). p(Delta_2 | A, W, S)
superlearner_lib	(Optional) vector that defines the SuperLearner library.
aamodel	The aamodel
asitemodel	The asitemodel
s_awz_model	formula for fitting S ~ A + W + Z
s_aw_model	formula for fitting S ~ A + W
aoutmodel	The aoutmodel
aq2model	The aq 2 model

Details

This estimator is also used in transport_cace.

Value

A list with three elements:

1. est: the parameter estimate.

2. var: the variance estimate of the EIC.

3. eic: the efficient influnce curve unit-level values.

References

Zheng (2012)

Examples

# TBD.

kararudolph/transport documentation built on June 15, 2020, 11:42 p.m.