ImpactIV-package: Identifying Causal Effect for Multi-Component Intervention...

In ImpactIV: Identifying Causal Effect for Multi-Component Intervention Using Instrumental Variable Method

Description

In this package, you can find two functions proposed in Ding, Geng and Zhou (2011) to estimate direct and indirect causal effects with randomization and multiple-component intervention using instrumental variable method.

Details

Package:	ImpactIV
Type:	Package
Version:	1.0
Date:	2010-12-12
License:	GPL (>=2)
LazyLoad:	yes

Author(s)

Maintainer: Peng Ding <[e-mail]>

References

Ding, P., Geng, Z. and Zhou, X. H. (2011). Identifying Causal Effect for Multi-Component Intervention Using Instrumental Variable Method: with A Case Study of IMPACT Data. Technical Report.

See Also

homo_IV1, heter_IV2

Examples

data(impact)
Z=impact$Z
A=impact$A
M=impact$M
Y=scale(impact$Y)
X=as.matrix(impact[,5:12])
##continuos variables of X
Xcon = X[, c(1,4,6,8)]
##discrete variables of X
Xdis = X[, c(2,3,5,7)]
##X^2
X2 = cbind(X, poly(Xcon, degree = 2, raw = TRUE), 
           Xcon*Xdis[,1], Xcon*Xdis[,2], Xcon*Xdis[,3], Xcon*Xdis[,4])

method1 = homo_IV1(Z = Z,A = A,M = M,Y = Y,X = X)
method2 = heter_IV2(Z = Z,A = A,M = M,Y = Y,X = X2, 
                   polydegree = 1, step1 = method1, 
                   truncate = 0.25, select ="AIC")

ImpactIV documentation built on May 1, 2019, 8:04 p.m.