Watson Glaser Test 2026: Sections and Timing

The Watson Glaser test (full name: Watson-Glaser Critical Thinking Appraisal) is a critical thinking test published by Pearson TalentLens. The current version, Watson-Glaser III, has 40 items in 30 minutes across five sections: inference, recognition of assumptions, deduction, interpretation and evaluation of arguments. UK commercial law firms use it widely for training contracts and vacation schemes. This guide covers each section's answer labels, the versions you may meet and how law firms use it, with original practice items for every section. For a second, different set of examples, see the critical thinking test guide.

Key facts

What the Watson Glaser test looks like

Each item gives you a short passage, then one or more statements to judge using only the passage. You treat the passage as true, even if it clashes with what you know. The five sections each have their own instructions and answer labels:

Deduction and interpretation share labels but not the standard. Deduction is strict logic: the conclusion must be forced by the premises. Interpretation allows a sensible reading of data (percentages, totals, comparisons) but never a jump to causes or to groups the data does not cover.

Versions you may meet

• Watson-Glaser III. The current online test: 40 items in 30 minutes, drawn from an item bank. Clifford Chance's London careers page describes exactly this: "The test lasts 30 minutes and includes 40 questions."
• Forms A and B (80 items). An academic paper gives about 55 minutes; UK law careers sites give 60. You are more likely to meet the 40-item version, but if your invitation says 80, budget the same 45 seconds per item.

How results are reported

The report gives an overall percentile against a norm group plus three areas, the RED model: Recognise assumptions, Evaluate arguments and Draw conclusions (inference, deduction and interpretation together). Employers choose the norm group and the cut-off.

Watson Glaser in UK law firm recruitment

• When: after the application form or alongside another online assessment, before the assessment centre. Some firms place it second, after an immersive or situational online assessment, with a deadline of a few days.
• Who: law careers sites report it at many large commercial firms in London, from the largest international practices to national and regional firms, for vacation schemes and training contracts. Processes change between cycles; check each firm's careers page.
• Pass marks: firms do not publish them, and the threshold can change year to year. Prep and law careers sites quote targets of 75 % to 80 % correct, but these are estimates, not firm policy.
• The Bar: the Bar Course Aptitude Test (BCAT), built from Watson-Glaser items, was a requirement for the Bar course until the Bar Standards Board abolished it from 31 July 2022.

Firms often pair it with a situational judgement test or an immersive simulation.

Original Watson Glaser practice questions

All items below are written for this guide in the real formats. Judge each one using only the passage.

Inference

Passage 1. In April, Riverside Council replaced its 25 old bicycle-hire stations with 40 new ones. From April to September, rentals averaged 2,600 a day, against 1,900 a day in the same months of the previous year. In a council survey of 500 renters, 70 % said they used the bikes to commute to work.

Rate each statement: True, Probably true, Insufficient data, Probably false, False.

1. Daily rentals from April to September were higher than in the same months a year earlier.
2. Some people who rented a bike after April had never rented one from the old network.
3. The new stations were paid for by a national grant.
4. On average, each station was used more often per day than before the change.
5. Demand for bikes in the city centre was lower after April than before.

Answers: 1 True, 2 Probably true, 3 Insufficient data, 4 False, 5 Probably false.

• 1 is stated: 2,600 against 1,900.
• 2: 700 extra rentals a day make new users likely, but the same people could be renting more often.
• 3: funding is never mentioned.
• 4: 2,600 ÷ 40 = 65 rentals per station now against 1,900 ÷ 25 = 76 before: the passage's numbers contradict it.
• 5: rentals rose by over a third. Lower demand is not impossible, but the facts point the other way.

Passage 2. A software firm with 800 employees offered an optional four-day week at 90 % of salary. In the first year, 120 employees took it up. Staff turnover across the firm fell from 14 % to 9 % that year; among four-day staff it was 3 %.

1. Fewer than one in five employees took up the four-day week.
2. Turnover among staff who stayed on five days was below 3 %.
3. Some employees took the four-day week to care for family members.

Answers: 1 True, 2 False, 3 Insufficient data.

• 1: 120 ÷ 800 = 15 %, below 20 %.
• 2: the firm-wide 9 % is an average of the four-day group (3 %) and the five-day group. An average lies between its parts, so the five-day rate must be above 9 %. Roughly: 9 % of 800 = 72 leavers, about 4 of them four-day staff, so about 68 of 680, or 10 %.
• 3: no reasons are given.

Recognition of assumptions

Statement 1. "We should translate our online shop into Spanish. That will increase our sales in Mexico."

1. Some potential customers in Mexico would be more likely to buy from a Spanish-language shop.
2. Customers in Mexico are able to order and receive our products.
3. Spanish is the only language spoken in Mexico.
4. Translation is cheaper than any other way of increasing sales in Mexico.

Answers: 1 made, 2 made, 3 not made, 4 not made.

• Higher sales need a language barrier that matters to some buyers (1) and the ability to sell to them at all (2).
• Spanish only has to be widely understood, not the only language (3). Cost comparisons (4) are a different argument.

Statement 2. "The new printer is twice as fast as the old one, so there will be no more queues at the printer."

1. Printing speed is at least part of the reason queues form.
2. The amount of printing will not rise enough to cancel out the faster speed.
3. The old printer will be removed.

Answers: 1 made, 2 made, 3 not made.

