The Watson Glaser Test in 2026: How to Pass the Critical Thinking Test That Guards Your Dream Job

You got the email. It mentioned Watson Glaser, gave you a link, maybe a deadline, and now your stomach is doing something unpleasant. Take a breath, because this is one of the most beatable hurdles in the whole hiring process.

The Watson Glaser Critical Thinking Appraisal is a standardised psychometric test built by Pearson TalentLens. It measures how well you reason from evidence, not how smart you are or how much you already know. That distinction is the whole game.

Here is the good news that most panicking candidates miss: this test follows fixed, learnable rules. It rewards discipline over instinct, and discipline is something you can build in a week. The people who fail usually fail because they trusted their gut instead of the rules.

This guide walks you through all five sections, the scoring math, the tech setup, and the exact traps that sink smart candidates. Treat it like the prep manual you wish the employer had sent. And if you want a broader warmup on selling yourself once you clear this gate, our take on the “tell me about yourself” question is worth a look for later.

How the Watson Glaser Assessment Actually Works

The modern version is the Watson Glaser III, which is item-banked. That means two candidates applying to the same firm might not see identical questions, so memorising a leaked answer key is pointless. An older 80-question, 55-minute version (Watson Glaser II, Forms D and E) still shows up at some employers, and that one uses fixed question sets.

The test is usually delivered online and unproctored, sent as a link after your initial application. It acts as a screening gate before anyone reads your CV closely. Some regulated employers and civil service roles may run it under proctored conditions, and the setup is decided by each employer, not by Pearson.

Five sections sit under what Pearson calls the RED model: Recognise assumptions, Evaluate arguments, Draw conclusions. The sections are Inference, Recognition of Assumptions, Deduction, Interpretation, and Evaluation of Arguments. Each one has its own response rules, and mixing them up under time pressure is a classic way to lose points.

Here is the part candidates underestimate: the test is designed to punish intuition. Plausible-sounding answers that feel right are often the trap. You are being scored on whether you can hold the line and reason strictly from the text in front of you.

Once you understand that this is a rules test rather than a knowledge test, your prep changes completely. You stop cramming facts and start drilling the response frameworks until they become automatic, the same way you would rehearse structured answers before a panel interview.

• Format: 40 multiple-choice questions across five fixed sections: Inference (5), Recognition of Assumptions (12), Deduction (5), Interpretation (6), and Evaluation of Arguments (12). An older 80-question version appears at some employers.
• Timing: Roughly 30 minutes for the 40-question WG-III, about 45 seconds per question. The older version runs 55 minutes for 80 questions. Some employers give a multi-day window to open the test, but once you start, the clock runs continuously.
• Scoring: Norm-referenced and percentile-based against a comparator group the employer chooses, often law graduates. On WG-III, question difficulty is weighted, so getting a hard item right or an easy item wrong carries extra significance. There is no universal percentage pass mark.
• Retakes: Policies vary by employer. Most law firms permit only one attempt per application cycle, usually one academic year. Some organisations enforce a waiting period of six months to a year before you can sit again. Always check your specific invite.

Top 12 Watson Glaser Assessment Questions With Sample Answers

1. Inference section: rating a conclusion on the five-point scale

What they’re really testing: This section tests disciplined probabilistic reasoning. You read a factual passage, then rate a proposed inference as True, Probably True, Insufficient Data, Probably False, or False, based only on the passage. It is checking whether you can resist filling gaps with your own knowledge.

Interview Guys Tip: Read the passage as the complete universe of facts. Reserve ‘True’ and ‘False’ only for statements the passage explicitly confirms or contradicts. Use ‘Probably True’ and ‘Probably False’ when the passage leans a direction without proving it. Use ‘Insufficient Data’ when the passage simply does not address the claim. Frame it in your head as: does the text give me evidence pointing this way, and how strongly?

Avoid this: Importing outside knowledge. If the passage never mentions something, no amount of real-world common sense makes the inference ‘Probably True’. That gap means Insufficient Data.

2. Inference example: telemedicine and patient preference

What they’re really testing: This item separates a strongly supported directional inference from one that needs data the passage never gives you. Passage: a hospital introduced telemedicine in January, cancellations fell 22% by March, and staff overtime fell 11%. Proposed inference: ‘Patients prefer telemedicine to in-person appointments.’

