Predictive validity of peer evaluation in problem-based learning: an exploration in medical education

Area of Science:

• Medical Education
• Collaborative Learning
• Assessment Methods

Background:

• Peer evaluation is a common tool in problem-based learning (PBL) for assessing collaborative learning and clinical reasoning.
• Evidence on the predictive validity of peer evaluation in medical education is limited.
• This study investigated the predictive power of peer evaluation for clinical performance and licensing exams.

Purpose of the Study:

• To determine if peer evaluation scores in PBL predict future clinical practice performance.
• To assess the relationship between peer evaluation and written Korean Medical Licensing Examination (KMLE) scores.
• To understand the distinct educational value of peer evaluation compared to other assessment methods.

Main Methods:

• Correlation analyses were used to examine predictive validity.
• Peer evaluations, tutor evaluations, and PBL comprehensive exam scores were analyzed.
• Relationships were assessed against clinical practice course scores and written KMLE scores.

Main Results:

• Peer evaluation scores strongly correlated with skills and attitudes domains in clinical practice courses.
• Comprehensive exam scores were more predictive of the knowledge domain.
• Peer evaluation showed lower predictive validity for the written KMLE compared to comprehensive exams.

Conclusions:

• Peer evaluation in PBL is linked to subsequent clinical practice, especially in skills and attitudes.
• Peer evaluation offers unique insights into collaborative competencies beyond tutor evaluations and written tests.
• Further research is recommended to optimize peer evaluation implementation and understand its underlying mechanisms.