A headline says:
Students who sleep less get lower grades.
A student rewrites it:
Lack of sleep causes poor grades.
Perhaps.
But the second sentence has turned a relationship into a cause.
Causal language should be proportional to causal evidence.
Association and causation are different claims
An association tells us that two variables are related in observed data.
Causation says that changing one variable helps produce a change in the other.
The second claim is stronger.
A familiar third-variable problem
Ice-cream sales and swimming accidents may both rise in hot months.
It would be wrong to conclude:
Buying ice cream causes swimming accidents.
Hot weather can increase both.
The association is real. The proposed causal story is not.
Use language that preserves what the evidence actually shows
Associational phrases include:
is associated with
is related to
is linked with
correlates with
These keep the relationship visible without pretending the causal mechanism is settled.
Causal verbs carry heavier commitments
Compare:
Heavy late-night phone use is associated with shorter sleep.
Heavy late-night phone use reduces sleep.
The second sentence says the phone use actively produces the reduction.
That requires stronger evidence.
Current research publishing guidance makes the distinction explicit
Major medical journals distinguish causal wording from associational wording according to study design.
This is not only a research-writing issue. It is a precision-English issue.
Correlation does not establish which causal story is true
An association may arise because:
- X causes Y;
- Y causes X;
- a third variable Z affects both;
- the pattern is partly due to chance or measurement.
The slogan “correlation does not imply causation” should not be misread as “correlation can never reflect causation”.
Reverse causation matters
If poor health is associated with use of a certain product, perhaps the product affects health.
But perhaps people with poor health are more likely to choose the product.
The causal arrow may point the other way.
Contributes to is useful in multi-causal systems
School performance can be shaped by:
sleep
prior knowledge
practice
stress
attendance
health
teaching
Writing:
Stress causes poor performance.
may be too broad.
Better:
High stress can contribute to poorer performance when it disrupts concentration, sleep or retrieval.
The phrase allows several causes to coexist.
Prediction is not automatically cause
A variable may predict an outcome because it is an indicator of another underlying process.
Students should not assume:
predicts = causes.
After is not because
Scores fell after the new timetable was introduced.
This establishes sequence.
It does not automatically establish:
The timetable caused the fall.
Temporal order is evidence, but not sufficient proof.
Mechanism strengthens explanation
Weak:
Noise causes poor concentration.
Stronger:
Continuous unpredictable noise can reduce concentration by repeatedly capturing attention and forcing the learner to reorient to the task.
The second sentence explains how the effect could occur.
Experiments can justify stronger causal inference
When researchers manipulate a variable, control relevant conditions and compare outcomes appropriately, causal claims can become better supported than in simple observation.
The language consequence is:
better causal design can justify stronger causal wording.
Cause language travels across subjects
Science:
increased temperature can increase evaporation under relevant conditions
History:
economic pressure contributed to political unrest
Geography:
prolonged rainfall can trigger landslides on vulnerable slopes
English:
repeated first-person pronouns can create a more personal tone
Different disciplines tolerate different causal strengths.
A useful causal ladder
Association:
associated with, linked to
Contribution:
may contribute to, can influence, is one factor in
Stronger causal relation:
leads to, results in, causes
Context and evidence decide where the sentence belongs.
A practical causal audit
- What is the evidence?
- Could the direction be reversed?
- Could a third factor explain both?
- Is there a plausible mechanism?
- Does my wording match the evidence strength?
Parent-friendly example
Claim:
Tuition improves grades.
Ask:
Always?
Under what conditions?
Refined:
Targeted tuition can improve performance when it repairs a genuine learning gap and the student transfers the learning into independent work.
Hilary Mantel as a high-level lens
Large historical outcomes rarely have one clean cause. Institutions, incentives, relationships, information and timing interact.
The transferable lesson is multi-causal restraint, not imitation.
AI and causal overreach
AI systems can turn association into causation, sequence into cause or a plausible mechanism into established fact.
A useful audit is:
Identify every causal claim. For each, state whether the evidence establishes causation, association or only a plausible mechanism.
Connections eduKateAI can learn
association → observed relationship
causation → directional production claim
third variable → possible confounding
reverse causation → causal arrow may run opposite direction
temporal sequence → not automatically cause
causal wording strength → evidence requirement
The public principle is that the strongest sentence is not the sentence with the strongest verb. It is the one whose verb says exactly what the evidence can support.
Related eduKateSG resources
Research basis
This article draws on current causal-language guidance from major research journals, experimental-reasoning resources and recent education research on correlation and causation.