Interview Guys Tip: The correct rating is Insufficient Data. The passage shows fewer cancellations and less overtime, but it says nothing about patient preference. Lower cancellations could stem from many causes, including scheduling changes or seasonal effects. With zero preference data in the text, you cannot rate this even ‘Probably True’.

Avoid this: Assuming that a good outcome proves people liked it. Feeling that telemedicine ‘probably’ pleased patients is exactly the intuition the test is built to punish.

3. Inference example: new bus route and car use

What they’re really testing: Another correlation-versus-causation trap. Passage: a city added a bus route in June and ridership on that corridor rose 18% by August. Proposed inference: ‘The new route caused commuters to switch from private cars.’

Interview Guys Tip: Rate this Insufficient Data. The passage shows ridership rose, but it gives no data on private car use at all. The riders could be former walkers, cyclists, or people making new trips. Without any information linking the increase to car users specifically, the causal claim floats unsupported.

Avoid this: Treating a plausible mechanism as evidence. It sounds reasonable that a new bus tempts drivers, but ‘sounds reasonable’ is not in the passage.

4. Recognition of Assumptions: is the hidden premise Made or Not Made

What they’re really testing: The largest section at 12 questions. You are given a statement plus a proposed assumption, and you decide whether that assumption is ‘Made’ or ‘Not Made’. It tests whether you can spot the unstated premise that the statement absolutely requires to make sense.

Interview Guys Tip: Run a two-step check. First, identify what the statement is actually trying to argue. Second, ask: would the statement fall apart if this assumption were false? If it collapses, the assumption is Made. If the statement can still stand, it is Not Made. Only logically necessary premises count.

Avoid this: Accepting assumptions that are merely likely or reasonable. Plausible is not the standard. The premise has to be genuinely required for the statement to hold together.

5. Assumptions example: advertising on social media

What they’re really testing: Statement: ‘We should advertise our product on social media to reach younger consumers.’ Proposed assumption: ‘Younger consumers use social media.’ This tests whether you can identify a premise that the statement literally depends on.

Interview Guys Tip: The assumption is Made. If younger consumers did not use social media, the entire recommendation would be incoherent, because you cannot reach them on a channel they do not use. Remove the premise and the statement breaks. That is the signature of an assumption that is Made.

Avoid this: Overthinking it into ‘Not Made’ because the link feels too obvious. Obvious does not mean absent. If the statement needs it to function, it is Made.

6. Deduction: does the conclusion follow necessarily

What they’re really testing: Five questions of pure formal logic. You get two premises, treat them as absolutely true regardless of real-world plausibility, and decide whether the conclusion ‘Follows’ or ‘Does Not Follow’. This is syllogistic reasoning, nothing more.

Interview Guys Tip: Treat the premises as gospel, even if they seem absurd. Sketch a quick mental Venn diagram to test whether the conclusion must be true given only those two premises. If there is any way the premises can be true while the conclusion is false, it does not follow.

Avoid this: Letting real-world knowledge override the premises. If the premises say all swans are purple, then within this question all swans are purple, full stop.

7. Deduction example: partners and the training session

What they’re really testing: Premises: ‘All partners at the firm attended the training. Sam did not attend the training.’ Conclusion: ‘Sam is not a partner at the firm.’ This tests whether you can apply a valid contrapositive.

Interview Guys Tip: The conclusion Follows. If every partner attended, then anyone who did not attend cannot be a partner. Sam did not attend, so Sam cannot be a partner. The logic is airtight given those exact premises.

Avoid this: Second-guessing with real-world doubts like ‘maybe Sam was sick that day’. The premises do not allow exceptions, so you cannot invent one.

8. Interpretation: does the conclusion follow beyond a reasonable doubt

What they’re really testing: Six questions, usually built on data or a short passage. You decide whether a conclusion ‘Follows’ or ‘Does Not Follow’, but the threshold is higher than Inference. Here the standard is ‘beyond a reasonable doubt’ based on the evidence given.

Interview Guys Tip: Ask whether the evidence strongly supports the conclusion, even if alternative explanations are theoretically possible. If the data points clearly in one direction and the conclusion matches that direction without overreaching, it Follows. Watch for conclusions that add claims the data does not cover.

Avoid this: Demanding absolute certainty. Interpretation does not require ironclad proof, so rejecting a strongly supported conclusion because it ‘might’ have another cause is a mistake.

9. Interpretation example: mentoring and promotion

What they’re really testing: A table shows 70% of mentored employees were promoted within two years versus 30% of non-mentored employees. Proposed conclusion: ‘Mentoring increases the likelihood of promotion.’ This tests the beyond-reasonable-doubt threshold.

