R/MirtSubset.R

In CambridgeAssessmentResearch/unimirt: Applications of Unidimensional IRT using the R Package 'mirt'

#' Create a mirt object based on a subset of items in another one
#' 
#' This function takes a fitted unidimensional IRT model (fitted using mirt or unimirt) 
#' as an input, and creates a smaller one based on a particular subset of items.
#' This may be useful to enable the shiny apps to focus upon a particular subsection of items.
#' Note that the item parameters are NOT re-estimated. This function is simply intended
#' to allow users to zoom in on results for particular sets of items.
#' For example this might be useful to help produce a Wright Map based
#' on a subset of the items.
#' 
#' Note that although the item parameters are unchanged in this process.
#' Fit statistics (or anything else dependent upon producing 
#' ability estimates as an intermediate step) 
#' will be affected
#' by focussing on a subset of items.
#' 
#' @param mirtobj An estimated IRT model (of class SingleGroupClass) estimated either using \link[mirt]{mirt} or \link[unimirt]{unimirt}.
#' @param which.items an integer vector indicating which items to include in the reduced object. 
#' 
#' @return A estimated "mirt" object.
#' @examples
#' \dontrun{
#' mirt1=unimirt(mathsdata,"Rasch")
#' MirtTidyCoef(mirt1)
#' mirt2=MirtSubset(mirt1,c(1,3,5,10))
#' MirtTidyCoef(mirt2)
#' }
#' @export
MirtSubset=function (mirtobj, which.items) {
    dat1 = data.frame(mirtobj@Data$data[, which.items])
    #remove any rows from data that are now completely blank
    dat1=dat1[!is.na(rowSums(dat1,na.rm=TRUE)),]
    names(dat1) = colnames(mirtobj@Data$data)[which.items]
    names1 = colnames(dat1)
    itetypes = extract.mirt(mirtobj, "itemtype")[which.items]
    nextmirt2v = mirt(dat1, 1, pars = "values", itemtype = itetypes)
    rowids = paste(nextmirt2v$item, "_", nextmirt2v$name, sep = "")
    pars1 = mod2values(mirtobj)
    rowids1 = paste(pars1$item, "_", pars1$name, sep = "")
    for (rowid in rowids) {
        if (rowid %in% rowids1) {
            nextmirt2v$value[rowids == rowid] = pars1$value[rowids1 == 
                rowid]
        }
    }
    nextmirt2 = mirt(dat1, 1, pars = nextmirt2v, itemtype = itetypes, TOL = NaN)
    return(nextmirt2)
}