Reliability and Validity of Peer Assessment at Tutorless PBL Classes

Introduction: Learning behaviors in PBL and team-based learning (TBL) at Mie University Faculty of Medicine have been measured with peer assessments in addition to tutor assessments. The change in class components posed a question on reliability of peer assessments at tutorless sessions in TBL.

 

Methods: Correlations between peer and tutor assessments in the PBL program comprised of tutor and tutorless sessions were analyzed with data from 377 students. The three domains for student’s performance, “self-learning in preparation”, “active participation in group discussion” and “cooperative attitude at group work”, were collected at 4 UNITs with the formal PBL and at 4 UNITs with the mixture of PBL and TBL.

 

Results: No statistically significant difference between peer and tutor assessments was found in all three domains in units with the mixture of PBL and TBL. The correlation analysis between peer and tutor assessments in units with the mixture of PBL and TBL demonstrated strong positive correlations in self-learning in preparation (r=0.62) and in active participation in group discussion (r= 0.63), and a weak positive correlation in cooperative attitude at group works (r=0.35).

 

Discussion: This study revealed that students evaluated their group-peers at the units comprised of both tutor and tutorless sessions similarly to those comprised of only tutor sessions. Peer assessments were well correlated with tutors’ assessment at the units comprised of tutor and tutorless session. Our study suggests that TBL might be expanded in the situation where teaching resource is limited.