How to answer predict questions is to state the most likely outcome based on a pattern, mechanism, model or evidence already available. Strong predictions are not guesses: they make the expected result explicit and connect it to the reasoning that supports it when the marks require justification.
Prediction appears across Science, Mathematics, data questions, reading and source work. The evidence may be a trend, a known mechanism, a sequence, a model or a set of conditions.
This guide explains how to distinguish prediction from guessing, select a defensible trend or mechanism, state the outcome clearly and qualify predictions when the evidence has limits.
This guide is written for students, parents and educators who want a practical answer to how to answer predict questions in science and exams. The purpose is to turn the command word into a repeatable operating system: identify the task, select the correct depth and evidence, construct the response, check it against the question and then practise again on a changed prompt.
The 60-Second Answer
Identify exactly what future or unobserved result is being asked about. Find the relevant pattern, rule or mechanism. State the predicted outcome directly. If justification is required, show the evidence or causal link supporting it. Keep the prediction within the range and conditions of the evidence, and qualify uncertainty when extrapolation is weak.
Working definition: A predict question asks the learner to state a likely outcome or value based on available evidence, patterns, rules, models or known mechanisms.
Why Predict Questions Matters
Prediction tests whether the learner can use existing information rather than merely repeat it.
A strong prediction often depends on recognizing what remains constant when the situation changes.
The learner must also respect limits: evidence within a known range may not justify a confident extrapolation far beyond it.
The useful standard is whether the learner can recognise the command word in an unfamiliar assessment and produce the right type of reasoning without relying on a tutor, template or memorised answer.
The Hidden Problem: A Prediction Can Sound Confident and Still Be a Guess
Students sometimes write what feels plausible without identifying the evidence behind it.
The predicted outcome may accidentally be correct, but the reasoning is not transferable.
A strong method ties prediction to a trend, mechanism or rule that can be inspected and challenged.
The Operating Model
Decode Outcome → Identify Evidence → Select Pattern/Mechanism → State Prediction → Justify → Check Conditions → Qualify.
Each stage has a different function. Interpretation identifies what the command requires. Selection chooses the relevant material. Construction turns that material into a markable response. Checking verifies scope, completeness and precision. Delayed retesting shows whether the method transfers beyond one familiar question.
How to answer predict questions: Step by Step
1. Identify what must be predicted
Clarify the future value, behaviour, result, response or event.
Do not predict unrelated consequences.
2. Find the relevant evidence
Use the trend, model, prior stage, mechanism or contextual clue that bears on the outcome.
Ignore decorative data.
3. Choose the governing pattern or mechanism
State what relationship is expected to continue.
Avoid unsupported intuition.
4. Make the prediction directly
Give the likely outcome in clear language.
Do not bury it after the explanation.
5. Justify when required
Use data, pattern or mechanism to show why the prediction follows.
Match depth to marks.
6. Check conditions
Ask whether the prediction stays within the known range and assumptions.
Extrapolation deserves more caution.
7. Quantify where appropriate
For numerical trends, estimate or calculate using the relevant relationship.
Include units.
8. Qualify uncertainty
Use likely, expected or approximately where evidence does not support certainty.
Do not weaken a well-supported prediction unnecessarily.
What Strong Performance Looks Like
- The target outcome is clear.
- Relevant evidence is selected.
- A pattern or mechanism supports the prediction.
- The outcome is stated directly.
- Justification matches marks.
- Conditions are considered.
- Numerical predictions use units.
- Certainty is calibrated.
These indicators are behavioural rather than personal. They can therefore be taught, rehearsed, checked and improved.
Common Failure Modes
1. Pure guessing
The prediction has no evidence.
Name the pattern or mechanism.
2. Evidence without outcome
The student explains the trend but never states what will happen.
Give the prediction first.
3. Over-extrapolation
A short trend is extended far beyond the data.
Qualify the claim.
4. Wrong variable
The learner predicts something different from what the question asks.
Identify the target explicitly.
5. Mechanism mismatch
A general rule is used outside its conditions.
Check assumptions.
6. False precision
An exact number is given when only a rough trend is supported.
Use appropriate precision.
7. No unit
A numerical prediction lacks units.
Include them.
8. Hedging everything
Well-supported predictions are weakened with excessive uncertainty.
Match confidence to evidence.
How the Strategy Changes Across Subjects
Science
Predict outcomes from mechanisms, variable relationships, experimental trends and models, while respecting limits and conditions.
