The Application of Optimal Test Assembly Methods to Items Developed in a Learning Progression Framework

• In Event: Roundtable Session 46
  In Roundtable Session: Assessment and Learning Progressions: Designs and Other Considerations

Abstract

This study applies optimal test assembly methods to the design and construction of formative assessments, using an item bank developed under a learning progression framework. Three test formats were assembled: (1) a fixed-form test, (2) a multistage adaptive test, and (3) a computerized adaptive test (CAT), following a shadow test approach. The precision for identifying varying levels of student performance will be compared. Student abilities and response patterns to a set of real items were simulated. Results indicated that while there is a slight improvement for the CAT method over the other two methods, much more work needs to be done on developing an item bank appropriate for the application of a learning progression within the classroom.

Authors

• Tina Chiu, University of California - Berkeley

• Seung W. Choi, Data Recognition Corporation

• Hao Ren, Data Recognition Corporation