CAT4 scores explained: how to read your child's report

A CAT4 report gives your child four scores, one for each kind of reasoning: verbal (with words), quantitative (with numbers), non-verbal (with shapes and patterns) and spatial (imagining shapes in the mind). It also gives a mean scoreThe average of your child's four CAT4 scores (add them and divide by 4): one number for the whole test. The school's report may call it the mean, the mean SAS or the overall score.. The mean score is the average of the four, so it sums up the whole test in one number.

On each of the four scores, 100 is exactly average for your child's age. About half of all children score between 89 and 111, and GL Assessment, the company that makes CAT4, calls that range average. A score of 112 or more is above average. Only 4 children in 100 score 127 or more.

The mean score works a little differently. 100 is still average, and so is any mean score from 89 to 111. But fewer children score far from 100 than on each of the four scores. So more children have a mean score in that range: about 6 in 10, by our calculation from GL's figures, not about half.

No score is exact. GL's own figures show that, for most children, their real level is within about 8 to 10 points of each of the four scores, nine times in ten. So a small difference between two of your child's scores may be just chance. We would treat a difference of 10 points or more as probably real: GL's figures show that a difference of 11 points is real nine times in ten.

If one score is lower than the others, first find it in the table of scores: any score from 89 to 111 is average. Then compare it with each of the other three. A gap of 10 points or more is probably real: your child is stronger in the kind of reasoning with the higher score.

So a lower score from 89 to 111 is still average for your child's age. It is lower than your child's other scores, but not low for a child of that age. For example, a verbal score of 94 and a spatial score of 112 are 18 points apart: the 94 is average, and your child is probably stronger at imagining shapes than at reasoning with words. Ask the school how it will use that strength.

A score of 88 or less is below average: ask the school whether it plans any help. We explain more under two scores far apart and a verbal score below the others.

The numbers on a CAT4 report

In this guide we use one example: Sam, a made-up child in Year 7The first year of secondary school in England and in British schools abroad, for children aged 11 to 12. Year 6 is the last year of primary school (age 10 to 11).. Year 7 is the first year of secondary school in England and in British schools abroad, when children are 11 to 12 years old. This is Sam's report, in the form most schools see it.

GL calls each of the four parts of the test a batteryGL's word for one of the four parts of CAT4: verbal (words), quantitative (numbers), non-verbal (shapes and patterns) and spatial (imagining shapes in the mind). Each part has two short tests.. The word is not used much outside CAT4. When you read "the verbal battery", it just means the verbal part of the test.

Standard Age Score (SAS)

The Standard Age ScoreThe Standard Age Score is the main CAT4 score, which compares your child with children of the same age in years and months. is the main number. It compares your child with children across the UK of the same age, counted in years and months. 100 is the middle. Sam's spatial score of 112 means Sam did better than most children of the same age on the spatial tests.

Because each score is set for your child's age, a child born in August is not compared with a child born the September before. So check the date of birth on the report. If it is wrong, tell the school, because every score depends on it.

Percentile, or NPR

A percentileHow many children in 100 of the same age scored the same as your child or lower. A percentile of 84 means your child did as well as or better than 84 in 100. tells you how many children in 100 of the same age scored the same as your child or lower. On the school's report it is called the NPRThe name for the percentile on the school's CAT4 report: how many children in 100 of the same age scored the same as your child or lower., which stands for national percentile rank. Sam's verbal percentile is 34. That means Sam scored the same as or better than 34 children in every 100 of the same age. A percentile is not the share of questions your child got right.

Stanine (ST)

A stanineOne of nine bands that group CAT4 scores, from 1 (lowest) to 9 (highest). Band 5 is the middle. A score of 115 is in band 7. is a band. GL puts all scores into nine bands, from 1 (the lowest) to 9 (the highest), and band 5 is the middle. Bands 4, 5 and 6 are average. Sam's spatial score of 112 is in stanine 7, which is above average. The table in the next section shows which scores fall in each band.

Group rank (GR)

Some reports also give a group rankWhere your child came among the pupils the school compared them with, often a class or the whole year. It says nothing about how your child compares across the country.. This is where your child came among the pupils the school compared them with, often your child's class, sometimes the whole year. It says nothing about how your child compares with children across the country. For that, use the SAS.

