get.shw: Read ConQuest showfiles
In eatRest: eatRest

Description Usage Arguments Details Value Author(s)

Function reads Conquest showfiles and transforms them into a R list of data frames.

1 2	get.shw(file, dif.term = NULL, split.dif = TRUE, abs.dif.bound = 0.64, sig.dif.bound = 0.43)

`file`	Character string of the Conquest showfile to be read in.
`dif.term`	Optional: Character string. Name of the term considered to be DIF-term. Must match corresponding term in showfile.
`split.dif`	Logical: When TRUE, DIF-Parameter are only given for Reference group.
`abs.dif.bound`	When DIF-Parameter are evaluated, this specifies the critical value for absolute DIF.
`sig.dif.bound`	When DIF-Parameter are evaluated, this specifies the critical value for confidence interval DIF.

Funktion searches for 'TERM'-statements in Conquest showfile and reads the tables associated with. If one statement is specified to contain DIF analyses, absolute DIF value is computed 2*[group-specific parameter]. Confidence intervalls for 90, 95 and 99 percent are computed via the standard error of specific parameters. If both criteria - absolute DIF exceeds abs.dif.bound and the confidence intervall does not include sig.dif.bound, item is considered to have DIF.

A list of data frames, named by the 'TERM'-statements in Conquest showfile, plus an additional data frame named regression with regression coefficients when latent linear regression model was specified in Conquest analysis, plus an additional data frame named cov.structure with covariance and correlation matrix of latent dimensions. If uni-dimensional model is specified, the variance of the latent dimension is given instead. If one term was specified as DIF-statement, the corresponding data frame is augmented with additional columns for confidence intervals and indicators specifying significant DIF.

Each data frame corresponding to a 'TERM' statement contains following columns:

`item.nr`	Item number
`item`	Name of item
`ESTIMATE`	Estimated difficulty of item
`ERROR`	Standard error of estimated item difficulty
`outfit`	Item's 'Outfit'
`outfit.ci.lb`	Lower bound of the outfit confidence interval
`outfit.ci.ub`	Upper bound of the outfit confidence interval
`outfit.t`	T-value for outfit
`infit`	Items's 'Infit'
`infit.ci.lb`	Lower bound of the infit confidence interval
`infit.ci.ub`	Upper bound of the infit confidence interval
`infit.t`	T-value for infit
`abs.dif`	Only for DIF analysis. Absolute DIF, computed as 2*[group-specific parameter].
`ci.lb`	Lower bound confidence interval for specific significance level of 90, 95 or 99 percent.
`ci.ub`	Upper bound confidence interval for specific significance level of 90, 95 or 99 percent.
`sig`	Indicates whether the corresponding item matches both DIF criteria. See details.

When latent regression was specified, the last element of the returned list is a data frame with regression coefficients, corresponding to the number of dimensions and the number of regressors. Regressor names, regression coefficients and its standard errors are given for each dimension.

Rows represent the regressors, columns represent the latent dimension to which the regression is fitted.

Sebastian Weirich

eatRest documentation built on May 2, 2019, 6:25 p.m.