IRT.residuals.Rd
Defines an S3 method for the computation of observed residual values.
The computation of residuals is based on weighted likelihood estimates as
person parameters, see tam.wle
.
IRT.residuals
can only be applied for unidimensional IRT models.
The methods IRT.residuals
and residuals
are equivalent.
IRT.residuals(object, ...) # S3 method for tam.mml IRT.residuals(object, ...) # S3 method for tam.mml residuals(object, ...) # S3 method for tam.mml.2pl IRT.residuals(object, ...) # S3 method for tam.mml.2pl residuals(object, ...) # S3 method for tam.mml.mfr IRT.residuals(object, ...) # S3 method for tam.mml.mfr residuals(object, ...) # S3 method for tam.jml IRT.residuals(object, ...) # S3 method for tam.jml residuals(object, ...)
object | Object of class |
---|---|
... | Further arguments to be passed |
List with following entries
Residuals
Standardized residuals
Expected value of the item response \(X_{pi}\)
Variance of the item response \(X_{pi}\)
Used person parameter estimate
Expected item response probabilities
Residuals can be used to inspect local dependencies in the item response data, for example using principle component analysis or factor analysis (see Example 1).
See also the eRm::residuals
(eRm) or
residuals
(mirt)
functions.
See also predict.tam.mml
.
if (FALSE) { ############################################################################# # EXAMPLE 1: Residuals data.read ############################################################################# library(sirt) data(data.read, package="sirt") dat <- data.read # for Rasch model mod <- TAM::tam.mml( dat ) # extract residuals res <- TAM::IRT.residuals( mod ) str(res) }