Training Principals to Reliably Score Classroom Observations for Teacher Evaluation: An Exploration

• In Event: Roundtable Session 4
  In Roundtable Session: Teacher Agency/Empowerment in Assessment and Policy Making

Abstract

Reliable scoring of classroom observations for teacher evaluation is a significant challenge. This study explores three reliability issues: to what degree the trainer matters, to what degree the specific teaching practice matters, and what is the reliability of a single observer. Principals (n = 150) were trained to score classroom observations using a rubric based on InTASC standards in an authentic, high stakes evaluation system. They took a video-based exam at the end of training. Reliability of their exam scores was explored using generalizability theory. Results suggest that who the trainer is does matter and scores for some teaching practices are less reliable than others. Reliability estimates were comparable with estimates from another study, even with a single rater.

Authors

• Christi Crosby Bergin, University of Missouri

• Sara Bernice Chapman, University of Missouri - Columbia

• Chia-Lin Tsai, University of Missouri

• Francis Howard Lim Huang, University of Missouri