A questionnaire table from a mirt model

Turns a unidimensional model fitted with mirt into the rows of a questionnaire csv (see read_questionnaire): the coefficients of each item (coef(fit, simplify = TRUE)), the mean and standard deviation of the latent trait, and the text and answer options of the items. Fit one model per scale, bind the tables of the scales with rbind() and save them with write_questionnaire.

Usage

questionnaire_from_mirt(
  fit,
  scale,
  question = NULL,
  options = NULL,
  key = NULL,
  reverse = FALSE,
  stem = NA,
  public = TRUE
)

Arguments

fit

A unidimensional model fitted with mirt::mirt(), with itemtype rsm, Rasch, graded, grsm, 2PL, 3PL, 4PL or nominal.

scale

The name of the scale.

question

The text of the items, in the order of the model; the item names when NULL.

options

The answer options: one character vector for all the items, or a list with one vector per item, in scoring order (the lowest category first). Numbered when NULL.

key

For dichotomous items given as several options: the number of the correct option of each item.

reverse

TRUE for items whose answers were reversed before the fit.

stem

An instruction shown above the items.

public

FALSE when the texts of the items are not public and question is only a placeholder: the CAT tab then highlights the correct option of keyed items.

Value

A data frame, one row per item.

See also

Other questionnaire: example_questionnaires(), irt_models(), read_questionnaire(), validate_questionnaire_against_mirt(), write_questionnaire()

Examples

if (FALSE) { # \dontrun{
fit <- mirt::mirt(responses, 1, itemtype = "graded")
q <- questionnaire_from_mirt(fit, "Anxiety",
  question = c("I worry a lot", "I feel tense"),
  options = c("Never", "Sometimes", "Often", "Always")
)
write_questionnaire(q, "anxiety.csv")
} # }