evalPbc: Evaluate discrimination statistics

In sachseka/eatPrep: Prepare Data for IRT Analyses

Evaluate discrimination statistics

Description

This function evaluates discrimination statistics for the categories of categorical variables, i.e. the correlations of each category with the total score on the test..

Usage

evalPbc(pbcs, mistypes = c("mnr", "mbd", "mir", "mbi"),
                    minPbcAtt = .05, maxPbcDis = .005, maxPbcMis = .07)

Arguments

pbcs	A data frame with the discrimination values for each category of categorical variables: output of catPbc. The data frame contains the following columns: item, cat, n, freq, freq.rel, catPbc, recodevalue, subunitType.
mistypes	Character or numeric vector. It will be checked, whether the missing types in this vector are defined for all items.
minPbcAtt	Numeric. Minimum correlation of attractors with total test score. If it is lower, the attractor will be flagged.
maxPbcDis	Numeric. Maximum correlation of distractors with total test score. If it is higher, the respective distractor will be flagged.
maxPbcMis	Numeric. Maximum correlation of missing with total test score. If it is higher, the respective missing type will be flagged.

Value

List. NULL (empty list) if everything is OK. Otherwise, returns list of character vectors containing the names of the flagged items:

zeroFreqAtt	Items with attractor frequency of zero.
zeroFreqDis	Items with distractor frequencies of zero.
lowMisPbcAtt	Items with too low (lower than minPbcAtt) or missing attractor correlations.
highPbcDis	Items with too high (higher than maxPbcDis) distractor correlations.
highPbcMis	Items with too high (higher than maxPbcMis) missing type correlations.

Author(s)

Karoline Sachse, Philipp Franikowski

Examples

data(inputDat)
data(inputList)

datRaw <- mergeData(newID = "ID", datList = inputDat, addMbd = TRUE)
datRec <- recodeData(datRaw, values = inputList$values,
    subunits=inputList$subunits)
pbcs   <- catPbc(datRaw, datRec, idRaw = "ID", idRec = "ID",
    context.vars = "hisei", values = inputList$values,
    subunits = inputList$subunits)
evalPbc(pbcs)

sachseka/eatPrep documentation built on June 9, 2025, 9:36 a.m.