Interview Guys Tip: The conclusion Follows. The data strongly supports the direction of the claim, and the conclusion is carefully worded as ‘increases the likelihood’ rather than ‘guarantees’ or ‘directly causes’. Because it matches the evidence without overstating it, it clears the bar.

Avoid this: Rejecting it because correlation is not causation. In Interpretation, a direction that is strongly supported by the data Follows, even when a stricter causal proof is missing.

10. Evaluation of Arguments: strong versus weak

What they’re really testing: Twelve questions, tied for largest. You are given a question and an argument, and you judge whether the argument is ‘Strong’ or ‘Weak’. A strong argument is directly relevant to the core question and substantial. A weak one is irrelevant, emotional, trivial, or peripheral.

Interview Guys Tip: Set your own opinion aside completely. Ask two things: is this argument directly relevant to the exact question, and does it carry real weight on the issue as a whole? Only a yes to both makes it Strong. Relevance plus substance, nothing less.

Avoid this: Marking an argument Strong because you personally agree with it. Agreement is not relevance. This is one of the most common scoring errors in the whole test.

11. Evaluation example: mandatory mental health training

What they’re really testing: Question: ‘Should all employees be required to complete annual mental health training?’ Argument: ‘No, because some employees already feel confident in this area.’ This tests whether you can spot an argument that addresses only a minority case.

Interview Guys Tip: The argument is Weak. It focuses on a subset of employees who feel confident and ignores the policy question as a whole, which is about all employees. Because it is peripheral rather than engaging the core issue, it fails the substance test even though it might be true.

Avoid this: Treating a true statement as automatically strong. An argument can be factually accurate and still Weak if it only nibbles at the edges of the actual question.

12. Pacing strategy across all five sections

What they’re really testing: This is not a question type, it is the meta-skill that decides scores. Candidates who over-invest in the psychologically demanding early Inference items often sprint through the 12-question Arguments section at the end and leak easy points.

Interview Guys Tip: Budget roughly 45 seconds per question and keep moving. If an Inference or Assumptions item stalls you, make your best rules-based judgement and move on. Protect enough time for the two big 12-question sections, Assumptions and Arguments, since that is where most of your marks live. Practising under a live timer is the only way this becomes instinct.

Avoid this: Perfectionism on the hard-feeling questions. Spending three minutes agonising over one Inference item can cost you five gettable Arguments answers later.

6 Insider Tips for Acing It

Master one section at a time before you mix them

Each section uses a different response rule: Inference has five options, the other four have two. Under time pressure candidates blend the rules and rate an Assumptions item on the Inference scale. Drill each framework in isolation until it is automatic, then practise them combined.

The passage is your whole universe

This is the habit that separates passers from failers. Treat the given text as the only facts that exist, even when it contradicts what you know is true in real life. Every time you catch yourself thinking ‘but in reality’, stop and reread the passage.

Stop hedging with ‘Insufficient Data’

Defaulting to Insufficient Data whenever you are unsure is not a safe strategy, it is a scored error. The five-point Inference scale demands genuine probabilistic judgement, and habitual hedging signals you did not engage with the passage. Use it only when the text truly says nothing about the claim.

Practise setting aside your beliefs on purpose

In Evaluation of Arguments, your personal opinion is poison. Deliberately rehearse marking arguments you disagree with as Strong when they are relevant and substantial, and arguments you love as Weak when they are peripheral. Neutrality is a muscle you train, much like staying composed in structured leadership interviews.

Do at least one full timed mock

Free timed mocks are available from Pearson TalentLens sample materials, AssessmentDay, JobTestPrep, and AllAboutLaw. Complete at least one full 30-minute mock under real conditions before test day, and use Watson Glaser-specific banks rather than generic verbal reasoning tests. You can build your foundation with free WGCTA practice from iPrep.

Review by the rule you broke, not the answer you missed

After each practice session, go back through every wrong answer and label the exact error: wrong scale, outside knowledge, or personal bias. This turns mistakes into fixed habits fast. Simply rereading the question teaches you nothing about why you slipped.

