Updated April 2026 · 12 min read · Pearson TalentLens · Standard test for legal & policy roles
| Provider | Pearson TalentLens |
|---|---|
| Test name | Watson-Glaser Critical Thinking Appraisal (W-GCTA) |
| Format | 40 questions · ~30 minutes · five sections |
| Used by | Linklaters, Allen & Overy, Clifford Chance, Slaughter and May, Hogan Lovells, UK Civil Service Fast Stream |
| Defining feature | The standard cognitive test for trainee solicitor positions and senior policy analysis roles |
The Watson-Glaser Critical Thinking Appraisal is the gold-standard critical thinking assessment for legal and policy careers. Developed by Goodwin Watson and Edward Glaser in 1925, the test has been continuously refined and remains the dominant cognitive screen for trainee solicitor recruitment at Magic Circle and top US law firms. The test does not measure general intelligence — it measures the specific cognitive skill of evaluating arguments, evidence, and conclusions.
You read a short passage and evaluate proposed inferences from it. Each inference is rated on a 5-point scale: True / Probably True / Insufficient Data / Probably False / False.
The trick: you must evaluate based only on the information in the passage, not on your real-world knowledge. If the passage says "75% of city dwellers commute by car" and the inference is "most people in cities own cars," that's "Probably True" — the passage suggests it but doesn't directly state it.
Each item presents a statement followed by proposed assumptions. You decide whether each assumption is or isn't being taken for granted in the original statement.
The skill: identifying unstated premises. If someone says "we should hire more lawyers because we have too much work," the assumption being made is that more lawyers will reduce work — not that work needs reducing.
Strict syllogistic reasoning. Given premises, you decide whether each proposed conclusion follows necessarily — yes or no, no middle ground.
The trap: conclusions that are likely or probable but not necessarily true. If "all bankers wear suits" and "John is a banker," then "John wears a suit" follows necessarily. But if "most bankers earn over $100K" and "John is a banker," then "John earns over $100K" does NOT follow — most leaves room for exceptions.
You read a passage and evaluate proposed conclusions, deciding whether each conclusion follows beyond reasonable doubt.
The standard is intermediate between Inference (probabilistic) and Deduction (strict). Conclusions must follow with high confidence but need not be logically airtight.
Each item presents a controversial statement and proposed arguments for or against it. You rate each argument as Strong (directly relevant and important) or Weak (irrelevant, trivial, or based on personal preference).
The mindset: think like a judge weighing evidence, not like an advocate. Arguments based on emotion, anecdote, or off-topic relevance are Weak regardless of which side they support.
Because all five Watson-Glaser sections test genuinely different reasoning skills rather than variations on one skill, it helps to see an illustrative item from each — these are original examples written to show the format and reasoning style, not real Watson-Glaser questions.
Passage: "A city-wide survey found that 75% of respondents who commute by car reported longer average commute times than those who use public transit."
Proposed inference: "Public transit is faster than driving for most commutes in this city."
Correct rating: Probably True. The passage supports this reading but doesn't establish it definitively — the survey covers reported commute time for car users specifically, not a direct city-wide comparison of all transit modes, so "probably true" rather than "true" is the calibrated answer.
Statement: "We should extend the store's opening hours because sales have been flat this quarter."
Proposed assumption: "Extended hours would increase sales."
Correct rating: Assumption made. The statement's recommendation only makes sense if this assumption holds — the speaker is taking it for granted that more available hours translates into more sales, even though that link is never stated explicitly.
Premises: "All senior partners at the firm have practiced law for at least 15 years. Maria has practiced law for 12 years."
Proposed conclusion: "Maria is not a senior partner."
Correct rating: Conclusion follows. Since senior partner status requires at least 15 years and Maria has only 12, the conclusion follows necessarily from the premises — this is valid strict deduction, unlike the "most bankers" example above which only supports a probabilistic link.
