UK Civil Service numerical test: format and preparation
The Civil Service Numerical Test asks you to do calculations and to interpret numerical information, presented as graphs, tables or other figures with several answers to choose from. GOV.UK describes it as untimed and adaptive, with no skipping and no going back. This page explains those conditions, works one original question in full, and opens free untimed practice.
Reviewed
Testonome is independent. It is not affiliated with, endorsed by or sponsored by the organisations whose tests this page describes.
What the numerical test asks you to do
Two things: whether you can do the calculations, and whether you can evaluate and interpret what the figures actually mean. The second is where the difficulty usually sits, because it decides which figure you divide by, which unit the question wants, and which direction a change runs in.
Knowing the format is worth something, and it is not what decides your answer. Your own method is. So the question further down is there to be tried before you read anything about it, and each of the four answers it offers is the result of a method somebody really uses. Whichever one you pick, this page names the step that produced it and opens free untimed practice in that step, in our numerical section.
How the test runs, and what that changes
| What you want to know | What the guidance says |
|---|---|
| Timing | Not timed. Most people take 15 to 45 minutes, and how long you take does not change your score. GOV.UK recommends allowing at least an hour for each one. |
| Length | Varies, and typically shorter than a fixed-length test. No question count is published. |
| Difficulty | Adapts as you go: the next question may get harder after a right answer, easier after a wrong one. |
| Moving through it | Answer every question. No skipping and no going back. |
| Calculator | You may want rough paper, a pen and a calculator to hand. |
| Your score | A percentile against a representative group who have also taken the test. |
| Pass mark | A minimum standard applies for the job level. |
| After the deadline | The recruiter looks at all scores and may raise that pass mark. |
| Practice | Practice tests are linked from the guidance, and are not scored. |
Four of those change what is worth preparing.
No clock means no target time, and the application deadline still runs. GOV.UK says your score is not affected by how long you take. So do not rehearse against a pace you invented, and do not carry one over from a timed test you sat elsewhere. Work at the speed that keeps you accurate. Plan the sitting all the same: GOV.UK recommends allowing at least an hour for each test. It also warns that if the application deadline passes before you finish, your test is not stopped, but your application has missed the submission deadline. So start early enough to finish before that deadline, rather than in the last minutes before it.
No going back means checking happens before you submit. Once an answer is in, it is in. The habit worth drilling is the one that runs before you commit: read which figure the question asks about, check the unit, and check the direction of the change.
The length varies, so you cannot tell how far through you are. The Cabinet Office publishes a transparency record for these tests: they use a computer adaptive algorithm, which fits the questions to the candidate and asks fewer of them than a fixed-length test would. There is no published count to work down from, so treat each question as the task in front of you rather than as a fraction of a set.
A percentile is a comparison, not a mark out of ten. Your score is set against a representative group who have also taken the test. A minimum standard applies for the job level first, and after the deadline the recruiter looks at all applicant scores and may raise that pass mark. GOV.UK is explicit that meeting the minimum is no guarantee of an invitation to the next stage, and no result on this site predicts any of those figures.
Your invitation email arrives with full instructions on how to proceed, and GOV.UK sets out the general arrangements around them. If the test goes wrong for you, or you need a reasonable adjustment, GOV.UK points you at your recruitment team, whose contact details are in the job advert, or at its own contact form. It adds that there is a legal duty to provide reasonable adjustments, so asking is the expected thing to do.
Try this question before you read the answer
One question cannot tell you how you would score. What it can show you is which method you reach for when nothing is at stake. This one is written for Testonome and it drills a single step: choosing the figure you divide by. It is not a Civil Service question, and it does not stand in for the rest of what the test asks.
| Month | Orders |
|---|---|
| March | 250 |
| April | 320 |
Practice question
By what percentage did the supplier's orders rise from March to April?
Independent practice written for Testonome, not Civil Service material. Nothing you decide here is recorded.
Check your answer
B, 28%. Take the difference first: 320 − 250 = 70 orders. Divide by March, the month the change started from: 70 ÷ 250 = 0.28. Read that as a percentage: 0.28 × 100 = 28 percent, a rise.
The rule underneath it is short: a change is the difference divided by the figure the change started from. The other three options come from using the wrong base, stopping at the wrong unit, or comparing April with March instead of measuring the move between them.
- A, 21.9%: you divided by April, 70 ÷ 320. The size of the change is right and the base is wrong, because a change is measured against where it started.
- C, 70: the change in orders. Correct arithmetic and the wrong unit, because the question asks for a percentage.
- D, 128%: April measured against March, 320 ÷ 250. It answers how big April is, not how far the orders moved.
What your answer tells you, and what to practise next
You answered B and can say why the base is March. That is the step this question tests, and you have it on this one. A single answer is not yet a habit, so the next move is the same either way: try it on questions you have not seen.
You answered A, C or D, or you were not sure. Each of those three comes from one identifiable step, and the explanation above names which. Work on that step rather than on everything.
You guessed. Landing on B by guessing tells you nothing about your method, and it is worth saying that plainly rather than counting it as a pass. Read the working above, then practise, where every question carries the same kind of explanation.
One drill answers this question, and two others are different skills that come up in the same tests. All three are free and untimed, written for Testonome; none is an official Civil Service test or a mock of one.
- Percentage change is the drill for this question. A, C and D all come from it, so start here if any of them looked right.
- Ratio sharing is a separate skill, not a diagnosis of anything above: splitting a total between parts, where the usual slip is dividing before counting the parts.
- Weighted average is separate too: averaging groups of different sizes, where ignoring the sizes can give the wrong figure.
The familiarisation test GOV.UK publishes is a different kind of preparation and worth doing as well. It is linked from the guidance for candidates, which tells you to take it beforehand. It is not scored, and what it shows you is the screen: how a question is put, how you answer it, and what happens when you move on. GOV.UK also notes it is the way to check that assistive technology works before the day.
The button below opens the percentage-change drill on this site: five original questions, untimed and free, each with the same kind of worked explanation you have just read.