• If queues had nothing to do with speed, a faster printer would not end them (1).
• If printing demand doubled, queues would return; "no more queues" takes for granted that it will not (2).
• Whether the old printer stays has no bearing on the claim (3).

Deduction

Premises 1. No contract signed before 2020 includes the new data clause. Some contracts in the Berlin file were signed before 2020.

1. Some contracts in the Berlin file do not include the new data clause.
2. No contract in the Berlin file includes the new data clause.
3. Every contract without the new data clause was signed before 2020.

Answers: 1 follows, 2 does not follow, 3 does not follow.

• 1: the Berlin file's pre-2020 contracts fall under "no contract signed before 2020", so they lack the clause.
• 2: the file may also hold later contracts, about which the premises say nothing.
• 3 reverses the first premise.

Premises 2. If the shipment arrives by Friday, the launch goes ahead on Monday. The launch did not go ahead on Monday.

1. The shipment did not arrive by Friday.
2. The launch was postponed because of the shipment.
3. If the launch had gone ahead on Monday, the shipment would have arrived by Friday.

Answers: 1 follows, 2 does not follow, 3 does not follow.

• 1: if the shipment had arrived, the launch would have happened. It did not, so the shipment did not arrive.
• 2: the shipment was late, but it need not be the reason; something else could also have stopped the launch.
• 3 reverses the condition: arrival guarantees the launch, but the launch does not require arrival.

Interpretation

Passage 1. A gym chain had 12,000 new members in 2024, and 18 % of them cancelled within three months. In 2025 it introduced a free induction session; it had 15,000 new members, and 12 % cancelled within three months.

1. Fewer new members cancelled within three months in 2025 than in 2024.
2. The free induction session reduced early cancellations.
3. A larger share of new members stayed beyond three months in 2025 than in 2024.

Answers: 1 follows, 2 does not follow, 3 follows.

• 1: 18 % of 12,000 = 2,160; 12 % of 15,000 = 1,800. Fewer, even with more members.
• 2 turns timing into cause. Prices, staff or the type of member may also have changed.
• 3: 82 % stayed in 2024 and 88 % in 2025.

Passage 2. Of 300 applicants who sat a written test, the 60 highest scorers were interviewed. Fifteen of those interviewed received an offer.

1. At least 5 % of the applicants received an offer.
2. One in four interviewees received an offer.
3. The written test predicts how well people perform in the job.

Answers: 1 follows, 2 follows, 3 does not follow.

• 1: 15 ÷ 300 = 5 %. "At least" is right because the passage does not rule out other offers.
• 2: 15 ÷ 60 = 25 %.
• 3: nothing is said about later job performance.

Evaluation of arguments

Question 1. Should universities record all lectures and publish them online for enrolled students?

1. Yes. Students who are ill or have to work can catch up on material they would otherwise miss.
2. No. Attendance in the lecture hall would fall, and with it the chance to ask questions and discuss.
3. Yes. Almost everything is online nowadays.
4. No. Lectures were never recorded in the 1990s, and graduates did fine.

Answers: 1 strong, 2 strong, 3 weak, 4 weak.

Treat each argument as true and ask: is it important, and is it about this proposal? 1 and 2 name real consequences for learning. 3 is a sweeping generalisation that says nothing about lectures in particular. 4 appeals to the past, which does not show that recordings would not help now.

Question 2. Should a company force all staff to change their passwords every 30 days?

1. Yes. Our competitors do it.
2. No. Research shows that forced frequent changes lead many people to choose weaker, predictable passwords, which can lower security.
3. Yes. If a password is stolen without anyone noticing, a regular change limits how long it can be misused.

Answers: 1 weak, 2 strong, 3 strong.

1 copies others without a security reason. 2 and 3 point in opposite directions, yet both are strong: each gives an important, directly relevant effect. A strong argument does not have to win.

Traps and how to avoid them

• Skipping hidden maths. Passages often hide a quick calculation (per-station use, a weighted average, an "at least"). If a statement compares rates, compute them.
• Category boundaries. "True" needs certainty; "Probably false" needs facts pointing against the statement, otherwise it is "Insufficient data".
• Mixing deduction and interpretation. Deduction is strict logic; interpretation accepts a sensible reading of numbers but never a cause.
• Reversed conditionals. "If A then B" plus "not B" gives "not A"; "B" never gives "A".
• Treating "strong" as "agree". Two opposing arguments can both be strong.
• Time. 45 seconds per item. Don't spend two minutes on one deduction; move on.

How to prepare for the Watson Glaser test

A one-week plan, about 6 hours in total:

On the day: one sitting, a quiet room, a laptop rather than a phone, scrap paper for quick sums.

Sources

• Pearson TalentLens, Watson-Glaser Critical Thinking Appraisal (items, timing, subtests, reports, languages)
• Pearson TalentLens France, Watson-Glaser III (40 items, 30 minutes, item bank)
• Pearson TalentLens, practice tests (free Watson-Glaser practice test, not scored)
• Practice Aptitude Tests, Watson Glaser test guide (RED mapping)
• The Lawyer Portal, Watson Glaser test (40 or 80 questions, pass marks)
• Prospects, "How to pass the Watson Glaser test" (answer options)
• Journal paper on the Watson-Glaser (Forms A and B: 80 items, about 55 minutes)
• Bar Standards Board, "BSB to abolish the Bar Course Aptitude Test from 31 July"

Escrito por

Adrian, especialista

MSc em Psicologia, 7 anos a fazer testes.

Última atualização: Oct 1, 2026