Application process
Civil Service Judgement Test: Format, Scoring and How to Pass It
Published
The Civil Service Judgement Test (CSJT) is an untimed online situational judgement test built on the Civil Service Behaviours selected for the vacancy. Two parts: a self-assessment worth 15% of the score, then three scenarios per behaviour, each with four actions you rate independently from Counterproductive to Effective. You pass by rating actions the way the behaviour descriptors do.
That last sentence is the whole method. The scoring key was written by people who know what each behaviour rewards at each grade: ownership, reasoning, proportionate escalation, honesty, and prioritisation rather than heroics. A candidate who has internalised the nine behaviours and their grade descriptors rates actions the same way the key does, without having seen the key.
The format, part by part
| Part | What you do | Scored how |
|---|---|---|
| Part 1: self-assessment | Answer questions about your ambition, motivation and flexibility | 15% of the overall score; may be discussed at interview |
| Part 2: scenarios | Three workplace scenarios per selected behaviour, as text or video. Each has four possible actions | Each action rated independently as Counterproductive, Ineffective, Fairly Effective or Effective, and compared with the scoring key |
The scenarios are written for the grade band you applied for (AA/AO, EO, HEO/SEO or G7), and you only see your own band. Which behaviours appear depends on the advert: the recruiter chooses the behaviours for the vacancy and the test draws its scenarios from those. Gov.uk does not publish the total number of scenarios, and the figures quoted on preparation sites are not official.
The test is one of eight that an advert can require. Where it sits among the others, and which of them are bankable, is covered in the guide to Civil Service online tests.
How the test is scored
Your responses are compared with a representative comparison group and reported as a percentile: a 44th percentile score means you did better than 44% of that group. There is a system-set minimum standard per grade level. After the application deadline, the recruiter reviews all scores and sets the pass mark, which can be higher than the minimum, depending on how many people they can take through and the impact on protected groups. Meeting the minimum does not guarantee an invitation to the next stage. No numeric pass mark is published anywhere official.
Four consequences follow from that.
- The score is not bankable. Verbal, Numerical and Work Strengths passes can be reused within a grade grouping for six months. The Judgement Test cannot, because different behaviours are selected for each vacancy. Expect to sit it on every application that requires it.
- There is no per-question feedback. You receive your percentile and automatically generated feedback in the application centre, but not which actions you misrated, for security reasons.
- The output includes a report for interviewers, with suggested follow-up questions. Your test performance can shape what the panel asks, and the panel marks your interview answers against the same descriptors on the 1 to 7 behaviour scale.
- Part 1 may be discussed at interview. Whatever you say about your ambition, motivation and flexibility in the self-assessment, say it honestly and expect to be consistent with it in the room.
Why rating each action independently changes your approach
Most situational judgement tests ask you to pick the best and worst option, or to rank all four. The CSJT does not. Each action gets its own rating on the four-point scale, and each rating is compared with the key on its own.
So two actions in the same scenario can both be Effective, and all four can be weak. If you approach a scenario by hunting for "the right answer" and then grading the other three downwards to make it stand out, you will misrate actions that the key marks as perfectly good. The question for each action is not "is this the best of the four?" but "would someone demonstrating this behaviour well, at this grade, do this?"
That reframing also protects you from a common trap: assuming the key wants the most dramatic intervention. It does not. It wants the proportionate one.
A worked scenario of the kind the test uses
The scenario below is illustrative, not a real test item, and the ratings are a reading of the Making Effective Decisions descriptors at EO level rather than an official key. At EO, the behaviour rewards timely, fair decisions within your own casework and the rules, with reasoning you can explain, and working out a sensible answer when guidance does not quite cover the situation. The full descriptor set and worked interview answers are in the Making Effective Decisions guide.
