IMPS 2022 Proceedings

Overview

This package is for generating simulation data for FLPS model which is introduced in IMPS 2022 proceedings. More information can be found in Fully Latent Principal Stratification: combining PS with model-based measurement models.

Install

devtools::install_github("sooyongl/IMPS2022")

Generate Data

library(IMPS2022)
library(rstan)

sdat <- makeDat(
  N = 200,        # number of total sample (Half is for the treatment group)
  R2Y = 0.2,      # Proportion of outcome explained by covariates
  R2eta = 0.5,    # Proportion of latent variance explained by covariates
  omega = 0.5,    # Effect of latent variable on outcome
  tau0 = 0.3,     # Difference in outcome between Treatment and Control
  tau1 = -0.15,   # Principal effects
  lambda = 0.5,   # Missingness on item responses
  nsec = 20,      # Number of items
  lvmodel = '2pl' # Latent variable models
  )

Run Stan for FLPS

# You can load the raw stan script corresponding to the type of measurement models from the package
stan_model <- mk_stanmodel("2pl") # "2pl", "grm", "gpcm" available

# Run stan based on the generated data and the FLPS stan script
fit <- rstan::stan(model_code = stan_code, data = sdat$stan_dt)

summary(fit)