The mean score

The mean scoreThe average of your child's four CAT4 scores (add them and divide by 4): one number for the whole test. The school's report may call it the mean, the mean SAS or the overall score. is the average of the four scores. Sam's four scores add up to 415, and 415 divided by 4 is about 104. It is a fair summary, but it hides the differences between the four parts. Sam's mean of 104 looks like an ordinary average score. Sam's verbal and spatial scores, 18 points apart, tell you much more.

Which report will you get?

Schools get the full numbers. A school may also send a report for parents.

The report for parents describes each score in words, as below average, average or above average. On GL's sample report for parents, a yellow band marks the average range, which is 89 to 111 in GL's table. If you would like the numbers, ask the school. GL advises schools to talk results through with parents, not just send them home.

Our CAT4 glossary explains the words on both reports, and in the school's letter, one by one.

Is your child's score good? Find it in this table

This is GL's table of the nine bands, or staninesOne of nine bands that group CAT4 scores, from 1 (lowest) to 9 (highest). Band 5 is the middle. A score of 115 is in band 7..^{1GL Assessment, Quick data guide: the bands, their names and the share of children in each. We worked out the example column from the scale GL describes (average 100, spread 15).} We added the last column: an example score from each band, and how many children in 100 of the same age it beats or equals.

So, in plain numbers:

• 100 is the exact middle. Half of children of the same age score below it and half above. It does not mean your child got half the questions right.
• 115 is a good score. It is above average, in stanine 7, and it beats or equals about 84 children in 100.
• 130 is a very high score. Only 2 or 3 children in 100 score that high.
• 89 to 111 is average. About 54 children in 100 score in this range. You may read elsewhere that 85 to 115 is average. By our calculation, that range holds about 68 children in 100, so it is wider than GL's. On a CAT4 report, GL's own average range is 89 to 111.

Numbers in 100 can be hard to imagine, so here is the same table as a class of 30 children, with a typical mix from across the UK.

In a class like this, 16 children have an average score, about 7 score 112 or more, and 1 scores 127 or more. We rounded GL's numbers to whole children. In a selective schoolA school that chooses its pupils by a test or by their results, rather than taking every child who applies., a school that chooses its pupils by a test, the class looks very different. Most pupils there score above 100, so a score of 105 can be near the bottom of the class and still be above the national average.

CAT4 scores and SATs scores are different

Many children in England also have SATsNational tests at the end of primary school in England (Year 6, age 10 to 11), in reading, maths, and grammar, punctuation and spelling. 100 is the expected standard. results. SATs are national tests at the end of primary school, taken in Year 6 at age 10 to 11. SATs scores also sit near 100, so the two are easy to mix up.

In SATs, 100 is the expected standard, and most children reach it. In 2026, by the first figures published, the average SATs score was 106 in reading and 105 in maths. In CAT4, 100 is the average itself. So a SATs score of 105 and a CAT4 score of 100 are both about average. A child with those two results is about average on both, so the lower number does not mean the child has slipped.

How exact is a score?

Less exact than a single number suggests. Every test score includes some luck: a good guess, a moment of poor focus, or a question that happened to suit your child. GL has measured how much the scores move because of this.^{2GL Assessment, CAT4 UK technical report, section on the standard error of measurement. GL calls a child's real level the "true score", and gives the range that holds it two times in three: 5.0 points for verbal, 5.2 quantitative, 5.4 non-verbal, 6.0 spatial and 3.4 for the mean. We multiplied each by 1.645 to get the range for nine times in ten, which is the one drawn here.}

Think of your child's real level as the score they would get on average if they could sit the test many times. For the verbal score, the real level is within about 8 points of the reported score, nine times out of ten. So when a report shows a verbal score of 100, the real level is very likely somewhere from 92 to 108. The spatial score is a little less exact, about 10 points either way. The mean scoreThe average of your child's four CAT4 scores (add them and divide by 4): one number for the whole test. The school's report may call it the mean, the mean SAS or the overall score. is more exact, about 6 points either way.

This range is the margin of errorHow far a score could be from your child's real level, because no test is exact. For a verbal score of 100, the real level is very likely between 92 and 108. GL calls this range a confidence band.. The teacher's report draws it as a bar next to each score, which GL calls a confidence band. Here are Sam's scores with their bars. The shaded area is GL's average range, 89 to 111.