Candidates sometimes assume Watson-Glaser is simply a rebranded verbal or logical reasoning test, but the skill it isolates — evaluating the validity of reasoning itself, independent of subject-matter knowledge — is genuinely distinct from standard verbal comprehension or abstract pattern-matching. A verbal reasoning test typically asks whether a statement is supported by a passage; Watson-Glaser goes a layer further, asking you to judge the soundness of the reasoning connecting evidence to a conclusion, which is exactly the skill legal argumentation and policy analysis require day to day. This is also why law firms and policy-focused civil service streams specifically favour Watson-Glaser over more generic aptitude batteries — it's a closer proxy for the actual on-the-job reasoning those roles demand than numerical or abstract reasoning would be.
Watson-Glaser's near-century of continuous use is unusual in psychometric testing, where most instruments get revised, rebranded, or replaced within a decade or two as validity research evolves. Its longevity comes down to the narrowness and stability of what it measures: critical-thinking skill, as isolated by the five-part structure above, doesn't shift with technology or workplace trends the way, say, "digital literacy" assessments would need to. Pearson (through its TalentLens division, after acquiring the rights) has periodically updated the item bank and short-form versions over the decades, but the underlying five-skill structure has remained essentially unchanged since Goodwin Watson and Edward Glaser's original design — which is also why the test is unusually well-documented in third-party prep material relative to newer, proprietary assessments.
Raw score out of 40, converted to percentile against a graduate norm group. Top law firms typically cut at the 75th-85th percentile. Civil Service Fast Stream analytical and policy streams typically require top-quartile performance (75th+).
| Employer | Typical cutoff |
|---|---|
| Magic Circle (Linklaters, A&O, CC, S&M, Freshfields) | ~80th percentile |
| Hogan Lovells, Herbert Smith Freehills | ~75th percentile |
| UK Civil Service Fast Stream (Policy) | ~75th percentile |
| US BigLaw with London offices | ~80th percentile |
Practice the five sections separately. Each section requires a different mental mode. Inference requires probabilistic thinking; Deduction requires strict logical thinking. Switching is hard if you haven't drilled each independently.
Watson-Glaser-specific prep books. The test is well-documented because of its age. Pearson sells official practice tests, and AssessmentDay, JobTestPrep, and PracticeAptitudeTests publish full Watson-Glaser practice batteries. Take 4-5 full timed practice tests before your real one.
Read carefully — twice if needed. Most wrong answers come from misreading the passage rather than wrong reasoning. The 40 minutes provides ~45 seconds per question, which is enough time to read each passage twice.
Beware over-confidence in Deduction. Strict deduction is unforgiving. When in doubt about whether a conclusion follows, default to "no" — most candidates over-include conclusions that "feel right" but don't strictly follow.
Policy varies by firm, but most Magic Circle and top law firms enforce a cooldown period before a repeat application — commonly six to twelve months — rather than an outright permanent ban, since firms generally recognise that candidates improve genuinely with practice and maturity. If your first attempt didn't meet a specific firm's cutoff, check that firm's stated reapplication policy directly rather than assuming either extreme (an instant retake or a permanent block), since both assumptions are common but neither is reliably accurate across firms. UK Civil Service Fast Stream applications have their own separate reapplication rules, generally allowing another attempt in a future recruitment cycle once a full year has passed.
Because Watson-Glaser is fundamentally a reading-and-reasoning test rather than a vocabulary test, strong general English reading comprehension matters more than a large specialised vocabulary — the passages are written in plain, direct language deliberately, since the test's purpose is to isolate reasoning skill from language sophistication. Candidates for whom English is an additional language sometimes find the Deduction section the most approachable of the five specifically because its logical structure is language-independent once the premises are understood, while Inference and Evaluation of Arguments lean more heavily on picking up subtle qualifying language ("likely," "some," "most") that changes the correct answer — these specific qualifying words are worth deliberately practising to recognise quickly, since misreading "some" as "most" or vice versa is one of the more common, avoidable sources of lost points regardless of native language.
Watson-Glaser Style Practice Pack
5 tests · 150 questions — Deduction, Interpretation, Recognition of Assumptions and Inference, with a worked explanation on every question.
Try 10 questions free →See the pack — $19.99