How Reliable and Valid Is the Assessment of Problem-Based Learning Performance?

• In Event: Longitudinal Assessment Issues in Health Professions

Abstract

This study examined the reliability and validity of an instrument assessing PBL performance in four domains. Two classes of medical students (N = 310) participated in the study to validate and cross-validate psychometric properties of the instrument. A generalizability study was executed for reliability examination. Pearson correlation coefficients of the PBL scores with six other outcome measures were calculated to examine convergent and discriminant validity. Psychometric findings provided support for reproducibility reliability. Although the findings of one cohort data supported convergent and discriminant validity, the cross-validation results indicated a need for further investigation. The study concludes that students’ performance in PBL is assessed fairly consistently. How PBL ratings are associated with other assessment measures remains to be further explored.

Authors

• Ming Lee, University of California - Los Angeles

• Paul F. Wimmers, University of California - Los Angeles