Using Projective IRT to Evaluate the Effects of Multidimensionality on Unidimensional IRT Model Parameters
Open Access
- 9 December 2024
- journal article
- research article
- Published by Taylor & Francis in Multivariate Behavioral Research
- Vol. 60 (2) , 345-361
Abstract
The application of unidimensional IRT models requires item response data to be unidimensional. Often, however, item response data contain a dominant dimension, as well as one or more nuisance dimensions caused by content clusters. Applying a unidimensional IRT model to multidimensional data causes violations of local independence, which can vitiate IRT applications. To evaluate and, possibly, remedy the problems caused by forcing unidimensional models onto multidimensional data, we consider the creation of a projected unidimensional IRT model, where the multidimensionality caused by nuisance dimensions is controlled for by integrating them out from the model. Specifically, when item response data have a bifactor structure, one can create a unidimensional model based on projecting to the general factor. Importantly, the projected unidimensional IRT model can be used as a benchmark for comparison to a unidimensional model to judge the practical consequences of multidimensionality. Limitations of the proposed approach are detailed.Keywords
Funding Information
- United States (US) Department of Veterans Affairs Health Services Research and Development Service and from the Loma Linda University Department of Psychiatry
This publication has 50 references indexed in Scilit:
- Psychometric evaluation of a patient-reported item bank for healthcare engagementQuality of Life Research, 2021
- Multi-Faceted Constructs in Abnormal Psychology: Implications of the Bifactor S - 1 Model for Individual Clinical AssessmentJournal of Abnormal Child Psychology, 2020
- Comparison of Three Unidimensional Approaches to Represent a Two-Dimensional Latent Ability Space2019
- Conditional Covariance Theory and Detect for Polytomous ItemsPsychometrika, 2006
- Cronbach’s α, Revelle’s β, and Mcdonald’s ωH: their Relations with Each Other and Two Alternative Conceptualizations of ReliabilityPsychometrika, 2005
- The Greatest Lower Bound to the Reliability of a Test and the Hypothesis of UnidimensionalityPsychometrika, 2004
- Graded Response Model1997
- Full-Information Item Bi-Factor AnalysisPsychometrika, 1992
- On the Relationship between iTem Response Theory and Factor Analysis of Discretized VariablesPsychometrika, 1987
- The Bi-Factor MethodPsychometrika, 1937