Effects of Problem-Based Learning: A Meta-Analysis
Dochy, F., Segers, M., Van den Bossche, P., & Gijbels, D. · 2003
grade Cmeta-analysisindependentmixed
Sample
43 tertiary (heavily medical) studies
Population
Tertiary/medical students; real curricula; knowledge vs skills.
Design
PBL vs conventional, inherently non-randomized cohort comparisons. Separates KNOWLEDGE and SKILLS outcomes.
Key findings
The classic PBL trade-off: better SKILLS (+0.46) but WORSE immediate factual knowledge (~-0.22), with the knowledge gap shrinking/reversing on delayed retention. Which method 'wins' depends entirely on outcome class and retention interval.
Genetic confound
Domain skills/retention; non-random cohort selection is the confound.
Replication notes
Canonical PBL outcome trade-off; broadened by Walker & Leary 2009 (overall only 0.13).
DOI / URL
10.1016/S0959-4752(02)00025-7
Effects
| Outcome | Metric | Value | Measure | Timing | Vs | Horizon | Class |
|---|---|---|---|---|---|---|---|
| PBL vs conventional on SKILLS | d | +0.46 (robust) | mixed | immediate+delayed | business-as-usual | unclear | domain-skill |
| PBL vs conventional on KNOWLEDGE | d | slightly negative initially (~-0.22, non-robust) but RETAINED better at delay | mixed | immediate vs delayed | business-as-usual | unclear | domain-skill |
Cited by
- Explicit instruction vs discovery/inquiry — how much guidance?moderate supportconf: mediumgc: low