The CAT4 Test: What Schools Use It For and How Practice Helps

If you have searched for the CAT4 test in the last week, you have almost certainly been told two incompatible things. GL Assessment, the company that makes it, states plainly that it "strongly advises against any kind of practice ahead of a CAT4 test" and does not "recommend or endorse any CAT4 practice materials". Schools repeat that sentence word for word in their parent FAQs. Then the next six results are prep companies selling CAT4 practice packs, some of which quote that very line before taking your card details.

Both camps are talking past each other because they are describing two different situations. CAT4 has two jobs in British schools, and they pull in opposite directions. In the overwhelming majority of cases it is a diagnostic instrument, sat in school, used to build a picture of how a child learns. In a much smaller number of cases it is an admissions gate, used by grammar schools for in-year and late-transfer entry where the usual 11+ route is not available, with published score thresholds attached.

This guide separates the two, then gives you the piece of arithmetic almost no page supplies: GL's own published conversion from Standard Age Score to stanine to national percentile, so that a parent holding a report showing 112 can work out exactly where their child sits. We will also correct a few claims that have been copied around the internet until they hardened into fact.

Myth: CAT4 is an entrance exam. Reality: it is a diagnostic your school bought

GL Assessment describes CAT4 as "an assessment that is designed to help students and their teachers understand how they learn and what their academic potential might be". It is not a curriculum test. It does not sample what was taught last term. It measures reasoning across four domains and produces, in GL's words, "an individual profile of learning preference along with a range of indicators of likely future performance".

The ownership detail matters more than parents realise: only schools can buy CAT4. You cannot enter your child for it privately, you cannot obtain an official paper, and no third party can issue you a genuine CAT4 score. Anything sold to parents as a CAT4 practice test is an imitation of the format written by someone outside GL. That is not necessarily dishonest, but it is worth knowing what you are buying.

Scale gives the results their meaning. GL says CAT4 was "standardised on 25,000 students and verified every year based on analysis of a quarter of a million students", and that over three quarters of a million primary and secondary pupils sit it annually. Your child's score is not a mark out of a hundred. It is a position relative to a nationally representative sample of children born at roughly the same time.

Why your school runs it: the four jobs CAT4 does

Schools rarely explain the purpose in the letter home, which is part of why parents assume the worst. In practice CAT4 does four things at once.

• Baselining a new cohort. Many schools test at the start of the autumn term, often across several year groups at once. Keble Prep, for example, assesses Years 4 to 8 at the beginning of the autumn term. It gives teachers a common reference point for children arriving from many different primaries.
• Setting target grades. CAT4 reports carry indicators for Key Stage 2, 9 to 1 GCSE, A level and the IB. These feed directly into the targets a child is later shown in reports and parents' evenings.
• Flagging children who need support. A low reasoning profile, or a sharp mismatch between reasoning and attainment, is one of the earliest signals a school gets that something needs investigating.
• Finding hidden ability. A child whose written work is weak but whose non-verbal and spatial scores are high is a very different pupil from one who is simply behind, and CAT4 is designed to surface exactly that contrast.

Three of those four jobs only work if the score is honest. Hold that thought.

Inside the test: the four batteries

CAT4 assesses four reasoning types, which GL summarises as shapes and patterns, words, numbers and spatial ability.

Verbal reasoning measures thinking with language: relationships between words, categories, analogies and word-based logic. It is the battery most affected by reading fluency and vocabulary, which is why a child with English as an additional language, or with dyslexia, may show a verbal score well below their other three.

Non-verbal reasoning measures thinking with shapes and patterns, with no language dependency at all. Sequences, matrices, odd-one-out and analogies built from abstract figures. If you want a sense of the underlying skill, our guide to non-verbal reasoning question types walks through the families of pattern with worked examples.

Quantitative reasoning measures thinking with numbers, but it is not a maths paper. There is no long division, no fractions method to remember. It tests whether a child can spot the rule governing a numerical relationship.

Spatial ability measures thinking in three dimensions: mentally folding, rotating and comparing shapes. This is the battery with the least overlap with anything taught in class, and it is often where an unexpectedly strong result appears.

Only the last of those four is genuinely alien to the 11+ world. Verbal, non-verbal and quantitative reasoning are the same skill families that sit at the heart of most selective entrance tests, which is the honest reason parents keep bumping into CAT4 while researching secondary transfer.

