Optimal Test Design Approach to Fixed and Adaptive Test Construction

Founded on the optimal test design framework, the TestDesign package allows for constructing fixed test forms and simulating adaptive tests to the same test specifications with complex sets of constraints. This approach can render a variety of testing formats with different levels of adaptativeness and relative efficiency. Several item banks from real-world testing scenarios are provided as examples.


Install the latest release (1.2.2) from CRAN:


Quick start

itempool    <- loadItemPool(itempool_science_data)
itemattrib  <- loadItemAttrib(itemattrib_science_data, itempool)
constraints <- loadConstraints(constraints_science_data, itempool, itemattrib)

Static (fixed-form) assembly

cfg <- createStaticTestConfig(
  item_selection = list(
    method          = "TIF",
    target_location = c(-1,  1),
    target_value    = c( 8, 10)

solution <- Static(cfg, constraints)


Shadow (adaptive) assembly

cfg <- createShadowTestConfig()

solution <- Shadow(cfg, constraints, true_theta = c(0, 1))

plot(solution, type = "audit" , examinee_id = 1)
plot(solution, type = "shadow", examinee_id = 2, simple = T)

The documentation is available at (https://choi-phd.github.io/TestDesign)