Your Technical Setup Checklist

• Use Chrome on a stable connection. Candidates report browser compatibility issues, especially with older versions of Firefox.
• Confirm your window and deadline the moment the invite arrives. Some employers give several days to open it (Linklaters has been reported to allow five), but the 30-minute clock runs continuously once you start.
• Pick a quiet block with zero interruptions, because you cannot pause the timer mid-test.
• Close every other tab and app to prevent lag or crashes during the sitting.
• Have scrap paper and a pen ready for quick Venn sketches in the Deduction section.
• Check whether your specific test is proctored. If it is, sort your webcam, ID, and clear-desk setup in advance.
• Charge your laptop or stay plugged in, and if possible use a wired connection over patchy wifi.

What Happens After You Submit

Once you submit, the employer scores your sitting against their chosen norm group and applies their own cutoff. Magic Circle and Silver Circle firms typically use a threshold around the 70th to 75th percentile against a law-graduate norm, which roughly corresponds to 29 or more correct out of 40, though firms do not publish exact numbers. For context, published norm data reported by AssessFirst puts the 80th percentile near 33 to 34 correct, the 90th near 36 to 38, and the 95th to 99th at 39 to 40.

Here is a reality check worth internalising: your raw score and your percentile are not the same thing, and the percentile shifts by firm. Candidates on The Corporate Law Academy forum describe scoring in the same raw range but landing at the 65th percentile at one firm and the 88th at another on live sittings taken close together. That is norm groups at work, and it means an 80% raw score is never an automatic pass. You can read those firsthand accounts on the TCLA discussion thread.

If you clear the gate, the next stages usually follow quickly, whether that is a video interview, an assessment centre, or a first-round call. It pays to line up your prep for what comes next, starting with the fundamentals in our roundup of common phone interview questions.

Frequently Asked Questions

Is there a pass mark for the Watson Glaser test?

No universal one. Scoring is percentile-based against a norm group each employer chooses, so the same raw score can pass at one firm and fail at another. Law firms commonly cluster their cutoff around the 70th to 75th percentile against a law-graduate norm, but none of them publish exact figures. Aim high rather than chasing a specific number.

Can I retake it if I do badly?

It depends entirely on the employer. Most law firms allow only one attempt per application cycle, usually one academic year, and some organisations enforce a waiting period of six months to a year before you can sit again. That is why first-attempt preparation matters so much. Always check the exact policy in your invite email.

Does my score carry over to other employers?

No. Scores are not portable, and there is no central score bank you can share across applications. Each employer administers, owns, and interprets their own sitting against their own norm group. Applying to three firms means sitting three separate assessments.

Can you actually prepare for Watson Glaser, or is it just innate ability?

You can absolutely prepare. It measures a specific set of trained critical-thinking skills, not raw intelligence, and the five sections follow predictable rules that improve substantially with targeted practice. The candidates who treat it as unbeatable are usually the ones who did not drill the frameworks. Preparation is the whole edge, much like rehearsing for any high-stakes conversation.

Will the questions be about law?

Not always. Law firms tend to use legal or business-flavoured passages, but civil service and regulatory employers like the Fast Stream and the FCA often draw on policy, economics, and social science content. The reasoning rules are identical regardless of subject matter, so do not let unfamiliar content rattle you.

How is Watson Glaser different from other aptitude tests I might face?

It focuses on critical thinking under strict rules rather than speed maths or vocabulary, so generic SHL or CEB prep will not train the five specific frameworks. It rewards disciplined judgement over quick pattern-matching. Think of it as one distinct gate in a wider process that can also include numerical tests and interviews, so treat your interview prep as a separate workstream.

The Watson Glaser test feels intimidating because it punishes the instincts you trust everywhere else. But that is also why it is so beatable: once you learn to reason strictly from the passage, apply each section’s rule cleanly, and manage your 45 seconds per question, the trap doors stop catching you. Discipline wins this one, not genius.

Put in one focused week of section-by-section drills, run at least one full timed mock, and review your errors by the rule you broke. Do that and you will walk into the sitting calm and ready. Clear this gate, then turn your attention to the human stages ahead, because that is where the offer is actually won.

Sources and Further Reading

• https://www.pearsontalentlens.com/products/watson-glaser-critical-thinking-appraisal
• https://www.allaboutlaw.co.uk/law-careers/selection-assessment/which-firms-use-the-watson-glaser
• https://www.careertestprep.com/knowledge/watson-glaser
• https://www.iprep.online/courses/watson-glaser-test-wgcta-free/
• https://www.thecorporatelawacademy.com/forum/threads/watson-glaser-tips-and-tricks.2488/
• https://yourlegalladder.com/resources/guides/watson-glaser-critical-thinking-test-prep