How the test is sat: parts, timings and format

CAT4 covers students aged 6 to 17 and over. For ages 8 and above, the standard format is three tests of 45 minutes each, 2 hours 15 minutes in total, with the four batteries distributed across those parts. The Young Learner format for ages 6 to 8 is two tests of 30 minutes each. Schools can administer it digitally, on PC or tablet, or on paper, and at any point in the academic year, which is why sitting dates vary so much between schools.

Each section opens with introductory information and practice examples. GL is explicit that these introductions are not timed and may be repeated, while the test itself is timed and run under formal conditions. The Radclyffe School, explaining the arrangement to its own Year 7 parents, notes that a reader may support a child through the introductory sections but that timed sections must be answered independently, and that each test is "generously timed". Radclyffe also states that no additional time is permitted for pupils with SEND, on the grounds that it would compromise the reliability of the result. That policy is set by the assessment design, not the school, and it surprises many families.

One more correction worth making: CATs are set by GL Assessment, not by the government. There is no national CAT4 day, no published national results, and no league table consequence.

Reading the report: SAS, stanine, percentile and group rank

A CAT4 report gives you the same underlying result expressed four ways.

• Standard Age Score (SAS). GL defines it as the raw score "adjusted for age and placed on a scale that makes a comparison with a nationally representative sample of students of the same age across the UK. The average score is 100." The age adjustment matters: a summer-born child is compared with other summer-born children, not with the year group.
• National Percentile Rank (NPR). The proportion of same-age students who scored the same as or below your child.
• Stanine. A compression of the whole scale into nine bands, 1 low to 9 high.
• Group Rank (GR). Position within the cohort that sat it at your school. This is local, not national, and it is the number most likely to mislead you if your school's intake is unusual in either direction.

Myth: "100 is average, 120 is strong." Reality: here is the actual conversion

Every competing page asserts that 100 is average and that 120 or so is good, then stops. GL publishes the full conversion in its own quick data guide, and it is far more useful than the folklore version. Reading down from the top:

• SAS 127 and above: stanine 9, the 97th percentile and above, about 4 per cent of children. GL labels this band very high.
• SAS 119 to 126: stanine 8, percentiles 90 to 96, about 7 per cent.
• SAS 112 to 118: stanine 7, percentiles 78 to 89, about 12 per cent. Stanines 7 and 8 together are GL's above average band.
• SAS 104 to 111: stanine 6, percentiles 59 to 77, about 17 per cent.
• SAS 97 to 103: stanine 5, percentiles 41 to 58, about 20 per cent. Stanines 5 and 6 form the average band.
• SAS 89 to 96: stanine 4, percentiles 23 to 40, about 17 per cent.
• SAS 82 to 88: stanine 3, percentiles 12 to 22, about 12 per cent. Stanines 3 and 4 are below average.
• SAS 74 to 81: stanine 2, percentiles 5 to 11, about 7 per cent.
• SAS 73 and below: stanine 1, the 4th percentile and below, about 4 per cent. Stanines 1 and 2 are the very low band.

A worked example: the report that says 112

Suppose a report comes home with a verbal reasoning SAS of 112, and you have been told by a forum that "anything over 110 is fine". Walk it through properly.

112 is the bottom edge of stanine 7. That places the child at roughly the 78th percentile, meaning about 78 in every 100 same-age children nationally scored the same or lower, and around 22 scored higher. In a typical mixed-ability class of 30, that is somewhere around the sixth or seventh strongest verbal reasoner. Comfortably above average, clearly not in the top 4 per cent that a 127 would indicate.

Notice how much the single point matters at a band edge. A 111 sits in stanine 6, the middle of the average band, near the 59th to 77th percentile stretch. A 112 sits in stanine 7. The child has not changed. This is why treating a one or two point difference between two of your child's batteries as meaningful is usually a mistake, and why schools look at the pattern rather than the decimal.

The learning profile: verbal bias, spatial bias and the spiky result

GL builds a student profile by comparing the Verbal Reasoning and Spatial Ability batteries, sorting children into seven categories: mild, moderate or extreme bias towards verbal learning, the same three towards spatial learning, or an even profile across the two.

This is the part of the report that actually changes teaching, and it is the part parents skim. A child with a strong spatial bias and a weaker verbal score often thrives when a concept is diagrammed before it is described, and struggles when new content arrives purely as dense written text. A verbally biased child may be the reverse, articulate in discussion but slower to build a mental model of a shape or a mechanism.

