Situational Judgement Test: Format, Examples and Preparation

A situational judgement test, often shortened to SJT, is a pre-employment assessment that asks how you would respond to realistic workplace situations. Instead of testing whether you can calculate a percentage or identify a pattern in a diagram, an SJT asks whether you can make a sensible, role-appropriate decision when something happens at work. You may be shown a short scenario about a frustrated customer, a disagreement with a colleague, a deadline conflict, a safety issue, a manager asking for help, or an ethical problem. Then you are asked to choose, rank, rate, or compare possible responses.

Candidates often find SJTs confusing because the questions can look subjective. In a numerical test, there is usually one answer. In a situational judgement test, several options may be partly reasonable. The challenge is to understand what the employer is measuring: judgement, prioritisation, teamwork, communication, integrity, customer focus, leadership, accountability, or problem-solving. The strongest answer is usually the one that handles the work problem constructively, follows the role’s responsibilities, avoids unnecessary escalation, and protects the organisation’s standards.

Every other test in a hiring process has an answer key. A situational judgement test usually does not. What it has instead is a profile: a set of ratings gathered from people already performing well in the role, or agreed by a panel working from the employer's competency framework. Your response is scored by how far it sits from that profile, not by whether it is right in any absolute sense.

That one structural fact undoes most of the advice written about these tests. There is frequently no single correct option, two options can both be good answers, and the option that feels most decisive can be the one the framework marks down because it steps outside the remit of the role you applied for. It also means the rubric is not secret: the competencies being measured are usually printed in the job advert, and reading them properly is worth more than another hundred scenarios worked blind.

The response formats sitting under one label

Situational judgement testing is a family of formats, not one question type. The scenario text looks similar across all of them; the response mechanics do not, and they change strategy completely.

Rate each response for effectiveness

You see one scenario and four to six possible responses, and you rate every response on its own scale, commonly anchored from counterproductive to highly effective across a range of 4 to 7 points. The mechanical point is that each rating is independent. Nothing forces you to separate the options, so you can legitimately rate two responses at the top and two in the middle. Scoring is by distance from the keyed rating for that response, which means a rating one point away from the key earns most of the available credit and a rating three points away earns little or none.

Choose the most effective and the least effective response

Here the test forces discrimination. You must name one option as best and one as worst, and the two choices are usually scored as separate marks. The least effective half is where candidates lose most of their marks, because they read the scenario hunting for the good answer and then treat "least" as whatever is left over. There is normally no partial credit on either half: picking the option the key rates second worst scores the same as picking the best one.

Rank all the responses in order

The hardest variant, because what a near miss costs depends on the distance metric behind the scoring. Some versions count how many pairs you placed in the correct relative order, some sum how far each option sits from its keyed position, and some award credit only for the correct top and bottom. Swapping two adjacent middle options can cost almost nothing under one metric and a meaningful share of the item under another.

Pick the single best response

Conventional multiple choice with one keyed answer. It is the fastest format to work through and the coarsest to be scored on: all or nothing, no partial credit, no mark for spotting the worst option, and one misreading loses the item.

Single scenarios against branching multi-stage scenarios

A single-shot item gives you one scenario, one question, and no consequences. A branching item gives you a scenario, takes your answer, and uses it to generate the next stage, so the situation you face at stage two is a product of what you chose at stage one. You cannot usually go back and revise. Some branching tests score each stage separately; others score the path as a whole, which means an early choice you would call minor is carried forward into everything that follows.

Scoring and timing, and what they mean for strategy

Formula scoring is rare in situational judgement testing. The dominant model is distance from key with partial credit, sometimes described as near-miss scoring: your answer is compared with the expert consensus rating and you earn credit in proportion to how close you were, rather than only on an exact match.

That has a direct consequence for guessing. Because an unanswered item scores nothing at all and a close answer scores something, leaving items blank is the worst available outcome. On a rating scale, committing to your best estimate beats skipping, and if you are torn between adjacent points the cost of picking the wrong one is small. If an instruction screen does tell you that marks are deducted for wrong answers and nothing is deducted for unanswered ones, that is formula scoring and the logic reverses: wrong answers are penalised and blanks are not, so guessing at the buzzer lowers your score. Read the instructions rather than assuming.

Timing models vary more than scoring does. Many situational judgement tests are untimed or timed generously, on the reasoning that the construct is judgement rather than speed. Some publishers apply a section timer across the whole set of scenarios, which lets you spend longer on a branching item and less on a short one. Per-question timers appear mostly in video and animated formats, where the scenario plays and a response window opens after it, and there is no way to bank time from an earlier item. Adaptive difficulty barely exists here, because scenarios do not sit on a single difficulty continuum an algorithm can climb.

A worked example

You are one of four people producing a client report due Friday. On Wednesday the colleague who owns the data section tells you privately that they are well behind, and asks you not to mention it to the project lead because they intend to catch up overnight.

A. Agree to say nothing and quietly take on part of the section yourself that evening. B. Tell your colleague you will hold off until Thursday morning, and if the section is not ready you will raise it with the project lead together. C. Email the project lead straight away, saying your colleague is behind and asked you to keep it quiet. D. Say nothing and concentrate on your own sections, since that one is your colleague's responsibility.

Most effective: B. Least effective: D.

B is keyed highest because it does two things at once: it leaves ownership with the colleague and gives them a real chance to recover, whilst capping how long the delivery risk stays hidden from the person accountable for it. Competency frameworks reward that pairing of support and risk control.

Each distractor is a real principle applied in the wrong place. A misapplies collegial support: the instinct is right, but absorbing the work silently means the lead still cannot plan, and the risk stays concealed rather than managed. C misapplies escalation: the lead does need to know about delivery risk, but raising it before your colleague has any chance to act, framed as reporting a person rather than a risk, is escalation at the wrong point. D misapplies ownership boundaries: the deliverable is shared, so treating a known risk to a client deadline as somebody else's problem is the only option involving no action at all, which is why it keys as least effective rather than weak.

How to prepare for this type specifically

Start with the competency framework, not with practice scenarios. If the advert lists behaviours such as collaboration, commercial awareness or resilience, those headings are the marking rubric, and the keyed answer to any scenario will be the option that demonstrates the most of them at once. Write down what each behaviour would look like in a concrete workplace decision before you practise anything.

Build a two-question habit for every option: is this within the remit of the role I applied for, and does it address the actual risk or only the awkwardness. Those checks separate keyed answers from plausible distractors better than general reasoning.

Practise the least effective half deliberately. It carries its own mark, it is where most marks leak away, and picking the worst option is a genuinely different judgement from picking the best.

If you are facing rating scales, decide in advance what your 1 and your top point actually mean. Candidates who have not calibrated cluster everything in the middle, and middle ratings are the ones that sit furthest from a key that uses the full range.

Answer at the level of the job, not your current one. Graduate scenarios reward checking with a supervisor; management scenarios often mark the same behaviour down as failing to take a decision you own.

Frequently asked questions

TestSolve is independent and not affiliated with any test provider or employer named on this page. All product names and trademarks belong to their respective owners.