Quick answer: Measurement is not always passive. Once a grade, ranking, target or examination becomes important, students, teachers, parents and institutions adapt behaviour around it. The measure then begins influencing the thing it was supposed to observe. This can be useful when the metric directs attention toward valuable capability. It becomes dangerous when people optimise the number while the underlying learning drifts away from what the number was meant to represent.
This article follows How an Examination Changes the Curriculum Before the Examination Happens, A Grade Is a Compression, and A Qualification Is a Signal.
A measure begins as a description. Once consequences attach to it, the measure can become part of the cause.
Why schools measure
Education systems need information.
- Does the student understand the topic?
- Is a programme working?
- Which learners need support?
- Has performance improved?
- Are standards comparable across schools?
Without measurement, institutions would be forced to act on impression alone.
Metrics therefore perform real and valuable work.
Consequences change the meaning of the metric
A low-stakes quiz can function mostly as feedback.
The same score becomes different when it determines admission, promotion, ranking, funding or reputation.
People naturally respond to incentives.
- Students practise what is tested.
- Teachers allocate more time to measured outcomes.
- Parents seek support in areas carrying high consequences.
- Institutions protect indicators used to compare them.
The metric is no longer merely reporting the system. It has entered the system.
Measure → Consequence → Behaviour Change → New Measured Reality
This can improve education
When a metric is well aligned with a valuable capability, attention to the metric can be productive.
If an assessment rewards clear explanation, students may practise explanation. If it requires transfer, surface memorisation becomes less sufficient. If schools monitor attendance, unnoticed absence may receive earlier attention.
Measurement can therefore direct energy toward real problems.
The danger is proxy drift
Most educational goals are too complex to observe directly in full.
We therefore use proxies.
- A test score stands in for some part of learning.
- A grade stands in for some part of performance.
- Attendance stands in for presence, not engagement.
- A qualification stands in for some part of capability.
The proxy becomes dangerous when optimisation improves the proxy faster than the underlying reality.
The number can improve while the thing we actually care about improves less—or even deteriorates.
Teaching to the test is one form of metric response
Preparing students for an examination is reasonable.
The problem appears when preparation begins exploiting predictable surface features instead of building the capability the examination intends to sample.
Students may become excellent at familiar templates but fragile when wording changes.
The score rises. Transfer does not.
That is a classic sign that the measure and the underlying capability have started separating.
A rank can change the person being ranked
Ranking does not merely report position.
It can change confidence, peer comparison, parental behaviour, subject choices and willingness to take risks.
The next performance therefore occurs in a learner who has now received information about their place in the group.
The ranking measures the learner, then becomes part of the environment acting on the learner.
Targets can create tunnel vision
When one target dominates, unmeasured goals can lose attention.
- Deep understanding may lose time to short-term score gains.
- Creative exploration may disappear near high-stakes periods.
- Students already near a threshold may receive more attention than those far above or below it.
- Easy-to-count outcomes may crowd out harder-to-measure qualities.
The issue is not malicious behaviour. Rational people can collectively create distortion simply by responding to the strongest incentive.
Gaming is the extreme case
At the extreme, people may learn to improve the metric without improving the underlying goal at all.
In education, that could include:
- memorising answer patterns without understanding;
- avoiding difficult students or tasks because they threaten averages;
- selecting only evidence that improves reported outcomes;
- overtraining a narrow assessed format.
Good systems therefore need more than a target. They need checks on whether the target remains connected to the real objective.
Multiple measures can reduce single-metric distortion
No single metric captures the full learner.
When decisions matter, a richer evidence stack may include:
- current score;
- component pattern;
- growth over time;
- student working;
- teacher observation;
- transfer to unfamiliar tasks;
- attendance and context where relevant.
Each measure has weaknesses. Together they can reduce the chance that one proxy becomes the entire reality.
Measures should have expiry dates in our minds
A result from last year can remain useful history.
It should not automatically override strong new evidence.
This is especially important when old measurements affect placement and opportunity.
A metric should become weaker evidence as the human changes and newer evidence accumulates.
Design the metric by starting with the human capability
A safer sequence is:
Capability → Evidence Needed → Measure → Consequence → Monitor Distortion
The dangerous sequence is reversed:
Available Number → Target → Optimise → Assume Capability
Schools need to remember why a measure exists before allowing it to dominate behaviour.
For parents
If a score improves, ask what improved underneath it. Was there stronger understanding, better retrieval, better pacing, or only greater familiarity with the format?
For teachers and tutors
Use metrics to direct attention, then keep checking whether students can perform when the surface changes. Transfer is one of the best safeguards against mistaking metric optimisation for learning.
For students
Learn how the examination works, but periodically ask: “If the question looked unfamiliar, would I still know what to do?”
The deeper principle
A good measure remains answerable to the reality it represents. The moment the number becomes more important than the capability, education risks teaching people how to win the metric instead of how to become capable.