Bootstrap Version of GVEM Confirmatory Analysis for M2PL

A bootstrap version of GVEM (i.e., GVEM-BS) can be implemented to correct the bias on item parameters and compute standard errors under confirmatory M2PL models

Usage

C2PL_bs(gvem_result, boots = 5)

Arguments

gvem_result: a list that includes exploratory or confirmatory GVEM results for M2PL models.
boots: the number of bootstrap samples; default is 5

Value

a list containing the following objects:

boots_a: item discrimination parameters corrected by bootstrap sampling, a \(J \times K\) matrix
boots_b: item difficulty parameters corrected by bootstrap sampling, a vector of length \(J\)
sd_a: stardard errors of item discrimination parameters, a \(J \times K\) matrix
sd_b: stardard errors of item difficulty parameters, a vector of length \(J\)

Author

Jiaying Xiao <jxiao6@uw.edu>

Examples

if (FALSE) { # \dontrun{
gvem_result <- with(C2PL_data, C2PL_gvem(data, model))
C2PL_bs(gvem_result, boots=10)} # }