Multivariate Item Response Models in Questionnaire Research

Abstract:

Item response theory (IRT) is widely used in educational, psychological research to model how participants respond to test items in isolation and in bundles, however there is still existing gap in socio-economic area. Nowadays, due to availability of computer software, we are able to analyse large datasets and it is possible to use multivariate IRT models as well. When the items measure two or more latent traits, the test structure is referred to as multidimensional. Multidimensional item response theory (MIRT) provides an ideal foundation for modeling performance in complex domains, taking into account multiple basic abilities simultaneously, and representing different mixtures of the abilities required for different test items. In this paper we present the theory, application and the use of multivariate IRT models for the analysis of socio-economic phenomena. We also present model selection criteria and graphical presentation of results using Item Characteristic Curve (ICC) and Item Information Curve. All calculations will be conducted in R available from CRAN.