LNIRT: Log-normal response time IRT modelling

In LNIRT: LogNormal Response Time Item Response Theory Models

Description

Log-normal response time IRT modelling

Usage

LNIRT(
  RT,
  Y,
  data,
  XG = 1000,
  burnin = 10,
  XGresid = 1000,
  guess = FALSE,
  par1 = FALSE,
  residual = FALSE,
  td = TRUE,
  WL = FALSE,
  ident = 2,
  alpha,
  beta,
  phi,
  lambda,
  XPA = NULL,
  XPT = NULL,
  XIA = NULL,
  XIT = NULL,
  MBDY = NULL,
  MBDT = NULL
)

Arguments

RT	a Person-x-Item matrix of log-response times (time spent on solving an item).
Y	a Person-x-Item matrix of responses.
data	either a list or a simLNIRT object containing the response time and response matrices and optionally the predictors for the item and person parameters. If a simLNIRT object is provided, in the summary the simulated item and time parameters are shown alongside of the estimates. If the required variables cannot be found in the list, or if no data object is given, then the variables are taken from the environment from which LNIRT is called.
XG	the number of MCMC iterations to perform (default: 1000).
burnin	the percentage of MCMC iterations to discard as burn-in period (default: 10).
XGresid	the number of MCMC iterations to perform before residuals are computed (default: 1000).
guess	include guessing parameters in the IRT model (default: false).
par1	use alternative parameterization (default: false).
residual	compute residuals, >1000 iterations are recommended (default: false).
td	estimate the time-discrimination parameter(default: true).
WL	define the time-discrimination parameter as measurement error variance parameter (default: false).
ident	set identification rule (default: 2).
alpha	an optional vector of pre-defined item-discrimination parameters.
beta	an optional vector of pre-defined item-difficulty parameters.
phi	an optional vector of predefined time discrimination parameters.
lambda	an optional vector of predefined time intensity parameters.
XPA	an optional matrix of predictors for the person ability parameters.
XPT	an optional matrix of predictors for the person speed parameters.
XIA	an optional matrix of predictors for the item-difficulty parameters.
XIT	an optional matrix of predictors for the item-intensity parameters.
MBDY	an optional indicator matrix for response missings due to the test design (0: missing by design, 1: not missing by design).
MBDT	an optional indicator matrix for response time missings due to the test design (0: missing by design, 1: not missing by design).

Value

an object of class LNIRT.

Examples

## Not run: 
# Log-normal response time IRT modelling
data <- simLNIRT(N = 500, K = 20, rho = 0.8, WL = FALSE)
out <- LNIRT(RT = RT, Y = Y, data = data, XG = 1500, residual = TRUE, WL = FALSE)
summary(out) # Print results
out$Post.Means$Item.Difficulty # Extract posterior mean estimates

library(coda)
mcmc.object <- as.mcmc(out$MCMC.Samples$Item.Difficulty) # Extract MCMC samples for coda
summary(mcmc.object)
plot(mcmc.object)

## End(Not run)

LNIRT documentation built on Jan. 20, 2021, 1:05 a.m.