Skip to contents

Scored responses (1 correct, 0 incorrect) of 1,000 simulated Grade 6 students to 60 multiple-choice items, generated with generate_isomorphs() from six item models of 10 isomorphs each. The items, their options and the values they were generated from are in aigfield_items.

Usage

aigfield

Format

A data frame with 1,000 rows (one per student) and 60 columns named by item id, PER-01 to TRI-10.

Source

Simulated; the script is data-raw/aigfield.R in the package's GitHub repository.

Details

The dataset is for learning to calibrate item families from automatic item generation: Checking that the isomorphs of a model are equally difficult, so that the family mean can stand in for field testing new items. Some problems were planted on purpose; lab_truth("aigfield") lists them and the true item parameters.

Examples

dim(aigfield)
#> [1] 1000   60
# share correct by item model
round(tapply(colMeans(aigfield), sub("-.*", "", names(aigfield)), mean), 2)
#>  EQN  OOO  PCT  PER  SUB  TRI 
#> 0.37 0.44 0.31 0.71 0.58 0.44