Prediction Observation Explanation¶

Strategy · All strategies
Evidence · 3 claims (3 for) · 9 studies (5 quant-synthesis, 3 causal, 1 theoretical), q2–q4 · 4 of 9 report an effect size

Description¶

Prediction Observation Explanation (POE) is a three-phase strategy, originally developed by White and Gunstone for science education. Learners first commit to an explicit prediction about the outcome of a demonstration or event, then observe the actual outcome, and finally write an explanation that reconciles prediction and observation — especially when the two diverge. The commitment step is essential: a public or recorded prediction creates the cognitive stakes that make the observation meaningful.

Design Implications¶

POE works because committing to a prediction activates prior knowledge — including misconceptions — and makes it available for revision Activating prior knowledge improves learning. [+M]. When observation contradicts prediction, the resulting cognitive conflict motivates conceptual change Cognitive disequilibrium motivates conceptual change. [+M]. The strategy is a structured form of Active Learning: every learner must commit before the reveal, preventing passive observation of a demonstration.

Context¶

Requirements¶

• An event, demonstration, or dataset with a genuinely uncertain or non-obvious outcome — if the outcome is trivially predictable, the prediction phase is busywork
• A mechanism for every learner to commit individually (written, clicker, or app-based) before any group discussion or reveal
• Time and structure for the explanation phase; the reconciliation writing is where learning consolidates, and truncating it reduces the strategy to a demonstration
• Instructor preparation for common misconceptions, so the explanation phase can target them directly

Constraints¶

• Ineffective when learners can predict correctly without understanding — surface cues let them succeed for the wrong reasons, and no conceptual change occurs [~M]
• If the observation is ambiguous or the demonstration fails, the discrepancy learners resolve is an artifact, not the target concept
• Learners with very low prior knowledge may guess randomly, generating no meaningful prediction to revise; a brief advance organizer or prior Activation activity is needed first
• Overuse with the same format breeds ritual compliance — learners predict to match what they think the instructor wants rather than their actual belief [-W]

Implementation Variability¶

• Peer Instruction variant: after individual commitment, learners discuss predictions with peers before the reveal, then re-vote — peer discussion substantially improves conceptual gains Active learning improves exam performance. [+S]
• POE with data: replace live demonstration with datasets or simulations (e.g., PhET interactive simulations) where learners predict graph or output behavior
• Delayed observation: in flipped settings, learners predict as pre-class homework and observe during class, freeing class time for the explanation phase

Target Learners¶

• Learners holding common misconceptions in domains like physics, chemistry, and statistics — the discrepancy is the mechanism that surfaces and confronts them Cognitive disequilibrium motivates conceptual change. [+M]
• Undergraduate lecture audiences at scale, where clicker-based commitment is feasible Active learning improves exam performance. [+S]
• Less suitable for complete novices with no basis for prediction; a minimal knowledge foundation is required for the prediction to be meaningful

Target Learning Goals¶

• Conceptual change: replacing intuitive but incorrect models with scientific ones
• Causal reasoning: articulating why an outcome occurs, not just what occurs
• Metacognitive monitoring: recognizing the gap between what one believes and what is true

Instructions¶

1. Present the setup of the event or demonstration without the outcome, and have each learner commit to a written or clicker prediction with a brief rationale.
2. Show or run the observation — live, via simulation, or with data.
3. Have learners write an explanation reconciling prediction and observation, individually first, then optionally in Class Discussion.
4. Close with instructor consolidation that names the target concept and explicitly addresses the misconception the discrepancy was designed to expose.

Related Strategies¶

• Peer Instruction — the clicker-based discussion variant of POE developed by Crouch and Mazur; adds peer debate between prediction and reveal
• Predict-Observe-Explain labs — the same structure applied to full laboratory investigations rather than single demonstrations
• Activating Prior Knowledge (Activating Prior Knowledge) — POE's prediction phase is a forced, commitment-based form of activation
• Use PDEODE tasks (Predict-Discuss-Explain-Observe-Discuss-Explain) to facilitate conceptual change in chemistry topics

Examples¶

• Peer Instruction at Harvard (Crouch & Mazur, 2001) — ConcepTests in introductory physics follow the POE cycle: individual clicker vote, peer discussion, re-vote, instructor explanation.
• PhET Interactive Simulations (https://phet.colorado.edu) — Learners predict circuit or gas-law behavior, then manipulate the simulation to observe the outcome and explain discrepancies.
• Chemistry "misconception demos" — Classic POE tasks such as predicting the mass of dissolved salt or the water level when a candle burns under a jar, each targeting a documented intuitive misconception.

Key Sources¶

• White, R., & Gunstone, R. (1992). Probing understanding. London: Falmer Press. (Chapter on Prediction–Observation–Explanation)
• Crouch, C. H., & Mazur, E. (2001). Peer Instruction: Ten years of experience and results. American Journal of Physics, 69(9), 970–977. doi:10.1119/1.1374249
• Freeman, S., Eddy, S. L., McDonough, M., Smith, M. K., Okoroafor, N., Jordt, H., & Wenderoth, M. P. (2014). Active learning increases student performance in science, engineering, and mathematics. PNAS, 111(23), 8410–8415. doi:10.1073/pnas.1319030111
• Vosniadou, S. (2013). Conceptual change in learning and instruction. In International Handbook of Research on Conceptual Change. Routledge. doi:10.4324/9780203154472.ch1