Look at how much the bars overlap. Sam's quantitative score (106) and non-verbal score (103) could easily swap places on another day. Sam's verbal and spatial bars touch only at their ends, so that difference is very likely real.

When is a difference between two scores real?

GL gives two guides. Its technical report says that when two scores are 11 points apart, the difference is real nine times out of ten. Its advice for teachers calls a verbal score 10 or more points below the others "significant". Here is how we would read the difference between any two of your child's scores:

Sam's quantitative and non-verbal scores are 3 points apart, so treat them as the same. Sam's verbal and spatial scores are 18 points apart, so that difference is very likely real.

Scores change between tests

GL suggests that schools give CAT4 again in Year 9 or Year 10, when children are 13 to 15. GL tested 3,883 pupils with the Year 7The first year of secondary school in England and in British schools abroad, for children aged 11 to 12. Year 6 is the last year of primary school (age 10 to 11). version of CAT4, and again with the version for Years 9 and 10 two years later. For 62 in 100 of them, the mean scoreThe average of your child's four CAT4 scores (add them and divide by 4): one number for the whole test. The school's report may call it the mean, the mean SAS or the overall score. moved by 5 points or less. For 90 in 100, it moved by 10 points or less.

Single scores move more. GL also gives figures for one part of the test between two sittings: about 15 pupils in 100 gain more than 9 points, and about 15 in 100 lose more than 9. So if your child's Year 9 score is 4 points below the Year 7 score, that is probably just normal movement.

The profile: words against shapes

The report also gives your child a profileA comparison of two CAT4 scores, verbal and spatial. The report for parents describes it in words; the school's report gives it a label, such as 'no bias' or 'mild spatial bias'.. The profile compares only two scores: verbal (reasoning with words) and spatial (reasoning with shapes and images in the mind). GL says these two can be seen as opposite ends of one range of ability. There are seven profiles:

• No bias (about even). The verbal and spatial scores are less than two staninesOne of nine bands that group CAT4 scores, from 1 (lowest) to 9 (highest). Band 5 is the middle. A score of 115 is in band 7. apart. About 66 children in 100 have this profile, so it is by far the most common.
• Mild, moderate or extreme verbal bias. The verbal score is the stronger of the two. The bigger the difference, the stronger the label.
• Mild, moderate or extreme spatial bias. The spatial score is the stronger of the two, in the same way.

Sam's verbal score is in stanine 4 and the spatial score in stanine 7, three bands apart. So Sam has a spatial bias. GL expects children with a spatial bias to enjoy working with images, diagrams, maps and 3D objects. They may feel less sure of themselves in subjects built on language, such as English.

Use the profile with care, for two reasons.

First, a profile is not a "learning style". GL's report for parents calls it your child's "learning preference". But GL's own guide for teachers says the idea of individual learning styles is wrong, and that spatial ability "is not a kind of learning style".^{3GL Assessment, CAT4 and strategies for learning, pages 7 and 90.} A major review of the research agrees: it found almost no good evidence that people learn better when taught in a "preferred" way. So read the profile as a sign of where your child is strongest, and where some help might be useful.

Second, the profile often changes. GL's guide says that over two years, a child with a verbal or spatial bias will probably not keep the same profile. Only the "no bias" profile tends to stay the same.

Predicted grades: "most likely" and "if challenged"

Many reports show predicted gradesThe grades that children with similar CAT4 scores got, on average, in later national exams. They are a guide, not a promise.. GL calls them indicators. They show the results that children with similar CAT4 scores got in later national exams: SATsNational tests at the end of primary school in England (Year 6, age 10 to 11), in reading, maths, and grammar, punctuation and spelling. 100 is the expected standard. at the end of primary school, or GCSEsThe exams most pupils in England take at age 16, one per subject, graded from 9 (highest) to 1. Older reports may show letter grades, from A* down.. GCSEs are the exams pupils in England take at age 16, one for each subject, graded from 9 (the highest) to 1. For GCSEs, GL used the results of about 91,000 students from 2019.

You will often see two columns:

• "Most likely" is the average result of children with your child's scores. It is not a promise. GL gives an example. Children with a mean scoreThe average of your child's four CAT4 scores (add them and divide by 4): one number for the whole test. The school's report may call it the mean, the mean SAS or the overall score. of 90 get, on average, 99 in their SATs maths test. But about half of these children get less than 99.
• "If challenged" is the result that the top quarter of those children reach. In the same example, 25 children in 100 get 103 or more.

