An item bank

An item bank holds calibrated items on one common scale, with their mirt parameters, a summary of their location and discrimination, where they came from and whether they are in use. Start a bank from a calibrated model or pipeline with mv_bank(), put a new calibration on the bank's scale with link_to_bank(), and add its new items with bank_items(). A bank is an ordinary R object: keep it between administrations with saveRDS() and readRDS().

Usage

mv_bank(x, name = "Item bank", source = NULL)

# S3 method for class 'mv_bank'
tidy(x, ...)

Arguments

x

a unidimensional model from mirt::mirt(), or a calibrated pipeline from calibrate() (if the pipeline was linked with link_to_bank(), its parameters are taken on the bank's scale).

name

a name for the bank.

source

a label for where the items came from, such as the administration. Defaults to the pipeline title, or "calibration".

...

not used.

Value

An object of class mv_bank: a list with name, items (a tibble with one row per item: item, itemtype, ncat, a (discrimination), b (location; the mean threshold for polytomous items), g (lower asymptote), source, added, status and the list column pars with the mirt slope-intercept parameters) and history.

Details

Supported item types are the unidimensional "Rasch", "2PL", "3PL", "graded" and "gpcm" models.

See also

bank_items(), link_to_bank()

Examples

# \donttest{
mod <- mirt(expand.table(LSAT7), 1, itemtype = "2PL", verbose = FALSE)
bank <- mv_bank(mod, name = "LSAT bank")
bank
#> <mv_bank> LSAT bank: 5 items (5 active)
#> Item types: 2PL 5 
#> Location (b) from -2.52 to -0.64; median discrimination 0.99
#> Sources: calibration 
# }