Functional Clustering for Diagnosing Person Misfit

• In Event: Rasch Models: Considerations and Use

Abstract

The purpose of this study is to illustrate the use of functional data analysis (FDA) as a general methodology for analysing person response functions (PRFs). FDA offers an approach for diagnosing person responses that have been identified by traditional fit statistics as aberrant. PRFs provide graphical displays that can provide insight into unexpected response patterns. In addition to examining individual PRFs, functional clustering techniques can be used to identify subgroups among persons flagged as misfitting that may be exhibiting categories of misfit such as guessing. The methodology is applied to simulated data, as well as data from a physical science assessment used in a southeastern state. FDA offers a promising methodology for evaluating whether meaningful person scores have been obtained.

Authors

• Kyle Turner, University of Georgia - Athens

• George Engelhard, University of Georgia