You are an Executive Officer processing applications. A colleague is off sick and you have picked up one of their cases, which must be decided today. One supporting document is dated after the cut-off in the guidance, but a covering letter says the original was lost and this is a reissued copy. Your manager is in meetings until late afternoon.
| Action | Rating | Why |
|---|---|---|
| Approve the application, because the covering letter explains the discrepancy and the applicant has clearly done their best | Ineffective | A decision on sympathy rather than evidence or rules. Nothing has been checked. It is also not obviously yours to make without the facts |
| Check the guidance and the case notes for how reissued documents are treated. If it is covered, apply it. If not, put the case on hold with a note and send your manager a specific question to answer before the end of the day | Effective | Facts first, then a decision within your authority, then escalation of a precise question rather than a raw problem. Ownership of the deadline throughout |
| Leave the case for your colleague to pick up on their return, since it was theirs and they know the history | Counterproductive | The deadline passes, the applicant is left waiting, and responsibility is pushed onto someone who is unwell. Ignoring the issue is the pattern the key punishes hardest |
| Phone the applicant to confirm when and why the document was reissued, record the answer on the file, and decide against the guidance once you have the facts | Effective | Also facts first, also a decision at your level, also recorded. A different route to the same standard of judgement |
Two actions rated Effective is not a mistake. It is the point. A candidate who ranks the four will demote one of the two good options to Fairly Effective to create an order, and lose the mark.
FailHow a candidate who ranks the options reasons
"The phone call is the best one because it resolves it fastest, so that's Effective. Checking the guidance and messaging the manager is good but slower, so Fairly Effective. Approving is wrong but at least it's decisive, so Ineffective. Leaving it is the worst, Counterproductive." Three of the four ratings match the key, and the one that does not is the mark lost by ranking. The candidate has rated the actions against each other, not against the behaviour.
PassHow a candidate who thinks in behaviour terms reasons
"What does Making Effective Decisions at EO reward here? A fair decision, within the rules, based on facts I have checked, made at my level, with the deadline owned. Phoning the applicant does all of that: Effective. Checking the guidance and escalating a specific question also does all of that: Effective. Approving on the letter alone skips the checking and decides on feeling: Ineffective. Leaving it abandons the decision and the deadline entirely: Counterproductive." Four of four. The candidate never compared the actions to each other.
The patterns that score low, and the ones that score high
Across scenarios and behaviours, the same shapes recur. Learn these and most items become legible on first reading.
| Reliably rated Counterproductive or Ineffective | Reliably rated Effective | The principle underneath |
|---|---|---|
| Ignoring the issue, or waiting for it to resolve itself | Taking the proportionate next step yourself | Ownership at your level |
| Escalating everything upward, unfiltered | Escalating a specific question with the facts attached | Escalation is proportionate, not reflexive |
| Acting outside your authority: approving, promising, overriding | Checking what is yours to decide, then deciding it | Decisions belong at the right level |
| Blaming a colleague, the system or the customer | Keeping the people affected informed, including the awkward message | Working Together, and honesty |
| Heroics: staying late to do everything regardless | Reprioritising, and saying plainly what will slip and why | Delivering at Pace rewards prioritisation over effort |
| Deciding on an assumption or on sympathy | Checking the facts first | Evidence before action |
| Quietly fixing a mistake and saying nothing | Flagging the error and owning the fix | Civil Service values |
Two of these catch strong candidates. The first is escalation: people who would be excellent in the role still rate "tell my manager immediately" as Effective in every scenario, because it feels safe. The key treats escalating a raw problem, with no facts gathered and no proposal attached, as weak from EO upward. The second is heroics. Absorbing the whole burden yourself reads as commitment, but the descriptors reward prioritising and being honest about what will slip, and an action that keeps colleagues informed will usually outrate one that quietly carries the load. The Working Together descriptors make this explicit from EO up.
Part 1: the self-assessment
Part 1 asks about your ambition, motivation and flexibility and is worth 15% of your score. Gov.uk does not publish the response scale or the number of questions, and claims about them elsewhere are unverified.
Treat it as the strengths element of Success Profiles rather than as a puzzle. Strengths assessment is built around what you genuinely do and enjoy, which is why it is hard to script, and why gov.uk warns that Part 1 may be discussed at interview. A self-assessment that paints an ambitious, highly flexible candidate will be tested by a panel whose strengths questions are designed to detect natural energy. Answer it as the candidate you are, and keep your interview preparation consistent with what you said.
