Investigating the Performance of Item Response Theory Proficiency Estimators Under Multistage Testing

• In Event: Marriott Roundtable Session 41
  In Roundtable Session: Multistage Testing

Abstract

Multistage testing (MST) has the advantages on content and quality control of a linear test as well as the flexibility and efficiency of a CAT test. It has been shown to increase the proficiency estimation precision and classification accuracy. Most research on MST has focused on how design features in MST impact the quality of proficiency estimation and classification accuracy. However, little research has investigated the performance of different proficiency estimators under MST design features. The purpose of this study was to investigate the performance of proficiency estimators under module difficulty difference and module length difference conditions in a 1-3 MST design.

Author

• Liuhan (Sophie) Cai, Measured Progress