Function to calculate test information
Given an estimated model compute the test information.
testinfo( x, Theta, degrees = NULL, group = NULL, individual = FALSE, which.items = 1:extract.mirt(x, "nitems") )
x |
an object of class 'SingleGroupClass', or an object of class 'MultipleGroupClass' if a suitable
|
Theta |
a matrix of latent trait values |
degrees |
a vector of angles in degrees that are between 0 and 90. Only applicable when the input object is multidimensional |
group |
group argument to pass to |
individual |
logical; return a data.frame of information traceline for each item? |
which.items |
an integer vector indicating which items to include in the expected information function. Default uses all possible items |
Phil Chalmers rphilip.chalmers@gmail.com
Chalmers, R., P. (2012). mirt: A Multidimensional Item Response Theory Package for the R Environment. Journal of Statistical Software, 48(6), 1-29. doi: 10.18637/jss.v048.i06
dat <- expand.table(deAyala) (mirt(dat, 1, '2PL', pars = 'values')) mod <- mirt(dat, 1, '2PL', constrain = list(c(1,5,9,13,17))) Theta <- matrix(seq(-4,4,.01)) tinfo <- testinfo(mod, Theta) plot(Theta, tinfo, type = 'l') ## Not run: #compare information loss between two tests tinfo_smaller <- testinfo(mod, Theta, which.items = 3:5) #removed item informations plot(Theta, iteminfo(extract.item(mod, 1), Theta), type = 'l') plot(Theta, iteminfo(extract.item(mod, 2), Theta), type = 'l') #most loss of info around -1 when removing items 1 and 2; expected given item info functions plot(Theta, tinfo_smaller - tinfo, type = 'l') ## End(Not run)
Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.