Even "most likely" hides a wide spread. Take GL's GCSE example of a child with a mean score of 100. In maths, that child has a 52 in 100 chance of grade 4 or below. The chance of grade 5 or above is 48 in 100. So the two results are almost equally likely.

The predicted grades assume that your child makes average progress in a typical secondary school. GL tells teachers to explain to pupils that the grades are a prediction from numbers, not a fixed future. If your child's school set target grades from CAT4, ask how they were set and how often they are reviewed.

What the scores can change for your child

A CAT4 score is not a pass or a fail. But schools use the scores to make decisions about your child, so the numbers matter. These are the main ones.

Sets

A setA class for one subject, such as maths, where pupils are grouped by how well they are doing: a top set, a middle set and so on. Some schools move pupils between sets during the year. is a class for one subject, such as maths, where pupils are grouped by how well they are doing. Some schools use CAT4 to help decide sets, and some use it together with SATsNational tests at the end of primary school in England (Year 6, age 10 to 11), in reading, maths, and grammar, punctuation and spelling. 100 is the expected standard. results. A child often stays in their first set. In one school that GL describes, only 4 or 5 pupils in 100 changed set when the school checked again eight weeks later.

Research in England found that nearly a third of pupils were in a higher or lower maths set than their SATs results would suggest. Compared with classes where pupils of all levels learn together, sets help pupils in the higher sets a little and hold back pupils in the lower sets a little, on average. The evidence for this is very limited. Pupils in lower sets may also lose confidence in the subject.^{4Education Endowment Foundation, Setting and streaming: "a small negative impact on low attaining learners, and a small positive impact for higher attaining pupils", from evidence it rates as very limited. Connolly and others (2019), The misallocation of students to academic sets in maths, British Educational Research Journal 45.}

Target grades

Some schools use CAT4 scores, with other information, to set targets. GL also describes using its predicted gradesThe grades that children with similar CAT4 scores got, on average, in later national exams. They are a guide, not a promise. to set targets. Then a higher score can mean a higher target, and this can shape what teachers expect of your child. GL says low scores "should never be used to put a ceiling on expectations". In plain words: a low score should never limit what teachers think your child can do.

A gifted and talented register

Some schools keep a list of very able pupils, called a gifted and talented registerA school's list of pupils it sees as very able, who may be given extra challenge. Each school decides who is on it., and give them extra challenge. GL's advice is that a score of 127 or more on any one of the four parts is evidence that a child may be gifted, even if the mean scoreThe average of your child's four CAT4 scores (add them and divide by 4): one number for the whole test. The school's report may call it the mean, the mean SAS or the overall score. is lower.^{5GL Assessment, CAT4 and strategies for learning: "a score in stanine 9 (an SAS of 127 or above and in the top 4% of the age group) on any one of the four CAT4 batteries provides evidence to suggest that the student is gifted".} Each school still decides who is on its own list.

A school place

Some schools use CAT4 to choose pupils. Ask the school what score it looks for, and whether it uses the mean score or the four separate scores.

What you can do

If a score looks too low to you, say so. You know your child better than one test does. Ask the school how it uses the scores, and when it will look at sets or targets again.

If your child will sit CAT4 again, for example for a place at a new school, you may want them to look at the types of question first. GL, which makes CAT4 for schools, advises against any practice: our guide to the test gives its advice in its own words.

Research on similar tests, mostly with teenagers and adults, finds that taking a test a second time raises scores by about 3 points on CAT4's scale, on average. That was a test given and marked by the people running it, with no coaching in between.^{6Hausknecht, Halpert, Di Paolo and Moriarty Gerrard (2007), Retesting in selection, Journal of Applied Psychology 92 (authors' manuscript): 0.21 of a standard deviation for a retest without coaching, for the scores people got (Table 4; 75 samples, 81,374 people; across all 107 groups in the review, the average age was about 19). The review's text gives 0.24, after a correction for errors of measurement; CAT4's standard deviation of 15 is for the scores children get, so we use 0.21. We multiplied it by 15: about 3 points (3.15). That conversion is ours.} That research did not measure practice at home, so it does not tell you whether practice at home changes a score, or by how much. Our CAT4 practice test explains this research, with 24 questions and a detailed explanation for each answer.

