Watson-Glaser Critical Thinking Appraisal: The Law Firm Gatekeeper
The Watson-Glaser is a different species from the CCAT or SHL. It does not test reasoning speed. It tests whether you can suspend your own opinions long enough to evaluate an argument on its own logic. That is a surprisingly hard skill, which is why Magic Circle law firms use it as a near-universal gate.
By PrepClubs Editorial Team, updated April 18, 2026
The Watson-Glaser Critical Thinking Appraisal (WGCTA) is a 40-question, 30-minute critical-reasoning assessment first published in 1925 and now sold by TalentLens, a Pearson division. It runs five sections (inference, recognition of assumptions, deduction, interpretation, evaluation of arguments) that each apply distinct decision rules to short passages and arguments. It is the dominant screen for UK trainee solicitor pipelines at Clifford Chance, Linklaters, Allen and Overy, Freshfields, and Hogan Lovells. Magic Circle cutoffs typically sit between the 60th and 75th percentile against a UK graduate norm, with elite streams pushing to the 80th.
Source: TalentLens (talentlens.com) WGCTA documentation and published Magic Circle trainee selection criteria.
Section-rule mastery, not generic reasoning practice
Why the Watson-Glaser exists and who uses it
The Watson-Glaser Critical Thinking Appraisal (WGCTA) was created in 1925 and is published today by TalentLens (a Pearson division). It is the most widely used critical thinking assessment globally and the dominant screen for UK trainee solicitor positions at Clifford Chance, Linklaters, Allen and Overy, Freshfields, and Hogan Lovells, among others.
The test is 40 questions across five sections in 30 minutes. Unlike cognitive tests, there is no time pressure inside each question, but the aggregate time is tight. Most candidates leave 2 to 4 questions unanswered without realizing.
What Watson-Glaser tests is whether you can separate what an argument actually says from what it sounds like it says. This is a skill trained lawyers develop over years. In hiring, firms use it to predict whether the candidate has the raw disposition for legal reasoning.
The five sections explained
Each section has its own instruction set. The instructions matter: candidates who skim them score 5 to 10 percentile points lower.
Inference
You are given a statement and a set of proposed inferences. For each, decide if it is True, Probably True, Insufficient Data, Probably False, or False based ONLY on the statement. The trap: confusing "Probably True in the real world" with "Probably True based on this passage."
Recognition of Assumptions
Each question gives a statement and a proposed assumption. Decide whether the assumption is Made or Not Made by the speaker. The trap: marking all reasonable assumptions as Made. Watson-Glaser wants assumptions that the stated argument REQUIRES, not ones that are plausible.
Deduction
Given a set of premises, decide whether a conclusion Follows or Does Not Follow. Pure syllogistic logic. The trap: assuming real-world knowledge. If premises say "All dogs are robots," treat dogs as robots.
Interpretation
Given a passage and a set of proposed conclusions, decide whether each conclusion Follows Beyond Reasonable Doubt from the passage. The trap: accepting conclusions the passage strongly suggests but does not establish.
Evaluation of Arguments
For each question and response, decide whether the response is a Strong Argument or Weak Argument. Strong arguments are directly relevant AND materially important. Weak arguments are irrelevant, emotional, or generic. Most candidates over-classify as strong.
Watson-Glaser vs LSAT logical reasoning vs GMAT critical reasoning: same family, different test
Magic Circle, US JD admissions, and consulting screen each pick a slightly different flavor of formal reasoning. Knowing which you are facing changes the prep approach.
| Spec | Watson-Glaser | LSAT logical reasoning | GMAT critical reasoning |
|---|---|---|---|
| Total questions | 40 | 24 to 26 per section | 10 to 13 of 31 verbal items |
| Time limit | 30 min | 35 min per section | ~65 min for verbal section |
| Per-question budget | About 45 sec | About 80 sec | About 100 sec |
| Question style | 5 distinct decision rules across 5 sections | Stimulus + question stem + 5 answer choices | Stimulus + question stem + 5 answer choices |
| Penalty for guessing | Form-dependent (correction-for-guessing on some) | No penalty | Adaptive: wrong answers lower difficulty |
| Primary audience | UK trainee solicitors, in-house legal hiring | US JD applicants | US MBA applicants, consulting laterals |
| Heaviest employer or program | Magic Circle, City law firms | US law schools | M7 MBA programs, MBB consulting |
| PrepClubs questions | 320+ | Not covered | See GMAT-style cluster |
Why Watson-Glaser is more brutal than its 30-minute time limit suggests
The five-section rule problem
Most aptitude tests apply a single underlying skill across all questions: speed math, verbal pattern recognition, or spatial rotation. Watson-Glaser is the opposite. Each of the five sections has its own scoring criteria, and the sections look almost identical at a glance. Inference and Interpretation share a five-option scale. Recognition of Assumptions and Evaluation of Arguments share a binary scale. Candidates who do not memorize the section rules cold mix them up under pressure and lose 4 to 8 raw points to rule confusion alone.
