Fills the templates of one or more item models (item_model()) with
combinations of their values that meet the constraints, and returns the
generated items with their options in random order. Each item keeps the
values it was generated from, which are its features for calibration and
for predict_difficulty().
Arguments
- model
an item model, or a list of item models.
- n
number of isomorphs per model;
NULLgenerates every combination of values.- seed
random seed for the choice of values and the order of the options, or
NULL.- shuffle
logical; put the options in random order. Otherwise the key is option A.
- drop_flawed
logical; drop items whose options are not all different.
Value
A tibble with one row per item: id, model, stem, the
options A, B, ..., key (the letter of the correct option), one
column per variable, and flag. It can be passed to add_items(),
item_features() and review_packet().
Details
Combinations that give two identical options, or that cannot be computed,
would make flawed items: By default they are left out of the pool that
isomorphs are drawn from, with a message, so that the constraints can be
fixed. With drop_flawed = FALSE, combinations with identical options
stay in the pool and their items are marked in the flag column.
Examples
speed <- item_model(
stem = "A car travels {s * t} km in {t} hours. What is its average speed?",
key = "{s} km/h",
distractors = c(multiplied = "{s * t * t} km/h", subtracted = "{s * t - t} km/h",
distance = "{s * t} km/h"),
values = list(s = seq(40, 90, 10), t = 2:5), name = "SPD")
generate_isomorphs(speed, n = 4)
#> # A tibble: 4 × 11
#> id model stem A B C D key s t flag
#> <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr> <dbl> <int> <chr>
#> 1 SPD-01 SPD A car travels 80… 40 k… 78 k… 80 k… 160 … A 40 2 ""
#> 2 SPD-02 SPD A car travels 10… 200 … 98 k… 50 k… 100 … C 50 2 ""
#> 3 SPD-03 SPD A car travels 14… 70 k… 140 … 280 … 138 … A 70 2 ""
#> 4 SPD-04 SPD A car travels 12… 360 … 120 … 40 k… 117 … C 40 3 ""