SMABS 2004 Jena University
SMABS 2004 Home Organization About Jena Sponsors Links Imprint / Contact SMABS Home

European Association of Methodology

Department of methodology and evaluation research

Jena University

Contributions: Abstract

In this presentation (to be given by the first author) we discuss two extensions to the GLMM framework for IRT that accomodate both multidimensionality and latent item predictors in item response models. We introduce multidimensional random effects to account for situations where an item or subset of items measures more than one latent variable. We introduce what we call latent item predictors to allow items to have different discriminative power with respect to the latent variable(s) they are measuring. Starting with the Rasch Model, we present these extensions to the GLMM framework sequentially, and show that when both are incorporated, the resulting item response model is analogous to classical factor analysis for continuous dependent variables. The item response models are illustrated using self-report data on verbal aggression tendencies (Vansteelandt, 2000).

Vansteelandt, K. (2000). Formal models for contextualized personality psychology. Ph.D. thesis, K.U.Leuven, Belgium.