ItemResponseFunctions.jl

ItemResponseFunctions.jl implements basic functions for Item Response Theory models. It is built based on the interface designed in AbstractItemResponseModels.jl.

You can install ItemResponseFunctions.jl from the General package registry:

] add ItemResponseFunctions

ItemResponseFunctions.jl exports the following functions for Item Response Theory models:

• irf: The item response function
• iif: The item information function
• expected_score: The expected score / test response function
• information: The test information function

Calling the function requires a model type M, a person ability theta and item parameters beta.
For a simple 1-Parameter Logistic model,

using ItemResponseFunctions

beta = (; b = 0.5)

irf(OnePL, 0.0, beta, 1)
iif(OnePL, 0.0, beta, 1)

evaluates the item response function and item information function for response y at ability value 0.0 for an item with difficulty 0.5.

Given an array of item parameters (a test) and an ability value, the test response function and test information can be calculated by

betas = [
    (; b = -0.3),
    (; b = 0.25),
    (; b = 1.0),
]

expected_score(OnePL, 0.0, betas)
information(OnePL, 0.0, betas)