Series: How to Prepare For — Global Examination Performance Edge Articles
Advanced Article P030
The question looks complete.
The calculation begins.
The argument develops.
Then two capable students produce different answers.
Neither made an arithmetic mistake.
Neither misread the obvious wording.
They made different assumptions.
One treated resistance as negligible.
One assumed the trend continued.
One assumed the source was representative.
One assumed all other variables remained constant.
One assumed the missing case behaved like the observed cases.
The assumption sat quietly underneath the solution and controlled the result.
This article owns that edge condition.
It follows How to Prepare for an Exam When the Evidence Is Incomplete | Make the Strongest Defensible Inference Without Inventing the Missing Pieces. That article teaches the learner not to fabricate missing information. This one examines the narrower advanced case in which some assumption may be necessary to proceed, but the answer depends on which assumption is chosen.
The 50-second answer
When an exam answer depends on an assumption:
- Separate what the question states from what you are adding.
- Name the assumption before it silently becomes a fact.
- Ask whether the assumption is required, conventional, optional or unjustified.
- Check whether the syllabus or model normally licenses it.
- Test what happens if the assumption changes.
- Identify whether the final answer is robust or assumption-sensitive.
- Use conditional language when several plausible assumptions remain.
- Prefer the least demanding assumption that is sufficient for the task.
- Do not use an assumption merely because it creates a convenient answer.
- Check boundary cases and counterexamples.
- Where appropriate, state the model limitation.
- Under time, make the assumption explicit and proceed rather than hiding it.
The central rule is:
An assumption is not dangerous because it is an assumption. It is dangerous when it controls the answer without being visible.
Alicia, Tricia and Kai Kai make different assumptions
Alicia assumes the missing condition is the usual one.
Tricia refuses to proceed because the condition is not stated.
Kai Kai asks:
Does the course model conventionally assume this condition unless told otherwise?
If yes, she states it and proceeds.
If no, she tests both plausible branches.
The answer becomes conditional rather than invented.
What this article owns
- Assumption detection: finding hidden premises.
- Assumption classification: separating conventional modelling assumptions from arbitrary guesses.
- Necessity testing: deciding whether an assumption is actually required.
- Sensitivity: testing whether the conclusion changes when the assumption changes.
- Minimality: using no more assumptions than the reasoning needs.
- Conditional answering: preserving multiple branches where evidence cannot choose.
- Model limitation: understanding what the assumption removes from reality.
- Timed transparency: making assumptions visible without consuming the paper.
First principle: separate given from assumed
Use two columns:
Given.
Assumed.
This simple move catches many hidden premises.
Second principle: an assumption is a bridge
Sometimes the evidence does not directly reach the conclusion.
An assumption supplies the missing bridge.
The answer is only as secure as that bridge.
Third principle: conventional assumptions differ from arbitrary assumptions
Some models deliberately simplify reality.
For example, a course may specify ideal conditions, negligible resistance or ceteris paribus reasoning.
These are legitimate when the model or question licenses them.
An arbitrary assumption added for convenience is different.
Fourth principle: ask who licensed the assumption
Was it stated in the question?
Defined by the model?
Established earlier in the problem?
A standard convention of the assessment?
Or merely supplied by the learner?
The source of the assumption affects its legitimacy.
Fifth principle: assumptions should be minimal
Use the smallest premise necessary to proceed.
Do not construct an elaborate invisible world around a simple missing condition.
Sixth principle: convenient assumptions deserve suspicion
If one assumption makes the algebra beautiful, the essay easy or the case obvious, that convenience is not evidence that it is correct.
Test it against the prompt.
Seventh principle: test whether the assumption matters
Change it.
If the answer remains the same, the conclusion is robust to that assumption.
If the answer changes, the assumption is decision-critical.
The sensitivity table
| Assumption | Case A | Case B | Conclusion changes? |
|---|---|---|---|
| Variable X constant | Yes | No | Yes |
| Rounding method | Standard | Alternative | No |
Eighth principle: decision-critical assumptions belong in the answer
If changing the assumption changes the conclusion, make the dependency explicit.
Assuming X, the answer is A; if X does not hold, B remains possible.
Ninth principle: low-impact assumptions can remain compact
Do not turn every routine convention into a paragraph.
Transparency should be proportional to decision impact.
Tenth principle: assumptions can hide inside words
“Typical.”
“Normal.”
“Representative.”
“Efficient.”
“Fair.”
“Constant.”
These words may contain unstated criteria.
Ask what operational meaning the answer requires.
Eleventh principle: extrapolation is assumption-heavy
Projecting beyond observed data assumes something about continuation.
State why the trend should continue or limit the claim.
Twelfth principle: interpolation usually requires fewer assumptions
Estimating within the observed range can be safer, but still depends on the relationship being sufficiently smooth or appropriate.
Do not treat interpolation as assumption-free.
Thirteenth principle: causal claims contain assumptions
To infer causality, the learner may be assuming alternative explanations are controlled or implausible.
Make that reasoning visible.
Fourteenth principle: source representativeness is an assumption
One source, sample or case may not represent a whole population.
If the conclusion generalises beyond the evidence, justify the generalisation.
Fifteenth principle: missing cases should not be assumed average
Unless the problem licenses it, a missing observation does not inherit the mean automatically.
Test bounds or alternative cases.
Sixteenth principle: independence can be an assumption
Probability and statistical reasoning often depend on whether events or observations are independent.
Do not multiply probabilities automatically without checking.
Seventeenth principle: linearity can be an assumption
A relationship that is approximately linear over one range may not remain linear outside it.
Check the domain.
Eighteenth principle: continuity can be an assumption
Some reasoning assumes no abrupt threshold or discontinuity.
If the problem contains a threshold, phase change, policy rule or categorical boundary, continuity may fail.
