Many-Faceted Rasch Measurement: Assessing Rater Errors in Performance Assessment

• In Event: Use of Various Rasch Measurement Models

Abstract

The process of rating is a complex and error prone cognitive process (Cronbach, 1990). In this complex process, raters tend to introduce errors that are not attributed to the actual ratees’ performance. This paper was intended to address the teacher supervisors (raters) potential errors through Many Faceted Rasch Measurement (MFRM). The analysis focused on examining the supervisors’ (1) severity in ratings, (2) central tendency, and (3) halo effect errors. Except for a difference in the severity of ratings, supervisors showed good understanding and application of both items and the rating scale. This result recommends minimizing rater bias by providing more training to the supervisor.

Authors

• Priyalatha Govindasamy, University of Denver

• Kathy E. Green, University of Denver

• Maria del Carmen Salazar, University of Denver