drdrtest_em {DRDRtest}

The base function for testing a effect modifier with user specified nuisance functions

Description

This is the function for testing whether a discrete covariate is an effect modifier with user specified nuisance functions

Usage

drdrtest_em(
  y,
  a,
  l,
  class_label,
  arange,
  pifunc,
  mufunc,
  h = NULL,
  b = 1000,
  dist = "TwoPoint",
  pi.low = 0.01,
  a.grid.size = 401
)

Arguments

y

A vector containing the outcomes for each observation

a

A vector containing the treatment levels (dosage) for each observation

l

A data.frame containing the observations of covariates

class_label

A vector containing the class label (label for the effect modifier) for each observation.

arange

A vector of length 2 giving the lower bound and upper bound of treatment levels

pifunc

A user specifid function or wapper that takes treatment a as the first argument and covariates l as the second argument and return propensit scores

mufunc

A user specifid function or wapper that takes treatment a as the first argument and covariates l as the second argument and return outcome regression values

h

bandwidth to be used in kernel regression. If not specified, will by default use "rule of thumb" bandwidth selector

b

number of Bootstrap samples to be generated

dist

distibution used to generate residuals for Bootstrap samples. Currently only have two options, "TwoPoint" and "Rademachar"

pi.low

Lower bound to truncate propensity scores

a.grid.size

size of equally spaced grid points over arange to be generate for numerically evaluating the integral in test statistic

Value

A list containing

p.value:

P value of the test result

test.stat:

Value of the observed test statistic

Bootstrap.samples:

A vector containing test statistic values from Bootstrap samples

bandwidth:

Bandwidth used in kernel regression

Examples

d <- 4
n <- 200
sigma <- 0.5
delta <- 1
height <-1
arange <- c(0,5)
triangle <- function(a,height){
   y <- exp(-a^2/((1/2)^2))*height
   return(y)
}
mu.mod<-function(a,l,delta,height){
   mu <- as.numeric(l%*%c(0.2,0.2,0.3,-0.1*delta))+
         triangle(a-2.5,height)+a*(-0.1*l[,1]+0.1*delta*l[,4])
   return(mu)
}
l <- matrix(rnorm(n*d),ncol=d)
l[,4] <- ifelse(l[,4]>0,1,0)
colnames(l) <- paste("l",1:4,sep="")

logit.lambda <- as.numeric(l%*%c(0.1,0.1,-0.1,0))
lambda <- exp(logit.lambda)/(1+exp(logit.lambda))
a <- rbeta(n, shape1 = lambda, shape2 =1-lambda)*5

mu <- mu.mod(a,l,delta,height)
residual.list <- rnorm(n,mean=0,sd =sigma)
y <- mu+residual.list

class_label <- l[,4]

pifunc <- function(a,l){
   l <- as.matrix(l)
   logit.lambda <- as.numeric(l%*%c(0.1,0.1,-0.1,0))
   lambda <- exp(logit.lambda)/(1+exp(logit.lambda))
   return(pmin(dbeta(a/5,shape=lambda,shape2=1-lambda)/5,100))    
}

mufunc <- function(a,l){
   return(mu.mod(a,as.matrix(l),delta,height))
}

out <- drdrtest_em(y,a,l,class_label,arange,pifunc,mufunc)