genItems function

genItems

Generates a list of item parameter matrices for use in function like conquestr::genResponses and conquestr::informationWrightMap

genItems(n, scores = NULL, deltadots, taus, discrims = 1, centre = NULL)

Arguments

• n: How many items?

• scores: When NULL it is assumed that all items have integer scoring, increasing for each category k, and beginning from 0. Otherwise a list where the elements are, in order:

  • a string naming a distribution function (for example runif, rnorm) to generate random deviates from (the scores).
  • a list of parameters to pass to the distribution function (for example, for runif, a list of length 2 defining "min" and "max"). This list is assumed to be in order to be directly passed into the function.
  • a boolean indicating whether the scores should be forced to be increasing across the response categories.
  • optionally a vector of item numbers to apply scores too. If not provided it is assumed that all items will be scored.

• deltadots: A list that describes the sampling distribution from which item location paramters are drawn. The elements of the list are, in order:

  • a string naming a distribution function (for example runif, rnorm) to generate random deviates from (the delta dots).
  • a list of parameters to pass to the distribution function (for example, for runif, a list of length 2 defining "min" and "max"). This list is assumed to be in order to be directly passed into the function. If the argument is missing, item location distribution is assumed to be ~U(-2, 2).

• taus: A list that describes the sampling distribution from which taus are drawn. Taus are deviations away from the average item location parameter. The elements of the list are, in order:

  • a string naming a distribution function (for example runif, rnorm) to generate random deviates from (the taus).
  • a list of parameters to pass to the distribution function (for example, for runif, a list of length 2 defining "min" and "max"). This list is assumed to be in order to be directly passed into the function.
  • a Boolean indicating whether the taus should be forced to be increasing across the response category boundaries (that is, no items exhibit disordered thresholds).
  • optionally, a vector that describes the number of response categories to apply to each of the n items being sampled. The length of this vector must be equal to n. For example if n=10, and the first 5 items are polytomous, then a vector of length 10, e.g., c(3, 3, 3, 4, 5, 2, 2, 2, 2, 2). In this example, the first three items have 3 categories each, the fourth item has 4 categories, the fifth item 5 categories, and the last five items are all dichotomies. When missing, a random vector is generated from sample(c(2:5), 10, replace = TRUE, prob = c(0.4, 0.3, 0.2, 0.1)) to create a mix of dichotomous and polytomous items. When the argument is missing, all items are assumed to be dichotomies.

• discrims: A list that describes the sampling distribution from which discrimination parameters are drawn.

  • a string naming a distribution function (for example runif, rnorm) to generate random deviates from (the discriminations).
  • a list of parameters to pass to the distribution function (for example, for runif, a list of length 2 defining "min" and "max"). This list is assumed to be in order to be directly passed into the function.
  • a Boolean indicating whether the discriminations are forced to be positive.
  • a Boolean indicating whether the discrimination is constant within item. When FALSE a unique discrimination is sampled for each category (that is, this can be one way of specifying the Bock Nominal model).
  • optionally a vector of item indices to sample a discrimination for. The length of this vector must be equal to or less than n. For example if n=10, and the user wants to sample discriminations for items 1, 5, and 10 then the vector is c(1, 5, 10). All other items will have a discrimination of 1. Otherwise it is assumed that all items have sampled discriminations. When missing all items are assumed to have constant discrimination equal to 1.

• centre: A number indicating the value to centre the generated values in deltadots on. Typically 0 for identification purposes. If NULL then values are left at their generated values (e.g., deviating from the expected mean proportional to sampling error).

Returns

A list of item matrices.

Examples

myItem <- matrix(c(0, 0, 0, 0, 1, 1, 0, 1), ncol = 4, byrow = TRUE)
  myItems <- list(myItem, myItem)
  myItems[[2]][2, 2] <- -1 # make the second item delta equal to -1
  myResponses <- genResponses(abilities = rnorm(100), itemParams = myItems)

See Also

simplef(), genResponses(), browseVignettes("conquestr")