Before you launch it
The test is untimed, but gov.uk suggests two to four minutes per scenario and recommends allowing at least an hour. Sit it in one quiet session where you can. It closes after 25 minutes of inactivity and can be relaunched, but a relaunch is an interruption to your concentration rather than a fresh attempt.
The deadline for completing the test is the same as the application deadline. Raise any technical query at least two working days before it. A larger screen is recommended over a phone, and if the scenarios are video, subtitles, BSL and transcripts are available.
If you need a reasonable adjustment, say so when asked at the application stage. The recruiter confirms it. Supporting documents may be requested, but gov.uk is clear that you should still ask.
Do the official practice test first. It is linked from the gov.uk online tests page, it is unscored, and it is shorter than the real test but looks exactly like it. There is no better rehearsal of the format, and no other practice material carries that guarantee.
If you fail, or score lower than you expected
A score below the pass mark means the application stops there. You can sit the test again in any future application that requires it, and since the score is not bankable anyway, every application is a clean attempt. Hiring managers may also require a supervised retake for candidates invited to interview, so the standard you reach at home should be one you can repeat in a room.
A low percentile is diagnostic even without per-question feedback. It says your model of what the behaviours reward is miscalibrated, usually in one of two directions: rating escalation and effort too highly, or rating independent action too highly where the grade expects a check first. The fix is not more practice questions. It is reading the descriptors for the behaviours and grade on the advert until you can say, for any workplace situation, what ownership, evidence and proportion look like at that level. The same reading prepares you for the interview, which is marked against the same text.
What the test is not
It is not a knowledge test. No specialist knowledge or experience is needed, there is no advantage from qualifications, and nothing in it rewards knowing how the Civil Service is organised. Everything you need to rate an action is in the scenario and in your understanding of the behaviours.
It is also not a test with a published pass mark, a published question count or a published number of attempts. Where a preparation site quotes those figures, it is guessing, and gov.uk's own guidance for the current test is the only authoritative description of the format. An older candidate guide from 2017 describes a previous, competency-based version and does not apply.
Next: the behaviours on your advert, read at your grade, are the scoring key you are allowed to see. Start there.
Common questions
An untimed online situational judgement test built on the Civil Service Behaviours chosen for the vacancy. It has a short self-assessment worth 15% of the score, then scenarios in which you rate four possible actions independently as Counterproductive, Ineffective, Fairly Effective or Effective. You only see scenarios written for the grade you applied for.
Gov.uk does not publish one. Your answers are compared with a comparison group and reported as a percentile. A system-set minimum standard applies per grade, and after the deadline the recruiter can raise the bar depending on how many candidates they can take forward. Figures quoted on preparation sites are not official.
No. The test is untimed, though gov.uk suggests two to four minutes per scenario and recommends allowing at least an hour. It closes after 25 minutes of inactivity but can be relaunched, and the deadline for completing it is the same as the application deadline.
No. Unlike the Verbal, Numerical and Work Strengths tests, the Judgement Test is not bankable, because different behaviours are selected for each vacancy. A pass on one application does not carry to the next, so you sit it again each time an advert requires it.
Your application does not progress, and you can sit the test again in any future application that requires it. You receive a percentile and automatically generated feedback in the application centre, but no feedback on individual questions. A low score is a signal to revisit the behaviour descriptors at your grade.
Yes. Part 1, which asks about ambition, motivation and flexibility, is worth 15% of your overall score, and gov.uk says it may be discussed at interview. Answer it honestly: a panel comparing your interview answers with your self-assessment will notice if the two do not match.
Yes. Gov.uk links an unscored practice test from its Civil Service online tests page. It is shorter than the real test but looks exactly like it, which makes it the only reliable way to see the format before you sit the scored version.
Practise before the real thing
Reading how panels score is step one. The candidates who pass are the ones who practised saying their answers out loud — and got scored feedback against the official Civil Service Behaviours framework before interview day.
Start Practising FreeStart a practice session3 free practice questions · No card needed · Cancel anytime