ROI: This function calculates the ratio of information (ROI) index...

In mustfa5/test.adaptation: Adaptation - level of adaptation of a computerized adaptive test

Description

This function calculates the ratio of information (ROI) index as described in Ju and Reckase (2019)

Usage

ROI(estimated.theta, items.administered, bank)

Arguments

estimated.theta	A data matrix that has the provisional ability estimates upon administering a sequence of items. In this input, rows are individuals taking the CAT. The columns are the items administered to individuals. The values in the cells are the provisional ability estimates after each item administration.
items.administered	A data matrix that has the set of item items administered to individuals. This input assumes that every row in the data frame corresponds to the set of item names/identifiers administered to an individual.
bank	A data matrix that have item parameters in the following order: discrimination, difficulty, guessing and slipping.

Value

This function returns a list with two outputs. (1) average ROI for the entire CAT administration and (2) ROI at individual level

References

Ju, U., & Reckase, M. D. (2019). New conditional measures of the amount of adaptation of adaptive tests. Paper presented at the annual meeting of the National Council on Measurement in Education, Toronto, Canada.

Examples

library(catR)
N=1000 #number of students
bank=250 #number of items
items=45
theta=rnorm(N,0,1) #level of trait
model="2PL" #IRT model to use
start <- list(theta = -1:1, randomesque = 1)
stop <- list(rule = c( "length"), thr = items)
final <- list(method = "ML")

test=list(method = "ML", itemSelect = "MFI")
bank=genDichoMatrix(items =bank, cbControl = NULL,
                    model = model)

res <- simulateRespondents(thetas = theta, bank,
                          start = start, test = test, stop = stop,
                          final = final, model = NULL)

estimated.theta=res$responses.df[,grepl("estimated.theta",names( res$responses.df ) ) ]
items.administered=res$responses.df[,grepl("items.administrated",names( res$responses.df ) ) ]
ROI(estimated.theta,items.administered, bank)

mustfa5/test.adaptation documentation built on Dec. 21, 2021, 11:03 p.m.