The learner knows the material.
The result changes when the clock starts.
Assessment condition sensitivity is the degree to which an observed result changes when relevant testing conditions—such as time, language, fatigue, environment, support or administration—change.
This is the third pillar in the second Assessment quartet beneath How Assessment Works. It owns one interpretation question: how much of this score belongs to capability, and how much belongs to the conditions under which capability had to appear?
Quick Read
Assessment performance is always produced somewhere, sometime, under specific constraints. Those conditions may be part of the intended construct or an irrelevant barrier. A timed examination legitimately measures performance under time pressure if that is part of the purpose. The same time limit may contaminate an assessment intended only to probe conceptual understanding. Strong interpretation therefore separates target conditions from incidental ones, compares performance when a condition changes, and narrows the claim when the result is highly condition-sensitive.
Conditions Are Not Automatically Bias
Some constraints belong to the real task.
A live oral presentation includes time, audience and pressure. Removing all of them would measure a different capability.
The question is whether the condition belongs to the construct we intend to infer.
Change One Condition to Learn More
A student scores poorly under time.
Give a comparable untimed task.
If performance rises sharply, the new evidence does not erase the timed result. It reveals a condition dependency that the original percentage hid.
Language Can Be a Condition or Part of the Target
In an English comprehension assessment, language demand is central.
In a science task intended to measure scientific reasoning, unnecessarily complex language may create an additional access barrier.
Support Changes the Meaning of the Result
With prompts, the result may represent supported capability.
Without prompts, it may represent independent capability.
Both can be useful. They answer different questions.
A Practical Condition-Sensitivity Check
CLAIM: understands algebraic method CONDITION A: timed mixed paper → 45% CONDITION B: untimed parallel items → 82% CONDITION C: oral explanation → strong method selection INTERPRETATION: conceptual route appears stronger than timed score alone suggests; inspect retrieval latency and execution under time.
Final Thought: A Score Carries Its Conditions With It
Assessment becomes more honest when the result is interpreted as performance under defined conditions, not as a context-free reading of the learner.
ASSESSMENT · SECOND FOUR PILLAR LEGS
Return to How Assessment Works, or continue through Response-Process Evidence, Score Aggregation Loss and Diagnostic Resolution.