The subject standard remains decisive. General exam technique can organise the response, but discipline-specific knowledge determines whether the content is actually correct and sufficient.
Mathematics
Predict terms, trends or graph behaviour from patterns and established relationships.
The subject standard remains decisive. General exam technique can organise the response, but discipline-specific knowledge determines whether the content is actually correct and sufficient.
English
Predict likely actions or developments from textual clues, character behaviour and context without inventing unsupported events.
The subject standard remains decisive. General exam technique can organise the response, but discipline-specific knowledge determines whether the content is actually correct and sufficient.
Humanities
Predict likely consequences cautiously from stated conditions or trends, distinguishing prediction from established historical fact.
The subject standard remains decisive. General exam technique can organise the response, but discipline-specific knowledge determines whether the content is actually correct and sufficient.
Primary School, Secondary School and Examination Years
Primary school
Use what happened before and what rule is operating to predict what comes next. Ask children to point to the clue supporting the prediction.
Use short prompts and concrete examples. Adults can model the response structure, then fade the scaffold as the learner becomes able to select the correct response type independently.
Secondary school
Students should justify predictions with mechanisms or data and distinguish interpolation from uncertain extrapolation.
Secondary learners should increasingly decide scope, evidence, depth and checking without waiting for a teacher to signal what belongs in the answer.
Before major examinations
Use unseen graphs, experimental scenarios and text passages so prediction depends on evidence rather than memorized outcomes.
Near major exams, mix command words in the same practice set so the learner must identify the required thinking before starting to write.
A Focused 60-Minute Practice Session
- 10 minutes — identify the target in five predict prompts.
- 10 minutes — locate supporting patterns or mechanisms.
- 15 minutes — write direct predictions plus justification.
- 10 minutes — practise numerical trend predictions.
- 10 minutes — compare interpolation and extrapolation.
- 5 minutes — audit certainty language.
The exact timings can change. What matters is the sequence: command identification, independent answer, feedback, targeted repair and fresh retest.
Diagnostic Checklist
- Is the target outcome clear?
- Is relevant evidence selected?
- Is a pattern or mechanism identified?
- Is the prediction stated directly?
- Is justification present when needed?
- Are conditions checked?
- Are units and precision appropriate?
- Is certainty calibrated?
Use the checklist to identify the earliest decision that failed. A weak response may begin with a command-word error long before the final sentence appears.
Practice Laboratory
Practice 1: Pattern continuation
Use a short table or sequence to predict the next value.
State the rule.
Practice 2: Mechanism prediction
Given a changed condition, predict the result from the causal mechanism.
Explain the link.
Practice 3: Evidence-first
Underline the clue before writing the prediction.
Reject unsupported possibilities.
Practice 4: Interpolation versus extrapolation
Classify predictions inside and outside the observed range.
Adjust certainty.
Practice 5: Precision check
Decide whether the answer should be exact, approximate or directional.
Match evidence.
Practice 6: Alternative outcomes
List two possible results, then choose the one best supported.
Explain why.
Practice 7: Text prediction
Use character behaviour or narrative evidence to predict the next action.
Avoid inventing unrelated plot.
Practice 8: Fresh-context transfer
Apply the same rule to a changed scenario.
Check whether conditions still hold.
Three Student Cases
Case 1: Maren guesses
Maren writes plausible future outcomes but cannot explain why.
Her tutor requires one evidence sentence before every prediction.
Guessing becomes evidence-based reasoning.
Case 2: Iona extrapolates too far
Iona extends a short linear graph far beyond the observed range.
She begins labelling interpolation and extrapolation explicitly.
Her certainty becomes more appropriate.
Case 3: Leonie explains but forgets to predict
Leonie writes a strong mechanism but never states the actual future result.
A prediction-first rule is introduced.
Her answers become easier to credit.
How Parents Can Help Without Taking Over
Parents can support predict questions by asking the learner to explain the command and response plan rather than supplying content. Useful prompts include: “What does the command word ask you to do?”, “Which evidence belongs?”, “How many parts are needed?”, “What would make this incomplete?”, and “How will you know when to stop?”
Once the learner can perform the decision reliably, remove the prompt. The goal is independent exam control.
How Tutors Can Use a Three-Student Small Group
In a three-student tutorial, predict questions becomes visible because one learner can answer, another can challenge whether the command has been satisfied, and the third can check against a rubric or mark scheme. Roles then rotate.
After discussion, each learner completes a fresh individual question. Shared explanation becomes useful only when it transfers into solo performance.
