A Unified True-Score Framework for Classification Indices with Application to Curriculum-Based Measures

• In Event: Classification

Abstract

We introduce a unified treatment of computing classification accuracy and consistency indices based on true-score theory using single-administration data. New and existing models for true scores and errors are assembled, and some model fit criteria are suggested. The proposed framework is applied to data from several K-3 curriculum-based measures (CBMs).

Authors

• Edison Choe, Renaissance

• Adam E Wyse, Renaissance