A New Multiprocess IRT Model With Ideal Points for Likert-Type Items. (June 2022)

• Record Type: Journal Article
• Authors: Jin, Kuan-Yu; Wu, Yi-Jhen; Chen, Hui-Fang
• Is Part Of: Journal of educational and behavioral statistics. Volume 47:Number 3(2022)
• Page Start: 297
• Page End: 321
• Publication Date: 2022-06
• Subjects: Rtree -- dominance -- ideal point -- unfolding model -- extreme response style; Educational statistics -- Periodicals; Social sciences -- Statistical methods -- Periodicals; 370.2
• DOI: 10.3102/10769986211057160
• Languages: English
• ISSNs: 1076-9986

Abstract

For surveys of complex issues that entail multiple steps, multiple reference points, and nongradient attributes (e.g., social inequality), this study proposes a new multiprocess model that integrates ideal-point and dominance approaches into a treelike structure (IDtree). In the IDtree, an ideal-point approach describes an individual's attitude and then a dominance approach describes their tendency for using extreme response categories. Evaluation of IDtree performance via two empirical data sets showed that the IDtree fit these data better than other models. Furthermore, simulation studies showed a satisfactory parameter recovery of the IDtree. Thus, the IDtree model sheds light on the response processes of a multistage structure.