How to Measure Progress
- Unsupported guessing decreases.
- Evidence links improve.
- Predictions become more direct.
- Extrapolation becomes more cautious.
- Numerical precision improves.
- Units appear consistently.
- Mechanism transfer improves.
- Certainty language becomes calibrated.
Marks are the final indicator, but process improvements often appear first: more direct openings, fewer irrelevant sentences, better evidence selection, stronger scope control and faster command-word recognition.
A Four-Week Implementation Plan
Week 1 — Evidence-backed predictions
Use simple patterns and mechanisms.
Require one supporting clue.
Week 2 — Add changed conditions
Predict outcomes when one variable changes.
Check assumptions.
Week 3 — Add numerical and extrapolation tasks
Practise precision and uncertainty.
Use mixed data.
Week 4 — Simulate
Complete predict questions across subjects under time.
Review whether failures began with target, evidence or mechanism.
Evidence and Responsible Use
Cambridge International’s emphasis on practical and experimental skills includes using evidence, patterns and scientific reasoning to work with unfamiliar situations. Cornell’s exam guidance likewise stresses decoding what a question asks before responding. Prediction therefore works best when it is anchored to a clear pattern or mechanism rather than treated as free guessing.
Command-word meanings can vary slightly by subject, school and examination board. Use official syllabus documents, teacher guidance and mark schemes as the final standard, while using this framework to build transferable reasoning habits.
- Cambridge International: Practical skills
- Cornell: How to Tackle Exam Questions
- Princeton: Preparing for Exams
Frequently Asked Questions
Is prediction the same as guessing?
No. A prediction should be supported by evidence, a pattern or a mechanism.
Do I always need to explain why?
Use the marks and wording; some predict questions ask only for the outcome, while others require justification.
Can I predict an exact number?
Only when the evidence supports that precision.
What is extrapolation?
Predicting beyond the observed data range, usually with more uncertainty.
Should I use ‘will’ or ‘may’?
Match certainty to the strength and scope of the evidence.
Can predictions be wrong and still earn reasoning marks?
That depends on the assessment, but transparent valid reasoning is always preferable to unsupported guessing.
How do I predict from a graph?
Identify the trend, scale and range, then extend cautiously if the task requires it.
How do I practise?
Use unfamiliar data and changed-condition questions, then compare predictions with outcomes or mark schemes.
Helpful Reading on eduKateSingapore
- How to Read Exam Questions
- How to Answer Describe Questions
- How to Answer Explain Questions
- How to Check Your Work
- How to Answer Data Interpretation Questions
- How to Answer Inference Questions
Prediction Is Evidence Used Forward
The answer should state what is likely to happen and show the pattern or mechanism that makes that outcome reasonable.
Identify the target. Find the evidence. Apply the rule. State the prediction. Check the conditions. Qualify only as much as the evidence requires.
Properly taught kids shine a bright light into the future.
Extended Diagnostic Workshops
Workshop 1: Identify what must be predicted × Evidence without outcome
Set up one fresh exam-style question where the learner must practise identify what must be predicted while watching specifically for evidence without outcome. Clarify the future value, behaviour, result, response or event. Require an independent response before any model is shown so the actual decision becomes visible.
The student explains the trend but never states what will happen. After correction, change one meaningful feature—topic, wording, evidence, mark allocation, data, source or context. Then use this related exercise: Underline the clue before writing the prediction. Reject unsupported possibilities. Finish with a delayed retest so success cannot be explained only by immediate memory of the correction.
Workshop 2: Find the relevant evidence × Mechanism mismatch
Set up one fresh exam-style question where the learner must practise find the relevant evidence while watching specifically for mechanism mismatch. Use the trend, model, prior stage, mechanism or contextual clue that bears on the outcome. Require an independent response before any model is shown so the actual decision becomes visible.
A general rule is used outside its conditions. After correction, change one meaningful feature—topic, wording, evidence, mark allocation, data, source or context. Then use this related exercise: Decide whether the answer should be exact, approximate or directional. Match evidence. Finish with a delayed retest so success cannot be explained only by immediate memory of the correction.
Workshop 3: Choose the governing pattern or mechanism × Hedging everything
Set up one fresh exam-style question where the learner must practise choose the governing pattern or mechanism while watching specifically for hedging everything. State what relationship is expected to continue. Require an independent response before any model is shown so the actual decision becomes visible.
