A MFRM Study Comparing Human and Machine Raters in Scoring Oral Reading

• In Event: GSIC Poster Session I

Abstract

The study assessed the consistency of scoring students' oral readings between human and machine raters. MFRM results indicated an acceptable fit for all raters, with infit and outfit values ranging from 0.50 to 1.50. The machine rater showed significant bias when rating English-L1 students.

Author

• Meng-Hsun Lee, University of Toronto