One student improves.
What does that prove?
Claim–evidence distance is the conceptual distance between what was directly observed and the broader conclusion drawn from it.
This is the second pillar beneath How Evidence Works. The master owns claim size broadly. This page owns the overreach problem: how far can a conclusion travel before new evidence is required?
Quick Read
A real observation can be strong evidence for a small claim and weak evidence for a larger one. One successful case can establish possibility, not prevalence. A before–after improvement can show change, not necessarily causation. A controlled study can support an average causal effect, but not guarantee the same outcome for every person. Claim–evidence distance grows when we move from observed case to population, association to cause, short term to long term, one setting to another, behaviour to mechanism, or average effect to individual prediction. Strong evidence practice either gathers the additional evidence needed for the larger claim or shrinks the wording back to what the route actually supports.
Possibility Is Not Frequency
One case proves something can happen.
It does not tell us how often it happens.
Change Is Not Automatically Cause
A learner improves after tuition.
Possible explanations include teaching, practice, maturation, changed effort, easier assessment or several together.
How Causation Works owns causal reasoning broadly.
Average Effects Do Not Determine Every Individual
A study may show an average benefit.
Individuals can differ because of baseline ability, implementation, context, dose or other conditions.
Short-Term Success Is Not Long-Term Retention
Immediate post-test performance can support learning under current conditions.
Long-term retention needs delayed evidence.
Familiar Performance Is Not Transfer
Success on rehearsed questions does not automatically establish capability on unfamiliar ones.
Changed tasks are needed to cross that distance.
Observed Outcome Is Not Mechanism
Two methods produce the same result.
They may do so through different mechanisms.
Mechanistic claims require evidence about the pathway, not merely the endpoint.
One Setting Is Not Every Setting
A result from one school, population or period may not travel unchanged elsewhere.
The fourth sibling, Transportability Boundary, owns that travel problem.
Evidence Conflict Can Reveal an Overwide Claim
Two credible studies disagree.
Sometimes the problem is not that one is wrong. The original claim may have ignored conditions under which the effect changes.
The third sibling, Evidence Conflict, owns that diagnostic process.
A Practical Distance Ladder
- Observation: this happened.
- Possibility: this can happen.
- Pattern: this happens repeatedly here.
- Frequency: this happens at roughly this rate.
- Association: these variables move together.
- Causation: changing X changes Y.
- Mechanism: this pathway explains how.
- Transport: the relationship persists elsewhere.
- Prediction: it applies to this new individual or case.
A 20-Lens Claim–Evidence Distance Audit
- What was directly observed?
- What exact claim is being made?
- Does the claim add frequency?
- Does it add causation?
- Does it add mechanism?
- Does it add duration?
- Does it add transfer?
- Does it add population generality?
- Does it add individual prediction?
- What comparison supports each added step?
- What alternative remains?
- What measurement uncertainty travels upward?
- What context was fixed?
- What context changed?
- Is one case being universalised?
- Is an average being applied deterministically?
- What new evidence would justify the next step?
- Can the claim be narrowed instead?
- Would the wording change if uncertainty were explicit?
- Does the final conclusion stay within the distance the evidence can carry?
Final Thought: Evidence Has a Carrying Distance
The further a conclusion travels beyond the observation, the more new structure it must earn. When the evidence runs out, shrink the claim or build another bridge.
EVIDENCE · FOUR PILLAR LEGS
Return to How Evidence Works, or continue through Evidence Dependency, Evidence Conflict and Transportability Boundary. Return to the How X Works Hub.