Well-supported predictions are weakened with excessive uncertainty. After correction, change one meaningful feature—topic, wording, evidence, mark allocation, data, source or context. Then use this related exercise: Use character behaviour or narrative evidence to predict the next action. Avoid inventing unrelated plot. Finish with a delayed retest so success cannot be explained only by immediate memory of the correction.
Workshop 4: Make the prediction directly × Over-extrapolation
Set up one fresh exam-style question where the learner must practise make the prediction directly while watching specifically for over-extrapolation. Give the likely outcome in clear language. Require an independent response before any model is shown so the actual decision becomes visible.
A short trend is extended far beyond the data. After correction, change one meaningful feature—topic, wording, evidence, mark allocation, data, source or context. Then use this related exercise: Use a short table or sequence to predict the next value. State the rule. Finish with a delayed retest so success cannot be explained only by immediate memory of the correction.
Workshop 5: Justify when required × False precision
Set up one fresh exam-style question where the learner must practise justify when required while watching specifically for false precision. Use data, pattern or mechanism to show why the prediction follows. Require an independent response before any model is shown so the actual decision becomes visible.
An exact number is given when only a rough trend is supported. After correction, change one meaningful feature—topic, wording, evidence, mark allocation, data, source or context. Then use this related exercise: Underline the clue before writing the prediction. Reject unsupported possibilities. Finish with a delayed retest so success cannot be explained only by immediate memory of the correction.
Workshop 6: Check conditions × Pure guessing
Set up one fresh exam-style question where the learner must practise check conditions while watching specifically for pure guessing. Ask whether the prediction stays within the known range and assumptions. Require an independent response before any model is shown so the actual decision becomes visible.
The prediction has no evidence. After correction, change one meaningful feature—topic, wording, evidence, mark allocation, data, source or context. Then use this related exercise: Decide whether the answer should be exact, approximate or directional. Match evidence. Finish with a delayed retest so success cannot be explained only by immediate memory of the correction.
Workshop 7: Quantify where appropriate × Wrong variable
Set up one fresh exam-style question where the learner must practise quantify where appropriate while watching specifically for wrong variable. For numerical trends, estimate or calculate using the relevant relationship. Require an independent response before any model is shown so the actual decision becomes visible.
The learner predicts something different from what the question asks. After correction, change one meaningful feature—topic, wording, evidence, mark allocation, data, source or context. Then use this related exercise: Use character behaviour or narrative evidence to predict the next action. Avoid inventing unrelated plot. Finish with a delayed retest so success cannot be explained only by immediate memory of the correction.
Workshop 8: Qualify uncertainty × No unit
Set up one fresh exam-style question where the learner must practise qualify uncertainty while watching specifically for no unit. Use likely, expected or approximately where evidence does not support certainty. Require an independent response before any model is shown so the actual decision becomes visible.
A numerical prediction lacks units. After correction, change one meaningful feature—topic, wording, evidence, mark allocation, data, source or context. Then use this related exercise: Use a short table or sequence to predict the next value. State the rule. Finish with a delayed retest so success cannot be explained only by immediate memory of the correction.
Workshop 9: Identify what must be predicted × Evidence without outcome
Set up one fresh exam-style question where the learner must practise identify what must be predicted while watching specifically for evidence without outcome. Clarify the future value, behaviour, result, response or event. Require an independent response before any model is shown so the actual decision becomes visible.
The student explains the trend but never states what will happen. After correction, change one meaningful feature—topic, wording, evidence, mark allocation, data, source or context. Then use this related exercise: Underline the clue before writing the prediction. Reject unsupported possibilities. Finish with a delayed retest so success cannot be explained only by immediate memory of the correction.
Workshop 10: Find the relevant evidence × Mechanism mismatch
Set up one fresh exam-style question where the learner must practise find the relevant evidence while watching specifically for mechanism mismatch. Use the trend, model, prior stage, mechanism or contextual clue that bears on the outcome. Require an independent response before any model is shown so the actual decision becomes visible.
A general rule is used outside its conditions. After correction, change one meaningful feature—topic, wording, evidence, mark allocation, data, source or context. Then use this related exercise: Decide whether the answer should be exact, approximate or directional. Match evidence. Finish with a delayed retest so success cannot be explained only by immediate memory of the correction.
Workshop 11: Choose the governing pattern or mechanism × Hedging everything
Set up one fresh exam-style question where the learner must practise choose the governing pattern or mechanism while watching specifically for hedging everything. State what relationship is expected to continue. Require an independent response before any model is shown so the actual decision becomes visible.
