Situational Judgement Test 2026: Formats

A situational judgement test (SJT) describes realistic work situations and asks you to judge possible responses. Depending on the test, you pick the most and least effective option, rank all the options, or rate each one. Banks, consultancies and public employers use SJTs to check how well your judgement fits their values. Many SJTs are untimed; the two described here took between a few minutes and about ten minutes. This guide shows two formats seen in real 2026 graduate assessments (most/least of three, and rank 1–5 of five), explains how SJTs are scored, and gives original example questions with reasoning.

Key facts

What it looks like: two real formats

Format 1: most and least effective of three (Korn Ferry)

In a 2026 application round, an employer-branded culture-match SJT on Korn Ferry's platform ran as 1 practice scenario plus 18 scored scenarios, with an "n / 18" counter.

• Screen: a stock photo and a 2–6 sentence scenario on the left; three response options on the right.
• Two steps: first select the option that is "am effektivsten" (most effective), then the one that is "am wenigsten effektiv" (least effective). The options then show "(Most Effective)" and "(Least Effective)" tags.
• Controls: a reset link appears after both picks, and the next button stays grey until both picks are made. There is no way back to earlier scenarios.
• Storyline: all 18 scenarios follow one new graduate through a rotation programme. Each scenario starts by saying how the previous one turned out.
• Timer: none.

With three options, choosing most and least ranks all three. So this format is really a three-way ranking.

Format 2: rank five options from 1 to 5 (Cappfinity)

In another 2026 round, a values-based SJT on Cappfinity's platform ran in English.

• Blocks by value. Each block opens with a short statement of one firm value and is followed by 2–3 scenarios. The browser tab names the value ("Question 1 - <value>"). Ten scenarios were seen, spread over four values. The fifth value introduced a separate slider questionnaire instead.
• Prompt: a scenario, then a question such as "What would you do?", "How do you respond?" or "How do you feel?". The instruction always asks you to rank by what you would most likely do (1) to least likely do (5).
• Answering: five option cards, A–E, sit in rank slots 1–5. You drag them by a handle to reorder them.
• Default order: the cards start in order A=1 … E=5. If you forget to drag, you submit A-B-C-D-E.
• Navigation: "Save and Show Previous" / "Save and Show Next" within a block, then "Save and Close". You can revise earlier rankings in the same block.
• Timer: none. Each scenario takes about half a minute to a minute to read and rank properly.

In both tests, every option was plausible. Typically one or two options fitted the value, one hedged, and one or two avoided the problem or went solo. Close pairs differed by small wording cues such as "immediately", "strongly", or "…but recognise you need the whole team".

Other formats you may meet

• Rate each option on a 4- to 6-point scale from "very ineffective" to "very effective".
• Pick one best option from four or five.
• "Would do" versus "should do". Some tests ask what you would do (behavioural tendency); others ask what you should do or what is most effective (knowledge). Read the instruction: the two can lead to different answers.
• Video or animated SJTs, and SJTs combined with a job simulation or an in-tray task.

How SJTs are scored

SJT publishers rarely publish their keys, and neither test described here showed feedback. The research literature describes the common methods:

• Expert keys. Experienced staff or subject-matter experts decide what the best response is, and the answer key follows their consensus. Employer-branded tests such as culture-match SJTs are keyed against the firm's own values and high performers.
• Empirical keys. Responses are keyed by how well they predicted good performance in a pilot group.
• Most/least scoring. You typically earn credit for matching the keyed best option and the keyed worst option. Some schemes deduct credit when you rate the keyed worst option as best.
• Ranking and rating scoring. Your ranks or ratings are compared with the key, and the closer you are, the more credit you get. A small mistake, such as swapping ranks 1 and 2, costs less than a large one, such as putting the key's rank 5 at the top.

Your result is usually compared with a norm group. It is then combined with your ability tests and screening, or used as a hurdle. SJTs predict job performance moderately well: one large meta-analysis estimated an average corrected validity of about .34. Knowledge-type instructions ("most effective") relate more to cognitive ability, and behavioural-tendency instructions ("would do") relate more to personality.

Question types with original examples

The scenarios below are original. The "answers" are reasoned keys of the kind employers use, not an official key.

Example 1: most and least effective of three

You're finishing a client deck for 9:00 tomorrow. At 18:00 you notice that one chart, built from a colleague's model, doesn't match the source table in the appendix. The colleague is on a flight until morning.

• (a) Correct the chart yourself to match the source table and send the deck without mentioning it.
• (b) Tell your manager tonight, show the two figures and the source, propose the correction, and leave a note for your colleague.
• (c) Send the deck as it is. The model is your colleague's responsibility.

Most effective: (b). Least effective: (c).

(b) fixes the problem, keeps the owner and the manager informed, and leaves a record. (a) probably gets the number right but changes someone else's work without telling them; if the source table is the one that's wrong, you have created a second error. (c) knowingly sends a wrong number to a client.

Example 2: most and least effective of three

On a call with a client, a senior manager gives a regulatory deadline that you are fairly sure is a month too early. You read the regulator's notice last week.

• (a) Correct the manager straight away in front of the client.
• (b) Send the manager a short private message during the call with a link to the notice, so they can correct it themselves.
• (c) Say nothing. You can mention it at the next team meeting.