Inference asks whether a proposed conclusion is True, Probably True, Insufficient Data, Probably False, or False based on the passage. Most candidates lean toward True or False. The correct distribution is heavily weighted to the middle three options, especially Insufficient Data.
Recognition of Assumptions asks whether the argument requires the assumption to hold. The trap is real-world plausibility. An assumption like "people enjoy summer" is plausible, but it is rarely required by the argument. Candidates over-mark assumptions as Made.
The TalentLens norm group changes the percentile by 10 to 15 points
Watson-Glaser raw scores convert to percentiles using one of several TalentLens norm groups. UK Graduate, US Graduate, Legal Professionals, and General Adult are the most common. The same raw score of 30 out of 40 lands at the 70th percentile against UK Graduate, the 75th percentile against US Graduate, and the 60th percentile against Legal Professionals.
Magic Circle firms specify which norm they score against. Most use UK Graduate or Legal Applicants. Asking which norm group a firm uses is a fair question for graduate recruitment teams, and the answer materially changes your prep target.
Some firms publish "minimum 60th percentile" without specifying the norm. Treat this as Legal Professionals norm by default, because it is the most conservative read. Hitting 60th there is harder than hitting 60th against UK Graduate.
Why speed prep loses points on Watson-Glaser
Most cognitive prep advice centers on time management: skip-and-move, race the clock, fill blanks. Watson-Glaser inverts that. Candidates who finish in 22 minutes typically score 4 to 8 percentile points lower than candidates who use the full 30. The reason is that section rule confusion and real-world bleed both increase under haste.
The optimal pacing is roughly 45 seconds per question, with 90 seconds reserved at the end for re-reading any flagged items. Re-reading on Watson-Glaser is unusually high-leverage because the rule-based answer often becomes obvious on a second pass when adrenaline drops.
PrepClubs Watson-Glaser mocks ship with section-by-section accuracy reporting so candidates see exactly which of the five rule sets is leaking points. That is the prep target for days 3 through 8 of a typical 14-day prep cycle.
Build the rest of your prep stack
How Watson-Glaser scoring works
Raw score is the count correct out of 40. TalentLens reports percentiles against one of several norm groups (UK graduate, US graduate, legal professionals, international). Most law firms compare you to a UK graduate or legal applicant pool.
The passing threshold at Magic Circle firms is widely reported as the 60th to 75th percentile, though candidates report cutoffs as high as the 85th at some Freshfields and Linklaters streams. A raw score of 30 out of 40 is roughly the 70th percentile against a UK graduate pool.
Unlike cognitive speed tests, guessing is not clearly free. Some versions of Watson-Glaser use correction-for-guessing, some do not. TalentLens does not publish which variant you took. Rule of thumb: answer everything you have time to reason, but do not random-guess blanks in the final seconds.
Who uses the Watson-Glaser?
Watson-Glaser is the law-firm gate. If you are applying for a training contract in the UK or to a corporate legal department at a Magic Circle or elite boutique firm, expect it.
A Watson-Glaser prep approach (10 to 14 days)
Days 1-2: Learn the five section rules cold
The single biggest prep return comes from memorizing the exact decision criteria for each section. Inference is not Recognition of Assumptions. Deduction is not Interpretation. Section rules are non-intuitive and candidates who apply Deduction rules to Inference questions lose 5 to 8 points.
Days 3-5: Drill one section at a time, untimed
Accuracy first. Work through 20 questions in Inference, scoring each against the official rule. Repeat for each section. Most candidates have one section where their first-try accuracy is under 60 percent. That section is the prep target.
Days 6-8: Timed section drilling
Once accuracy stabilizes above 75 percent per section, add timing: roughly 45 seconds per question. Do not go faster than that, speed does not win points on Watson-Glaser.
Days 9-12: Full-length mocks
Two or three 30-minute mocks under strict conditions. Compare section scores across mocks. Stability is a better signal than high average.
Days 13-14: Rest and review
On the final day, review only your wrong answers from mocks, grouped by section. The test rewards settled clarity, not last-minute cramming.
Watson-Glaser-specific traps
Importing real-world knowledge
The single biggest failure mode. If a passage says "X company increased sales after launching Product Y," you may NOT reason that "X company is doing well." The passage did not say that. Stick to the text.
Over-assuming in Assumptions
Recognition of Assumptions asks what the argument REQUIRES, not what is plausible. "The sky is blue" is plausible and not made by the argument. Most candidates mark too many assumptions as Made.
Flattening the answer scale in Inference
Inference has 5 answer options, not 2. Candidates default to True or False and miss the shades (Probably True, Insufficient Data, Probably False). Most Inference answers are in the middle three options.
Rushing
Unlike CCAT or PI, Watson-Glaser does not reward speed. Candidates who finish in 22 minutes typically score worse than those who finish in 28. If you are done early, re-check.
Related reading
Watson-Glaser FAQs
Magic Circle will not wait. Start your Watson-Glaser prep.
Section-specific drills plus full-length timed simulations with percentile reporting.
Start Watson-Glaser Practice