test_LR: runs Anderson's likelihood ration test using the LRtest()...

In exhaustiveRasch: Item Selection and Exhaustive Search for Rasch Models

runs Anderson's likelihood ration test using the LRtest() function of eRm.

Description

runs Anderson's likelihood ration test using the LRtest() function of eRm.

Usage

test_LR(
  items = NULL,
  dset = NULL,
  na.rm = TRUE,
  model = NULL,
  p.par = NULL,
  modelType = NULL,
  splitcr = "median",
  alpha = 0.1,
  estimation_param = NULL
)

Arguments

items	a numeric vector containing the index numbers of the items in dset that are used to fit the model
dset	a data.frame containing the data
na.rm	a boolean value. If TRUE, all cases with any NA are removed (na.omit). If FALSE, only cases with full NA responses are removed
model	on object of a fit Rasch model, estimated with the packages 'eRm' (classes 'RM', 'PCM' or 'RSM'), 'psychotools' (classes raschmodel, 'pcmodel' or 'rsmodel') or 'pairwise' (class 'pers'), matching the value of modelType. If 'model' is provided, this model is used. If NULL, a model is fit using 'dset' and 'items'.
p.par	a person parameter object matching the class of 'model'. If NULL, the person parameters will be estimated.
modelType	a character value defining the rasch model to fit. Possible values: "RM", "PCM", "RSM".
splitcr	as defined by eRm::LRtest. Split criterion for subject raw score splitting. "all.r" corresponds to a full raw score split, "median" uses the median as split criterion, "mean" performs a mean split. Optionally splitcr can also be a vector which assigns each person to a certain subgroup (e.g., following an external criterion). This vector can be numeric, character or a factor. A random split, as in pairwise, is also a possible option.
alpha	a numeric value for the alpha level. Will be ignored if use.pval is FALSE
estimation_param	options for parameter estimation using estimation_control

Value

if the p-value of the test is not significant AND if no items were excluded in the test due to missing patterns, a list containing 3 elements is returned: the item combination that was tested, a list of the class the model was estimated with (depending on modelType and estimation_param$est) with the fit model and a list with a person parameter object (depending on estimation_param$est). If the test is significant, NULL is returned.

exhaustiveRasch documentation built on April 3, 2025, 6:18 p.m.