Well-supported predictions are weakened with excessive uncertainty. After correction, change one meaningful feature—topic, wording, evidence, mark allocation, data, source or context. Then use this related exercise: Use character behaviour or narrative evidence to predict the next action. Avoid inventing unrelated plot. Finish with a delayed retest so success cannot be explained only by immediate memory of the correction.
Workshop 12: Make the prediction directly × Over-extrapolation
Set up one fresh exam-style question where the learner must practise make the prediction directly while watching specifically for over-extrapolation. Give the likely outcome in clear language. Require an independent response before any model is shown so the actual decision becomes visible.
A short trend is extended far beyond the data. After correction, change one meaningful feature—topic, wording, evidence, mark allocation, data, source or context. Then use this related exercise: Use a short table or sequence to predict the next value. State the rule. Finish with a delayed retest so success cannot be explained only by immediate memory of the correction.
Workshop 13: Justify when required × False precision
Set up one fresh exam-style question where the learner must practise justify when required while watching specifically for false precision. Use data, pattern or mechanism to show why the prediction follows. Require an independent response before any model is shown so the actual decision becomes visible.
An exact number is given when only a rough trend is supported. After correction, change one meaningful feature—topic, wording, evidence, mark allocation, data, source or context. Then use this related exercise: Underline the clue before writing the prediction. Reject unsupported possibilities. Finish with a delayed retest so success cannot be explained only by immediate memory of the correction.
Workshop 14: Check conditions × Pure guessing
Set up one fresh exam-style question where the learner must practise check conditions while watching specifically for pure guessing. Ask whether the prediction stays within the known range and assumptions. Require an independent response before any model is shown so the actual decision becomes visible.
The prediction has no evidence. After correction, change one meaningful feature—topic, wording, evidence, mark allocation, data, source or context. Then use this related exercise: Decide whether the answer should be exact, approximate or directional. Match evidence. Finish with a delayed retest so success cannot be explained only by immediate memory of the correction.
Workshop 15: Quantify where appropriate × Wrong variable
Set up one fresh exam-style question where the learner must practise quantify where appropriate while watching specifically for wrong variable. For numerical trends, estimate or calculate using the relevant relationship. Require an independent response before any model is shown so the actual decision becomes visible.
The learner predicts something different from what the question asks. After correction, change one meaningful feature—topic, wording, evidence, mark allocation, data, source or context. Then use this related exercise: Use character behaviour or narrative evidence to predict the next action. Avoid inventing unrelated plot. Finish with a delayed retest so success cannot be explained only by immediate memory of the correction.
Workshop 16: Qualify uncertainty × No unit
Set up one fresh exam-style question where the learner must practise qualify uncertainty while watching specifically for no unit. Use likely, expected or approximately where evidence does not support certainty. Require an independent response before any model is shown so the actual decision becomes visible.
A numerical prediction lacks units. After correction, change one meaningful feature—topic, wording, evidence, mark allocation, data, source or context. Then use this related exercise: Use a short table or sequence to predict the next value. State the rule. Finish with a delayed retest so success cannot be explained only by immediate memory of the correction.
Workshop 17: Identify what must be predicted × Evidence without outcome
Set up one fresh exam-style question where the learner must practise identify what must be predicted while watching specifically for evidence without outcome. Clarify the future value, behaviour, result, response or event. Require an independent response before any model is shown so the actual decision becomes visible.
The student explains the trend but never states what will happen. After correction, change one meaningful feature—topic, wording, evidence, mark allocation, data, source or context. Then use this related exercise: Underline the clue before writing the prediction. Reject unsupported possibilities. Finish with a delayed retest so success cannot be explained only by immediate memory of the correction.
Workshop 18: Find the relevant evidence × Mechanism mismatch
Set up one fresh exam-style question where the learner must practise find the relevant evidence while watching specifically for mechanism mismatch. Use the trend, model, prior stage, mechanism or contextual clue that bears on the outcome. Require an independent response before any model is shown so the actual decision becomes visible.
A general rule is used outside its conditions. After correction, change one meaningful feature—topic, wording, evidence, mark allocation, data, source or context. Then use this related exercise: Decide whether the answer should be exact, approximate or directional. Match evidence. Finish with a delayed retest so success cannot be explained only by immediate memory of the correction.
