fitStackGPCM: Up-dates association parameters of the GPCM by fitting model to stacked data

Description

Discrete choice model (conditional multinomial logistic regression) is fit to stacked data to up-date matrix of association parameters of the LMA that corresponds to the generalized partial credit model. This function is called from 'fit.gpcm', which is called from 'ple.lma'. It is unlikely that it would be run outside of these wrappers. It is only slightly different from 'fitStack' for nominal models.

Usage

fitStackGPCM(
  Master,
  item.log,
  phi.log,
  fstack,
  TraitByTrait,
  starting.sv,
  npersons,
  nitems,
  ncat,
  nless,
  ntraits,
  Maxnphi,
  pq.mat,
  LambdaNames,
  PhiNames
)

Arguments

Master

Master data set from which stacked data is created

item.log

Needed to get most recent values of scale values (item.log)

phi.log

History of estimates parameters from stacked regression

fstack

Forumla for stacked regression

TraitByTrait

inTraitAdj matrix

starting.sv

Fixed category scores

npersons

Number of persons

nitems

Number of items

ncat

Number of categories per item

nless

Number of unique lambdas and unique nus per item

ntraits

Number of latent traits

Maxnphi

Number of phi parameters to bet estimated (NULL for 1 dimensional)

pq.mat

Used to compute rest-scores and totals

LambdaNames

Needed for formula and data for up-dating phi (stacked regresson)

PhiNames

Null for 1D models

Value

Phi.mat Up-dated matrix of phi parameters

item.log of iterations for LogLike, Lambda and phi parameters