prod_qmat1: Produces weight matrix for special LLTM application (e.g....

In christinehohensinn/DataprepMisc: collection of independent functions for data prepration

Description

Function prod_qmat1 produces a weight matrix for the LLTM (e.g. estimable with package eRm) e.g. for the examination of position effects. Function is very restricted and assumes a data matrix with special colnames for producing the Q matrix. Warning: function does not proof Q for full rank!

Usage

prod_qmat1(virtdat, nr_itemstem, nr_pos)

Arguments

virtdat	data matrix with "virtual items" (e.g. resulting from function virtblock), data matrix has to be 0/1 matrix and colnames must have the form 'it30_pos4' that is the number of item and the number of position
nr_itemstem	number of item stems
nr_pos	number of different positions

Value

qmat weight matrix (warning: check for full rank!)

Author(s)

Christine Hohensinn

Examples

library(DataprepMisc)

it1 <- c(rep(1,5),rep(NA,7))
it2 <- c(rep(NA,5),rep(0,4), rep(NA,3))
it3 <- c(rep(NA,9),rep(0,3))
it4 <- c(rep(NA,5),rep(0,4), rep(NA,3))
it5 <- c(rep(1,5), rep(NA,7))
it6 <- c(rep(NA, 9), rep(1,3))
exqd <- cbind(it1,it2, it3, it4, it5, it6)
colnames(exqd) <- c("it1_pos1", "it1_pos2", "it1_pos3", "it2_pos1", "it2_pos2", "it2_pos3")

prod_qmat1(exqd, 2,3)

christinehohensinn/DataprepMisc documentation built on May 13, 2019, 7 p.m.