The science, explained
Glossary: the words behind cognitive assessment
The terms our science pages use, each in a line or two of plain words, with a link to the page that explains it properly.
What a cognitive assessment measures
The words for what a cognitive assessment looks at, before any score is given.
- Cognition and learning
Cognition is how a person takes in, holds, works with and stores information: the mental tools under every task on a course or at work. "Cognition and learning" is also one of the four broad areas of need in the SEND code of practice, which is why sector staff already use the phrase.
- Cognitive abilities (the Cattell-Horn-Carroll model)
What the research calls the parts of cognition, mapped in the Cattell-Horn-Carroll (CHC) model, the internationally recognised map. The domains are built against that model.
- Cognitive domains
The areas an assessment measures and scores one by one, such as reading or remembering.
What a score means
The words for who a score is compared with, the scale it sits on and why every learner takes it the same way.
- Age norms
A learner's score is compared with people of the same age, not with everyone, because what is typical at 17 is not what is typical at 45.
- Below average and low average
The exam boards' names for a score of 84 or below and a score of 85 to 89. They describe a score, never a learner.
- Index scores
An index score pools the scores of several related domains into one, so it leans on no single task. The cognitive assessment gives three (Language & Numbers, Visual & Fluid, and Memory): the index scores say whether support may be needed; the domain scores say what kind.
- Inter-rater reliability
Whether two people marking the same work give it the same mark. When a fixed rule does the scoring, there's only one marker.
- Norm sample
The group of people every score is compared with, like the children on the growth chart. Also called the normative sample or the standardisation sample.
- Norms
A score means nothing on its own; it means something when it is compared with people like the learner. That comparison group is the norm sample, and its size and make-up decide whether the comparison is fair.
- Percentile
The share of people of the same age who score lower. A score of 84 is at the 14th percentile: 14 in every 100 people that age score lower.
- Standard deviation
How much scores tend to vary. When it is small, there is little variation: most people score close to the average. When it is large, scores vary a lot and are spread out well above and below it. On the standard score scale it is 15 points, so the typical range, 85 to 115, is one standard deviation either side of 100.
- Standard scores (standardised scores)
Most people score between 85 and 115 for their age: that is the typical range, with 100 in the middle. A score of 84 is the first score below the typical range.
- Standardisation
Standardisation means every learner gets the same instructions, the same practice, the same tasks and timing, and the same scoring, so a difference in score comes from the learner, not from who gave the assessment or who marked it.
- Standardised administration and scoring
The same instructions, practice, tasks and timing for everyone (the administration), and the same rule turning answers into a score (the scoring).
Whether you can trust a score
The words for the evidence behind a score, and the ones to ask any assessment about, ours included.
- Construct validity (internal structure)
The tasks group together the way the theory of how people think says they should. Tasks measuring related abilities move together; unrelated ones don't.
- Criterion or predictive validity
Scores link to an outcome that matters to you, such as whether a learner completes their programme.
- High-stakes benchmark (.80)
The British Psychological Society's standard for an assessment whose scores are used to make decisions about people: a reliability of .80 or above.
If the learner took it again, would they get a similar score?
- Internal consistency
The tasks that make up a score agree with each other, so the score isn't resting on one task.
If the learner took it again, would they get a similar score?
- Known-groups validity
Groups of learners that research says should score differently on a particular task do, and score close to typical everywhere else.
- Margin of error
No score is exact. Every score carries a little give either way, so two sittings rarely match point for point.
If the learner took it again, would they get a similar score?
- Performance validity (engagement checks)
Performance validity checks, built into each task, catch a learner who is clicking through, not engaging or interrupted. A score that fails its check is set aside rather than trusted.
- Performance validity tests (PVTs)
The checks themselves, built into the tasks. It's the term a procurement form or a psychologist may use.
- Reliability
Reliability is how consistent a score is: a learner who took the same assessment again a few weeks later should get much the same result. The more reliable a score, the smaller its margin of error, and the more safely you can act on it.
If the learner took it again, would they get a similar score?
- Test-retest reliability
The same learners, assessed again weeks later, get much the same results.
If the learner took it again, would they get a similar score?
- Validity
Validity is the evidence that an assessment measures what it claims to, for learners like yours. It is shown three ways: the tasks group together the way the theory says they should, the right learners score differently, and the scores link to outcomes that matter.