A spiky profile, with one battery well clear of another, is information, not a defect. What it should trigger is a question: is the low battery a genuine reasoning difference, or is something masking it? A weak verbal score in a child who reads fluently in another language at home means one thing. The same score in a child with an undiagnosed processing difficulty means something else entirely, and the school's next step should be to find out which.

Myth: CAT4 is the 11+ under another name. Reality: they are different tests with opposite rules on practice

This is the most common confusion, and the clearest way to resolve it is to look at another GL product. Buckinghamshire's Secondary Transfer Test, the county's 11+, is also produced by GL Assessment. The council explains that it "measures verbal, non-verbal and mathematical skills", that children sit two papers of about 60 minutes each with a break between, and, crucially, that there is "also a shorter practice test, usually taken two days before the main test, to help children learn what to expect".

So here is the same assessment publisher, testing overlapping reasoning skills, in one case supplying familiarisation material as a matter of policy and in the other advising against practice entirely. The difference is not the skills. It is the purpose. A selection test is meant to rank children fairly, so everyone is given the same head start. A diagnostic is meant to describe a child accurately, so a head start is contamination.

If you are working out which selective test your child will actually sit, formats vary considerably by county and by school, and our parent guide to the 11+ and our comparison of GL, CEM and ISEB are better starting points than any CAT4 page.

When CAT4 becomes an admissions gate

The exception to everything above is real, and it is where prep sellers have a point that the purists ignore. Some grammar schools use CAT4 for in-year and late-transfer applications, where a child arrives outside the normal Year 7 entry round and no 11+ sitting is available.

Tonbridge Grammar School states that "all applicants are expected to sit pre-admission CAT4 tests to determine eligibility and written papers in English, Maths and Languages to assess suitability", and publishes an eligibility bar of at least 118 in each CAT4 test. The school also notes it is usually full in all year groups, with places arising mainly through family relocation. So the number is a threshold to be eligible, not an offer.

Weald of Kent Grammar School shows the other pattern. Its in-year applicants sit "a CAT4 test (quantitative, non-verbal and verbal reasoning)", three batteries rather than four, with spatial ability omitted, and the school publishes no numeric threshold at all. Applicants must instead achieve results "consistent with the standards of the current cohort".

Those two schools, both grammars, both in Kent, sit fifteen miles apart and run the process differently. That should permanently retire the prep-site claim that there is a universal CAT4 admissions score. There is not. Check the individual school's admissions page, and if it publishes nothing, ring and ask what they look for.

Myth: there is one right answer to "should we practise?" Reality: it depends which CAT4

If your child is sitting a school diagnostic CAT4, do almost nothing. GL's analogy to parents is an eye examination: practising the chart beforehand does not improve your eyesight, it just produces a wrong prescription, and "valuable information may be missed". A drilled score can raise a child's target grades beyond what is reasonable, or hide the very gap that would have unlocked support. The cost of inflation lands on your child, not on the school.

What is legitimate is removing avoidable friction, and it takes about twenty minutes:

• Tell your child what the test is and, more importantly, what it is not. No pass mark, no revision, nothing to fail.
• Explain that the questions get harder on purpose and that nobody finishes every item correctly.
• Make sure they know they can read the untimed introductory examples as many times as they like.
• Sleep, breakfast, glasses if they wear them. This is genuinely most of it.

If your child is sitting an admissions CAT4 for an in-year or late-transfer grammar place, the calculus changes. A published bar such as 118 in each battery is a gate, and pretending otherwise is not integrity, it is just unhelpful. There is no dishonesty in a child knowing what a matrix question or a cube-folding question looks like before they meet one under timed conditions. What still will not work is cramming: reasoning does not respond to a fortnight of drilling the way spelling does.

The sensible middle is steady work on the underlying skills over months, not a panic pack in the final week. Our guides to verbal reasoning question types cover the same families of thinking, and timed mock tests are useful mainly for teaching a child to keep moving when a question resists rather than for the score they produce.

What we would tell you tonight

If the letter says your school is running CAT4 next Tuesday, close the tab, buy nothing and let your child sit it cold. The report will be more useful to them because you did.

If you are chasing an in-year place at a school that uses CAT4 as an eligibility screen, find the number on that school's own admissions page, note whether they test three batteries or four, and build reasoning practice slowly across the terms you have. Those are different situations, and anyone giving you the same answer to both has not asked which one you are in.