est.MAP: Estimate traits based on genuine likelihood

In susanne-frick/MFCblockInfo: compute block information for multidimensional forced-choice questionnaires

Estimate traits based on genuine likelihood

Description

Estimate traits based on genuine likelihood

Usage

est.MAP(
  FUN,
  responses,
  int,
  loads,
  uni,
  perms,
  SE = TRUE,
  lh.fun = lh,
  starts = NULL,
  box = 3,
  ...
)

Arguments

FUN	function to compute response probability
responses	matrix of block responses, rows = persons, columns = blocks. Responses should be given as indices for rank orders, corresponding to the columns in perms.
int	vector of pair intercepts (i.e., intercepts for binary outcomes of pairwise comparisons)
loads	matrix of item loadings, rows = items, columns = traits
uni	matrix of item uniquenesses (diagonal)
perms	matrix of permutations (i.e., rank orders). Can be obtained from calling permute()
SE	logical. Obtain standard errors from generalized inverse of the negative hessian at the log-likelihood? defaults to TRUE.
lh.fun	function to calculate likelihood across blocks. Defaults to lh.
starts	matrix of starting values for the latent traits, rows = persons, columns = traits. If NULL, all starting values are zero. Defaults to NULL.
box	numeric vector of length 1. Box constraints for the latent traits are set as $\pm$ box for all traits. Defaults to 3.
...	additional arguments passed to FUN.

Value

list with 5 entries: traits = matrix of point estimates for the latent traits, row = persons, columns = traits. ses = matrix of standard errors for the trait estimates, if SE = FALSE, all entries are NA. errors, warns, messages = vectors of any errors, warnings and messages that occured during estimation, in the order of their occurence,

susanne-frick/MFCblockInfo documentation built on Oct. 20, 2024, 8:26 p.m.