When the report looks wrong

A low verbal score

The non-verbal and spatial tests use no words or numbers. The verbal tests depend on reading. So a child who finds reading hard, or who is still learning English, can do well on the shape tests and less well on the verbal tests.

For a child still learning English, a verbal score well below the other three is "quite typical", in GL's words. GL's advice for teachers adds that predicted gradesThe grades that children with similar CAT4 scores got, on average, in later national exams. They are a guide, not a promise. based on all four scores are then likely to be too low. GL also recommends a reading test when the verbal score is in stanineOne of nine bands that group CAT4 scores, from 1 (lowest) to 9 (highest). Band 5 is the middle. A score of 115 is in band 7. 3 or lower, which means a score of 88 or less. Ask the school whether one is planned. At home, in our view, the best help is reading together and talking about words and their meanings.

Two scores far apart

If two scores are 10 points or more apart, the difference is probably real. Neither score is wrong: your child is probably stronger in one kind of reasoning than in the other.

A mean scoreThe average of your child's four CAT4 scores (add them and divide by 4): one number for the whole test. The school's report may call it the mean, the mean SAS or the overall score. can hide this. A mean of 100 can come from four scores of 100, or from scores of 115, 115, 85 and 85. GL advises teachers to use a child's strengths to help with the weaker areas. Ask how the school will do that.

CAT4 lower than SATs or schoolwork

First, compare the two scales correctly. A SATsNational tests at the end of primary school in England (Year 6, age 10 to 11), in reading, maths, and grammar, punctuation and spelling. 100 is the expected standard. score of 105 is about the national average. A CAT4 score of 105 is a little above CAT4's average of 100.

Then think about the day. CAT4 is a set of timed tests, often given in the first weeks at a new school. Illness, nerves or an instruction read wrongly can pull a score down. Tell the school what you know: GL expects teachers to use CAT4 together with other information about your child.

Many questions left blank

The school's report shows how many questions your child answered in each part. If you have only the report for parents, ask the school for these numbers. If your child left many blank, GL's first reading is that your child's reasoning in that area is probably at a low level. The score may still be right, but GL says it cannot be relied on with confidence. GL suggests testing again after at least six months, which may show more clearly what your child can do.

CAT4 higher than schoolwork

GL says CAT4 can find a very able child who may have been missed, or who is at risk of not doing as well at school as they could. In our view, there are many possible reasons, from a difficulty with reading or writing to being bored in class. Ask the school what it thinks the difference means.

Very high or very low scores

CAT4 comes in versions for different ages, which GL calls levelsThe version of CAT4 for an age group, named by a letter: in England and Wales, usually D for Year 7, E for Year 8, F for Years 9 and 10. Each child is compared with children of the same age.. For most levels, the highest possible score is 141. A child who reaches it may be able to do more than the test can show. Ask whether the school offers extra challenge.

For a very low score, first check how many questions your child answered (see "Many questions left blank" above). And whatever the reason, GL is clear that a low score should never limit what teachers expect of a child.

Talking to your child about the scores

If your child asks about the numbers, start with their best score. Every child has one. Avoid labels such as "clever" or "not a maths person". Scores change between tests, and labels tend to stay.

Tell your child that the scores say nothing about effort, which is the part of school they control. With a younger child, it is fine to keep the numbers to yourself.

What to ask the school

• Can I see the scores for each of the four parts, and talk them through with a teacher?
• How many questions did my child answer in each part?
• Is my child's date of birth correct on the report?
• What will the school do with the results: sets, target grades, extra help, extra challenge?
• When will sets and targets be looked at again?
• Is the difference between any two scores large enough to act on?
• If the verbal score is low, is a reading test planned?

Frequently asked questions

What is the highest CAT4 score? 141, for most levels of the test. The levels for the youngest children, X and Y, go from 69 to 131.

What is the average CAT4 score? 100 on every part. The scores are set so that the typical child of each age scores 100.

Is 130 gifted? It can be. For GL, a score of 127 or more on any one part is evidence that a child may be gifted. By our calculation from GL's scale, only 2 or 3 children in 100 score 130 or more. Whether your child goes on a gifted list is still the school's decision.

