Using Functional Data Analysis to Analyze Unfolding Models

• In Event: Advances in Rasch Measurement

Abstract

Unfolding models are frequently used for the measurement of attitudes. This study describes an approach for estimating unfolding models using functional data analysis (FDA). Furthermore, this study illustrates the use of graphical displays within a functional data analysis framework to evaluate model-data fit in the context of unfolding models. To illustrate the use of FDA to analyze unfolding models, data collected about attitudes toward capital punishment (N=54) are used (Andrich, 1988). Preliminary results suggest that the use of FDA to estimate person response functions (PRFs) in the context of unfolding models combined with graphical displays of the derivatives of the smooth functions is a promising approach. The examination of derivative of the PRFs gives additional insight regarding person fit.

Authors

• Kyle Turner, University of Georgia - Athens

• George Engelhard, University of Georgia

• Jue Wang, University of Science and Technology of China