Most effective: (b). Least effective: (c).

(b) gets the correct information into the conversation while the call is still running, without embarrassing anyone. (a) is clumsy, but at least it fixes the error. (c) lets a wrong deadline stand with a client, which is the worst outcome. This item illustrates a general rule: the least effective option is usually the one that leaves the problem in place.

Example 3: rank 1–5

You have just joined a project. Your manager asks you to build a progress tracker by Friday in a planning tool you have never used. What would you do?

• A. Look for the internal guide or training for the tool, start building, and ask a colleague when you get stuck.
• B. Ask your manager to give the task to someone who already knows the tool.
• C. Build the tracker in a spreadsheet you know well, without mentioning that you didn't use the tool.
• D. Ask a colleague who uses the tool for 15 minutes to show you the basics, then work on your own.
• E. Wait until your manager has time to explain the tool to you in detail.

Suggested ranking: 1 A, 2 D, 3 E, 4 B, 5 C.

A and D both show ownership and learning. A ranks higher because it uses the self-service route first and someone else's time second. E is passive and will probably cost you the deadline. B avoids the task. C is the lowest because it quietly ignores the request, which is a transparency problem, not just a skills gap.

Example 4: rank 1–5

During a very busy week, a teammate who is usually careful has sent two pieces of work with obvious errors. They seem stressed. What would you do?

• A. Ask them privately how they are doing and offer to take one task off their list.
• B. Tell your manager that the teammate's work quality has dropped.
• C. Quietly redo their work yourself without saying anything.
• D. Leave it. Everyone is busy this week.
• E. Suggest in the next team meeting that the team reviews how work is shared out during peaks.

Suggested ranking: 1 A, 2 E, 3 B, 4 C, 5 D.

A deals with the person and the work at once. E addresses the cause at team level, but slowly and impersonally. B is legitimate if things don't improve, but as a first step it skips the colleague. C protects the output but hides the problem and burns your own time. D ignores both the person and the quality.

Example 5: values under pressure (most and least of three)

A prospective client asks you to send them "an example of a similar analysis you did for another company" before they sign.

• (a) Send a past client's report with the company name blacked out.
• (b) Explain that client work is confidential. Offer a public case study or a sample built on dummy data, and check with your manager what can be shared.
• (c) Send the full report. It's the same industry, so it shows exactly what they would get.

Most effective: (b). Least effective: (c).

Confidentiality is not a matter of style. Blacking out a name (a) often still identifies the client from the details. (c) is a clear breach. (b) protects the existing client and still helps the prospect.

Traps and how to avoid them

• Choosing the "nicest" answer. The keyed best option is usually the one that solves the problem and informs the right people, not the most polite one.
• Escalating too early. Forwarding everything to your manager without looking at it first rarely wins. A few minutes of your own investigation, then a clear report, usually does.
• Escalating too late. For compliance, legal, IT or anything with client risk, the right specialist beats your own improvisation.
• Passing internal problems to the client. Options that ask the client to redo something because of your mistake, or that hide a risk from the client, score low.
• Forgetting to drag (ranking format). If the cards start in A–E order, check that the final order is really yours before you submit.
• Mixing "would" and "should". Read the instruction on every screen. In the Cappfinity test, "How do you feel?" still came with a "most likely do" ranking instruction.
• Inconsistency. Linked storylines and value blocks repeat themes. If you value speaking up in one scenario, don't choose silence in a near-identical one.
• Overthinking near-twins. When two options differ by one word ("immediately" versus "at the next opportunity"), ask which one fits the value statement shown at the top of the block, then move on.

How to prepare

SJTs reward judgement and fit, so preparation means understanding the format, knowing the employer's values, and answering honestly and consistently.

Step 1: research the values (1 hour per employer). Find the employer's published values or behaviours and translate each into graduate-level behaviour:

Step 2: learn the format (30 minutes). Know whether you will choose most and least, rank, or rate. Do a few practice items in that exact format.

Step 3: practise with reasons (2–3 hours). Work through 30–40 mixed items. For each, write one line saying why the least effective option is least effective. That is where most marks are lost.

Step 4: build a personal checklist. Understand the problem → do a quick check yourself → inform the right person → protect the client and the rules → follow up and improve.

Step 5: on the day. Take the test in one sitting, somewhere quiet. Read every option before you choose. Don't rush just because there is no timer.

Sources

• McDaniel, M. A., Morgeson, F. P., Finnegan, E. B., Campion, M. A., & Braverman, E. P. (2001). Use of situational judgment tests to predict job performance: A clarification of the literature. Journal of Applied Psychology, 86, 730–740 (mean corrected validity ≈ .34).
• McDaniel, M. A., Hartman, N. S., Whetzel, D. L., & Grubb, W. L. (2007). Situational judgment tests, response instructions, and validity: A meta-analysis. *Personnel Psychology, 60*, 63–91 (knowledge versus behavioural-tendency instructions). Summary
• Whetzel, D. L., & McDaniel, M. A. (2009). Situational judgment tests: An overview of current research. Human Resource Management Review, 19, 188–202 (keying and scoring methods).

Escrito por

Adrian, especialista

MSc em Psicologia, 7 anos a fazer testes.

Última atualização: Oct 1, 2026