Is CAT4 an IQ test? Not in GL's words. GL describes CAT4 as a test of "developed abilities", and says that staninesOne of nine bands that group CAT4 scores, from 1 (lowest) to 9 (highest). Band 5 is the middle. A score of 115 is in band 7. avoid the wrong idea that the scores are "IQ scores".

But CAT4 is a reasoning test, and researchers use tests like it to measure general intelligence. One large study in England used an earlier version of CAT in this way. Our guide to the CAT4 test explains how CAT4 scores relate to later exam results.

When do results come back? On screen, the tests are scored as soon as your child finishes. Paper tests are scored within about 10 working days of reaching GL. GL's guidance allows up to 15 in September and October, its busiest months. Whether, and when, parents see the results is the school's decision.

Can my child sit CAT4 again to improve a score? Only if the school arranges it. If a child left many questions blank or got very few right, GL suggests testing again after at least six months, because the score may not be reliable. GL also suggests testing again in Year 9 or 10. If the score was for a school place, ask that school whether it allows a second try.

Our guide to the CAT4 test explains the eight types of question and what happens on the day.

All our CAT4 guides

Practice questions for your child's year

Worked examples and questions of each type, every answer explained. Each level page also has a PDF of its questions, with the answers. In Year 7 (Grade 6), we suggest starting with the Level D page, then the 24 questions for more practice.

• Year 7, ·Grade 6, Level D
• 24 free questions, all eight types, Years 6 and 7, ·Grades 5 and 6
• Year 8, ·Grade 7, Level E
• Years 9 and 10, ·Grades 8 and 9, Level F
• Other years have no page yet. Our guide to the CAT4 test shows one example of each type of question. For Years 2 and 3 (Grades 1 and 2), we suggest games, not test papers.

About CAT4

• The CAT4 test, What your child will sit, the test day, and how schools use results
• CAT4 in the UAE, Which grades take it, at which level, and whether it affects admission
• CAT4 in Ireland, 6th class: which level, and why it cannot decide a school place

Your child's scores

• CAT4 scores explained, What a score of 100, 115 or 130 means, and what to ask the school, This page
• CAT4 glossary, SAS, stanine, NPR and the other words on reports and school letters

Further reading

• GL Assessment, CAT4: a quick guide to the data. The company's own short guide to the scores on the report, with the table of the nine bands.
• Pashler, McDaniel, Rohrer and Bjork (2008), Learning styles: concepts and evidence. Four researchers looked for studies showing that children learn better when teaching matches their "learning style". They found almost none of the right kind, and the few they found did not support the idea.

Notes

1. GL Assessment, Quick data guide: the bands, their names and the share of children in each. We worked out the example column from the scale GL describes (average 100, spread 15). ↩

2. GL Assessment, CAT4 UK technical report, section on the standard error of measurement. GL calls a child's real level the "true score", and gives the range that holds it two times in three: 5.0 points for verbal, 5.2 quantitative, 5.4 non-verbal, 6.0 spatial and 3.4 for the mean. We multiplied each by 1.645 to get the range for nine times in ten, which is the one drawn here. ↩

3. GL Assessment, CAT4 and strategies for learning, pages 7 and 90. ↩

4. Education Endowment Foundation, Setting and streaming: "a small negative impact on low attaining learners, and a small positive impact for higher attaining pupils", from evidence it rates as very limited. Connolly and others (2019), The misallocation of students to academic sets in maths, British Educational Research Journal 45. ↩

5. GL Assessment, CAT4 and strategies for learning: "a score in stanine 9 (an SAS of 127 or above and in the top 4% of the age group) on any one of the four CAT4 batteries provides evidence to suggest that the student is gifted". ↩

6. Hausknecht, Halpert, Di Paolo and Moriarty Gerrard (2007), Retesting in selection, Journal of Applied Psychology 92 (authors' manuscript): 0.21 of a standard deviation for a retest without coaching, for the scores people got (Table 4; 75 samples, 81,374 people; across all 107 groups in the review, the average age was about 19). The review's text gives 0.24, after a correction for errors of measurement; CAT4's standard deviation of 15 is for the scores children get, so we use 0.21. We multiplied it by 15: about 3 points (3.15). That conversion is ours. ↩