DGM: Generating RCT data or observational data for the examples...

In RCTrep: Validation of Estimates of Treatment Effects in Observational Data

Generating RCT data or observational data for the examples used in the package

Description

Generating RCT data or observational data for the examples used in the package

Usage

DGM(
  trial,
  n,
  var_name,
  p_success,
  tau,
  y0,
  log.ps = NULL,
  binary = FALSE,
  noise = 1,
  ...
)

Arguments

trial	Logical indicating whether the treatment is randomly assigned in the generated data. If TRUE, RCT data is generated. Otherwise, observational data is generated.
n	A numeric value indicating the number of observations in the generated data
var_name	A character vector indicating the names of variables
p_success	the success probability of binary variables
tau	a character indicating the generation of the true treatment effect of each individual
y0	a character indicating the generation of the potential outcome under control
log.ps	a numeric value indicating the logit of propensity score
binary	logical indicating whether the outcome is binary or continuous variable
noise	a numeric value indicating the standard error of noise term of continuous outcome
...	an optional argument indicating pairwise correlations between variables

Value

a data frame; column names are variables names, z, y

Examples

n_rct <- 500; n_rwd <- 500
var_name <- c("x1","x2","x3","x4","x5","x6")
p_success_rct <- c(0.7,0.9,0.2,0.3,0.2,0.3)
p_success_rwd <- c(0.2,0.2,0.8,0.8,0.7,0.8)
tau <- "6*x2+x6+2"
y0 <- "x1"
log.ps <- "x1*x2+x3*x4+5*x5+x6"
rho1 <- c("x1","x2",0)
rho2 <- c("x2","x3",0)

target.data <- RCTrep::DGM(trial=TRUE, n_rct, var_name,
                           p_success_rct, tau, y0, log.ps=0,
                           binary = FALSE, noise=1, rho1, rho2)
source.data <- RCTrep::DGM(trial=FALSE, n_rwd, var_name,
                           p_success_rwd, tau, y0, log.ps,
                           binary = FALSE, noise=1, rho1, rho2)

RCTrep documentation built on Nov. 2, 2023, 6:10 p.m.