Designing Scoring Reliability and Instructional Support into Classroom-Based Math Assessments

• In Event: Reliability and Scoring

Abstract

This study examines how classroom assessments employing open-ended items can attend to scoring reliability in a manner both instructionally supportive and validity enhancing. The approach is implemented for a classroom-focused, math assessment. Data from human-scored student responses are analyzed, with implications for construct validity and score reliability discussed.

Authors

• Xingyao Xiao, Berkeley School of Education, UC BERKELEY

• Richard Patz, UC BERKELEY

• Mark Wilson, Berkeley School of Education, UC BERKELEY