Workshop 19: Choose the governing pattern or mechanism × Hedging everything
Set up one fresh exam-style question where the learner must practise choose the governing pattern or mechanism while watching specifically for hedging everything. State what relationship is expected to continue. Require an independent response before any model is shown so the actual decision becomes visible.
Well-supported predictions are weakened with excessive uncertainty. After correction, change one meaningful feature—topic, wording, evidence, mark allocation, data, source or context. Then use this related exercise: Use character behaviour or narrative evidence to predict the next action. Avoid inventing unrelated plot. Finish with a delayed retest so success cannot be explained only by immediate memory of the correction.
Workshop 20: Make the prediction directly × Over-extrapolation
Set up one fresh exam-style question where the learner must practise make the prediction directly while watching specifically for over-extrapolation. Give the likely outcome in clear language. Require an independent response before any model is shown so the actual decision becomes visible.
A short trend is extended far beyond the data. After correction, change one meaningful feature—topic, wording, evidence, mark allocation, data, source or context. Then use this related exercise: Use a short table or sequence to predict the next value. State the rule. Finish with a delayed retest so success cannot be explained only by immediate memory of the correction.
Workshop 21: Justify when required × False precision
Set up one fresh exam-style question where the learner must practise justify when required while watching specifically for false precision. Use data, pattern or mechanism to show why the prediction follows. Require an independent response before any model is shown so the actual decision becomes visible.
An exact number is given when only a rough trend is supported. After correction, change one meaningful feature—topic, wording, evidence, mark allocation, data, source or context. Then use this related exercise: Underline the clue before writing the prediction. Reject unsupported possibilities. Finish with a delayed retest so success cannot be explained only by immediate memory of the correction.
Workshop 22: Check conditions × Pure guessing
Set up one fresh exam-style question where the learner must practise check conditions while watching specifically for pure guessing. Ask whether the prediction stays within the known range and assumptions. Require an independent response before any model is shown so the actual decision becomes visible.
The prediction has no evidence. After correction, change one meaningful feature—topic, wording, evidence, mark allocation, data, source or context. Then use this related exercise: Decide whether the answer should be exact, approximate or directional. Match evidence. Finish with a delayed retest so success cannot be explained only by immediate memory of the correction.
Workshop 23: Quantify where appropriate × Wrong variable
Set up one fresh exam-style question where the learner must practise quantify where appropriate while watching specifically for wrong variable. For numerical trends, estimate or calculate using the relevant relationship. Require an independent response before any model is shown so the actual decision becomes visible.
The learner predicts something different from what the question asks. After correction, change one meaningful feature—topic, wording, evidence, mark allocation, data, source or context. Then use this related exercise: Use character behaviour or narrative evidence to predict the next action. Avoid inventing unrelated plot. Finish with a delayed retest so success cannot be explained only by immediate memory of the correction.
Workshop 24: Qualify uncertainty × No unit
Set up one fresh exam-style question where the learner must practise qualify uncertainty while watching specifically for no unit. Use likely, expected or approximately where evidence does not support certainty. Require an independent response before any model is shown so the actual decision becomes visible.
A numerical prediction lacks units. After correction, change one meaningful feature—topic, wording, evidence, mark allocation, data, source or context. Then use this related exercise: Use a short table or sequence to predict the next value. State the rule. Finish with a delayed retest so success cannot be explained only by immediate memory of the correction.
Workshop 25: Identify what must be predicted × Evidence without outcome
Set up one fresh exam-style question where the learner must practise identify what must be predicted while watching specifically for evidence without outcome. Clarify the future value, behaviour, result, response or event. Require an independent response before any model is shown so the actual decision becomes visible.
The student explains the trend but never states what will happen. After correction, change one meaningful feature—topic, wording, evidence, mark allocation, data, source or context. Then use this related exercise: Underline the clue before writing the prediction. Reject unsupported possibilities. Finish with a delayed retest so success cannot be explained only by immediate memory of the correction.
Workshop 26: Find the relevant evidence × Mechanism mismatch
Set up one fresh exam-style question where the learner must practise find the relevant evidence while watching specifically for mechanism mismatch. Use the trend, model, prior stage, mechanism or contextual clue that bears on the outcome. Require an independent response before any model is shown so the actual decision becomes visible.
A general rule is used outside its conditions. After correction, change one meaningful feature—topic, wording, evidence, mark allocation, data, source or context. Then use this related exercise: Decide whether the answer should be exact, approximate or directional. Match evidence. Finish with a delayed retest so success cannot be explained only by immediate memory of the correction.