LordChi: Lord’s Chi-Square DIF Procedure

In cswells1/MeasInv: Collection of Methods to Detect Dichotomous and Polytomous Differential Item Functioning (DIF)

Description

Lord’s Chi-Square DIF Procedure

Usage

LordChi(
  par.Ref,
  par.Foc,
  cov.Ref,
  cov.Foc,
  fix.g = FALSE,
  same.scale = FALSE,
  purify = FALSE,
  sig.level = 0.05,
  delta.a = 0,
  delta.b = 0,
  delta.c = 0
)

Arguments

par.Ref	a data frame for the reference group including the columns of item ID, number of score categories, IRT model, a-parameter, b-parameter, and g-parameter
par.Foc	a data frame for the focal group including the columns of item ID, number of score categories, IRT model, a-parameter, b-parameter, and g-parameter
cov.Ref	a list of covariance matrices for each item for the reference group
cov.Foc	a list of covariance matrices for each item for the focal group
fix.g	logical: whether g is fixed or constrained to be equal between the groups (Default is FALSE.)
same.scale	logical: whether the item parameter estimates are on the same scale (Default is FALSE.)
purify	logical: purification for the anchor (Default is FALSE.)
sig.level	numeric: the significance level (Default is FALSE.)
delta.a	(Default is 0.)
delta.b	(Default is 0.)
delta.c	(Default is 0.)

Value

A list that contains the metadata information for the reference group ($par.est.Ref), the metadata information($par.est.Focal.t), and test statistics for the Lord’s chi-square($Lord.TS) and Wald tests ($Wald.TS) for the focal group with the transformed item parameter estimates. The data frame for Lord.TS contains for each item the item ID, number of score categories, and IRT model, as well as the Lord’s chi-square test statistic, degrees of freedom, its corresponding p-value, noncentrality parameter, and whether the item was flagged as DIF using the significance level supplied by the user.

cswells1/MeasInv documentation built on Dec. 19, 2021, 7 p.m.