35 lines
1.2 KiB
R
35 lines
1.2 KiB
R
% Generated by roxygen2: do not edit by hand
|
|
% Please edit documentation in R/select_weight.R
|
|
\name{select_weight}
|
|
\alias{select_weight}
|
|
\title{Compute confounding weights for the PCBSM.}
|
|
\usage{
|
|
select_weight(
|
|
df = NULL,
|
|
grp = NULL,
|
|
X = NULL,
|
|
instr = NULL,
|
|
type.res = "deviance",
|
|
std = F
|
|
)
|
|
}
|
|
\arguments{
|
|
\item{df}{data.frame containing the data}
|
|
|
|
\item{grp}{string containing the name of the column where the group membership variable is stored in df}
|
|
|
|
\item{X}{vector of strings containing the name of confounders to be included in the model}
|
|
|
|
\item{instr}{vector of strings containing the name of instrumental variables to be included in the model}
|
|
|
|
\item{type.res}{string containing the type of glm residuals to be used. Default is "deviance"}
|
|
|
|
\item{std}{boolean indicating whether residuals should be standardized. Default is TRUE}
|
|
}
|
|
\value{
|
|
A vector of weights to be included in a PCBSM
|
|
}
|
|
\description{
|
|
This function computes weights to be included in a PCBSM as a covariate accounting for unobserved confounding. Obtained by extracting response residuals from a probit model with grp as dependent variable and confounders and instruments as independent variables.
|
|
}
|