The edstan package for R provides convenience functions and pre-programmed Stan models related to item response theory (IRT). Its purpose is to make fitting common IRT models using Stan easy.
The following table lists the models packaged with edstan. Each of these may optionally included a latent regression of ability. The Stan code for these models is documented in a series of case studies, linked in the table.
| Model | Stan file |
|---|---|
| Rasch | rasch_latent_reg.stan |
| Partical credit | pcm_latent_reg.stan |
| Rating scale | rsm_latent_reg.stan |
| Two-parameter logistic | 2pl_latent_reg.stan |
| Generalized partial credit | gpcm_latent_reg.stan |
| Generalized rating scale | grsm_latent_reg.stan |
The next table lists the functions packaged with edstan.
| Function | Purpose |
|---|---|
irt_data() |
Prepares data for fitting |
labelled_integer() |
Create vector of consecutive integers |
irt_stan() |
Wrapper for Running MCMC |
print_irt_stan() |
Show table of output |
rhat_columns() |
Create plot of convergence statistics |
edstan relies on the rstan package, which should be installed first. See here for instructions on installing rstan.
edstan is available on CRAN and may be installed with
install.packages("edstan").
Alternatively, the development version may be installed directly from Github as follows.
# Install edstan development version of edstan
install.packages("devtools")
devtools::install_github("danielcfurr/edstan")The R code below is an example how prepare data, fit the Rasch model, and then view results. It uses an example dataset packaged with edstan.
# Load packages
library(rstan)
library(edstan)
# Make the data list
data_dich <- irt_data(y = aggression$dich,
ii = labelled_integer(aggression$description),
jj = aggression$person)
# Fit the Rasch model
fit_rasch <- irt_stan(data_dich, model = "rasch_latent_reg.stan",
iter = 200, chains = 4)
# View convergence statistics
rhat_columns(fit_rasch)
# View summary of parameter posteriors
print_irt_stan(fit_rasch, data_dich)
# Add a latent regression to the previous model
data_lr <- irt_data(y = aggression$dich,
ii = labelled_integer(aggression$description),
jj = aggression$person,
covariates = aggression[, c("male", "anger")],
formula = ~ 1 + male*aggression)
fit_lr <- irt_stan(data_lr, model = "rasch_latent_reg.stan",
iter = 200, chains = 4)