Nineteenth principle: equilibrium can be an assumption
Models in economics, physics, chemistry and systems analysis may assume equilibrium or steady state.
If the question concerns transition dynamics, that assumption can distort the answer.
Twentieth principle: ceteris paribus has limits
Holding other things equal can isolate a relationship.
But real cases may change several variables at once.
Know whether the exam asks for the model result or the real-world qualification.
Twenty-first principle: model assumptions are not model defects by default
A model simplifies because simplification can make a mechanism visible.
The question is whether the simplification is appropriate for the task.
Twenty-second principle: every model has a validity region
Ask:
Under what conditions does this model provide a useful approximation?
Then keep conclusions inside that region.
Twenty-third principle: boundary cases test assumptions
Set the variable to zero, maximum, minimum or another special case.
Does the assumed model still behave sensibly?
If not, the assumption may have a limited domain.
Twenty-fourth principle: counterexamples expose hidden universals
If the learner’s assumption is “this always happens,” one valid counterexample can force a narrower rule.
Twenty-fifth principle: assumptions can be compared by explanatory cost
If two assumptions both fit the evidence, prefer the one that adds less unsupported structure, unless the domain gives reason otherwise.
This is not a universal law of truth.
It is a useful exam heuristic against unnecessary invention.
Twenty-sixth principle: assumptions can be ranked
Classify:
- explicitly given;
- standard model convention;
- strongly implied;
- plausible but optional;
- unsupported.
The lower the assumption sits, the more cautiously it should control the answer.
Twenty-seventh principle: use conditional branches when assumptions remain unresolved
If two assumptions are both defensible and the prompt does not choose between them:
If X holds, A follows. If Y holds, B follows.
Then state which branch is better supported, if any.
Twenty-eighth principle: conditional answers can be complete answers
Students sometimes think conditional language is evasive.
It is evasive only when the evidence actually supports a unique conclusion.
When the premise is genuinely unresolved, conditional structure is precision.
Twenty-ninth principle: do not smuggle assumptions through diagrams
A diagram can accidentally impose symmetry, scale, direction or geometry not stated in the question.
Mark sketches as schematic unless the problem licenses exact structure.
Thirtieth principle: do not smuggle assumptions through notation
Naming two quantities with the same symbol family or writing a proportional sign can imply a relationship that has not been established.
Notation should record reasoning, not create it.
Thirty-first principle: do not smuggle assumptions through language
Words such as “therefore,” “obviously” and “clearly” can hide a missing step.
Replace them with the actual relationship.
Thirty-second principle: assumptions can enter through prior questions
A result from Part (a) may be intended for Part (b).
But if Part (a) depended on a stated condition, carry that condition forward.
Do not preserve the result while forgetting its assumptions.
Thirty-third principle: linked questions inherit dependencies
If a later answer depends on an earlier assumption, the uncertainty propagates.
Track it.
Do not report downstream precision greater than the upstream premise allows.
Thirty-fourth principle: uncertainty can propagate without exploding
Sometimes a range of upstream values still produces the same downstream decision.
Use robustness tests before assuming every uncertainty destroys the answer.
Thirty-fifth principle: assumptions can be strategic in case studies
A recommendation may require assumptions about resources, compliance, demand or implementation.
State the assumptions most critical to success and identify how they would be monitored.
Thirty-sixth principle: recommendations should have assumption triggers
Write:
Proceed with A while condition X holds; if evidence shows Y, switch to B.
This converts an assumption into a monitored decision rule.
Thirty-seventh principle: essays contain assumptions about criteria
If a question asks whether something was “successful,” “important” or “effective,” define the criterion.
Different criteria can produce different judgements.
Make the evaluative assumption visible.
Thirty-eighth principle: define the criterion before weighing evidence
Success by short-term output may differ from success by long-term resilience.
Importance by scale may differ from importance by causal necessity.
The criterion is part of the argument architecture.
Thirty-ninth principle: assumptions can be challenged by counterevidence
Do not protect the assumption from the data.
If evidence contradicts what the assumption predicts, update or abandon it.
Fortieth principle: assumptions should earn continued use
A useful assumption simplifies while preserving the relevant structure.
If it repeatedly produces wrong predictions or contradictions, it is no longer helping.
Forty-first principle: timed exams need an assumption budget
Do not spend five minutes writing philosophical caveats around a routine model assumption.
State the assumption briefly, proceed, and reserve extended evaluation for assumptions that materially change the result.
Forty-second principle: identify the highest-impact assumption first
If several assumptions exist, ask which one could most change the answer.
Test that one first.
Forty-third principle: use sensitivity before speculation
Rather than imagining endless scenarios, vary the critical assumption within a plausible range and observe whether the conclusion changes.
Forty-fourth principle: robust answers need fewer caveats
If the conclusion survives reasonable assumption changes, say so.
Robustness is evidence.
Forty-fifth principle: fragile answers need conditional language
If the answer flips when one plausible premise changes, expose the dependency.
Do not present a fragile conclusion as universal.
Forty-sixth principle: an assumption can be the actual object of the question
Some advanced questions ask learners to evaluate a model, argument or experiment.
The hidden assumption may be exactly what the examiner wants identified.
Train assumption hunting as a deliberate reading mode.
Forty-seventh principle: ask what would happen if the assumption failed
This reveals the role the assumption plays.
If nothing changes, it may be unnecessary.
If everything changes, it is central.
Forty-eighth principle: assumption failure can reveal the next model
When a simplified model breaks, ask what variable or relationship must be added.
This is how advanced answers move from simple models to richer ones without discarding the useful core.
Forty-ninth principle: the best assumption is not the one that makes you right
It is the one most justified by the question, model, evidence and domain.
Outcome convenience is not