LNIRT {LNIRT}

Log-normal response time IRT modelling

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)