Parents arrive at the subject of intelligence testing from very different directions. Some are responding to a school suggestion, some are worried that a bright child is drifting, some are trying to make sense of a report that already exists, and some are simply curious. The formats and the interpretation rules differ enough from adult assessment that a little background prevents a great deal of misreading.
In UK practice, formal cognitive assessment of a child is nearly always instrumental rather than descriptive. It happens because a decision needs supporting evidence, not because anyone wants a number for its own sake.
The common triggers are identifiable. A child is struggling in one area while performing normally elsewhere, and the question is whether a specific learning difficulty is involved. A child is achieving far below what teachers observe informally, and the gap needs explaining. Support is being considered and the local authority requires evidence. Or a child is markedly ahead of the curriculum and the school is deciding how to respond.
In each case the useful output is not the composite figure but the pattern beneath it. A profile with strong verbal comprehension, average visual reasoning and unusually weak working memory says something actionable. A single number says almost nothing.
The single most important idea in child assessment is that raw performance is meaningless without an age reference. Reasoning capacity, memory span and processing speed all develop rapidly through childhood, so the same raw total means entirely different things at seven and at thirteen.
Every scored assessment therefore compares a child only against peers in a narrow age band, often as tight as three or four months in the younger ranges. A score of 100 means performance typical for that specific age band, not typical for children in general.
This is also why the historical formula that gave the measure its name has been abandoned. The original quotient divided a mental age by a chronological age and multiplied by a hundred, which behaves reasonably in childhood and falls apart in adulthood. Modern instruments use deviation scoring — position within an age-matched distribution — for every age group, which is why the same scale can describe a six-year-old and a sixty-year-old coherently.
Professional assessment in the UK draws on a small number of established batteries, chosen by age.
The Wechsler Intelligence Scale for Children covers roughly six to sixteen and is the workhorse of school-age assessment. It produces index scores for verbal comprehension, visual spatial reasoning, fluid reasoning, working memory and processing speed, which is precisely the profile detail that makes it useful. Its preschool counterpart handles younger children with formats adapted for shorter attention spans and limited literacy.
The British Ability Scales are widely used by educational psychologists in the UK and are normed on a British sample, which matters for interpretation. Stanford-Binet remains in circulation, particularly where a very wide range needs covering. Non-verbal instruments built entirely from abstract pattern items are used where language would confound the result — for children with limited English, hearing impairment or significant speech difficulties.
Sessions are one-to-one, run by a qualified psychologist, and typically last an hour to ninety minutes with breaks. Much of the value comes from what the examiner observes: how the child approaches difficulty, whether they persist or disengage, whether performance drops when tasks are timed.
A large number of free web assessments are marketed for children, and their quality varies enormously. Most are not normed on children at all — they apply an adult scale and adjust it crudely, which produces figures that are essentially decorative.
The general limits described in the material on unsupervised online assessment apply with extra force here. A child's performance is far more sensitive to fatigue, mood, hunger and distraction than an adult's, and a home sitting controls none of these. A child who does not understand an instruction will often guess rather than ask, and nobody is present to notice.
None of this makes them worthless as an activity. Pattern puzzles are enjoyable, and working through a well-built iq test together can be a perfectly good afternoon and a fair introduction to how such items are constructed. The same sensible conditions that apply to any self-administered sitting — one uninterrupted block, a proper screen, no distractions — matter even more with a child. The problem is only ever the interpretation. Treat the output as a game score, not as information about the child.
A proper report contains far more than a headline number, and the surrounding detail is where the value sits.
A markedly uneven profile is often more informative than a high or low composite. Strong reasoning alongside weak working memory, for instance, describes a child who understands concepts but loses multi-step instructions — a pattern with clear classroom implications and clear remedies.
Less stable than most people assume, particularly in the early years. Assessment before school age correlates only loosely with later performance, because the abilities being measured are still consolidating and because young children are highly variable from session to session.
Stability improves substantially through primary school and is reasonably good from around ten onwards, though movement of several points between sittings remains entirely ordinary. This is also the age from which high-IQ societies begin accepting candidates, and the standards involved in supervised society assessment reflect that same threshold of reliability. A result from age five should never be treated as a permanent characteristic, and a single session at any age is a snapshot taken under one set of conditions. The broader discussion of measurement error and confidence intervals applies to children with even wider margins than it does to adults.
Growth also matters. Children who believe ability is fixed respond very differently to difficulty from those who see it as improvable, and that difference shows up in performance. The material on beliefs about ability covers the mechanism, and the related work on attention and self-regulation explains why the same child can produce noticeably different results on different days.
Assessment is a tool for answering a specific question, and there are situations where it answers nothing useful while introducing costs of its own.
Testing a child purely to satisfy parental curiosity rarely helps anyone. The number changes nothing about what happens next, and once it exists it tends to shape expectations in ways that are difficult to unwind. Testing during a period of upheaval — a bereavement, a house move, a difficult transition between schools — will produce a result that reflects the upheaval rather than the child. Testing a child who is exhausted, unwell or resistant to the whole idea produces the same problem in a shorter timeframe.
There is also a timing question. If a school has raised a concern but no decision depends on the answer yet, waiting a term and gathering classroom evidence often produces better information than a session would. Assessment works best when a specific question has already been framed and the result will determine a specific response.
How results are framed matters more than the results themselves. A child told they are gifted often becomes cautious, avoiding tasks that risk contradicting the label. A child told they scored poorly may conclude that effort is pointless.
The safer framing is descriptive and specific. The assessment looked at particular kinds of thinking, it found some areas easier than others, and it suggested some things that will help at school. That is accurate, it is useful, and it leaves the child's sense of their own capacity open rather than fixed.
Avoid sharing the composite figure with a young child at all. It carries no information they can act on and considerable weight they cannot yet handle.
Cognitive batteries measure reasoning efficiency across a defined set of tasks. They do not measure curiosity, persistence, kindness, practical skill, artistic capacity or the ability to work with other people — and those qualities predict adult outcomes at least as strongly.
They also miss divergent and creative thinking almost entirely, since every item has one correct answer. A child who generates unusual solutions rather than the expected one may score below their actual capability. Anyone commissioning an assessment should be clear about what question it answers, and equally clear about what it leaves untouched. The broader case for structured problem-solving skills exists precisely because those skills are trainable in a way that a composite score is not.
Used for the purpose it was built for — identifying specific strengths and difficulties so that support can be targeted — cognitive assessment of children is genuinely valuable. Used as a verdict on a child's potential, it is both unreliable and unkind.
Continue with the rest of the series on assessment and reasoning: