Cut Points and Item Grouping for Classification Accuracy in Three-Parameter Logistic (3-PL) Computerized Adaptive Tests

• In Event: 72.066 - Poster Session 16
  In Poster Session: 72.066-2 - Poster Session: Practical Issues in Measurement

Abstract

Computer adaptive testing offers greater precision in scores over traditional paper-pencil testing. With emphasis on meeting performance standards in today’s large-scale assessment programs, precision of scores is not enough. The precision of performance classification from computerized adaptive tests (CATs) is as important as the score itself since the classification often drives interpretations of academic achievement. This research study investigates measurement characteristics of CAT item pools and CAT design elements on the relationship to performance classification accuracy. For this study, items are clustered around performance cut points in an attempt to increase classification accuracy over item pools that match the overall student ability. The results of this study and practical implications are discussed.

Authors

• Che-Ming Allen Lau, Pearson

• Marc Johnson, Pearson