Watson Glaser
Updated July 1, 2026How do you pass the Watson Glaser test?
The Watson Glaser Critical Thinking Appraisal is a high-stakes psychometric assessment widely used by global law firms, investment banks, and professional services companies to evaluate analytical reasoning. In the UK, it is a gatekeeper for elite training contracts and graduate schemes; in the US, it screens applicants for summer-analyst programs and leadership tracks. Passing requires more than general intelligence. It demands a systematic mastery of formal logic and strict adherence to specific evaluation rules across five distinct subtests. Candidates who fail to understand these precise mechanical constraints are frequently eliminated before human recruiters ever see their CV or resume.
30 to 45 minutes
Typical time limit
Varies by test version and employer
30 to 40
Number of questions
Commonly reported across modern variants
Top 20% to 30%
Typical pass percentile
Approximate cut-off for elite firms
1 minute
Average time per question
Crucial pacing benchmark
Quick answer
To pass the Watson Glaser test, you must isolate the provided text from all real-world knowledge and evaluate statements using only formal logical rules. Success depends on mastering the unique decision criteria for each of the five subtests (Inference, Assumptions, Deduction, Interpretation, and Arguments) while maintaining a strict pace of roughly one question per minute.
Key points
- Accept every statement in the passage as absolute truth, even if it contradicts your existing real-world knowledge or professional expertise.
- Treat the five subtests as separate logical games, each requiring a completely different set of decision-making rules and truth scales.
- Identify and eliminate outside assumptions, emotional biases, and extrapolations that are not explicitly supported by the text.
- Maintain a strict pacing strategy, spending no more than 60 seconds on any single question to avoid running out of time.
The Golden Rule of the Watson Glaser Test
The single most common reason high-achieving candidates fail the Watson Glaser test is the introduction of external knowledge. This assessment does not measure general knowledge, moral judgment, or real-world business acumen. It measures your ability to operate within a closed logical ecosystem. You must accept every statement in the provided text as absolute truth, even if it states something factually incorrect about the real world. If a text claims that the moon is made of cheese, you must treat that as an undisputed scientific fact for the duration of that question.
Outside knowledge is an active liability on this test. Test developers deliberately design questions around hot-button social, economic, or legal topics because they know candidates have pre-existing opinions on them. When you read a passage about corporate taxation, environmental policy, or labor laws, your brain naturally attempts to fill in the blanks using articles you have read or concepts you studied at university. To pass, you must consciously suppress this instinct. Treat the passage like an isolated legal contract: if a claim is not explicitly written within the four corners of the text, it does not exist.
You must also develop extreme sensitivity to absolute qualifiers. Words like "all", "never", "always", "none", and "entirely" carry massive logical weight. Conversely, words like "some", "often", "many", "likely", and "sometimes" introduce probabilistic wiggle room. A conclusion that is "likely" true cannot be deduced as "absolutely" true. Paying close attention to these modifying words will help you differentiate between a conclusion that is genuinely guaranteed by the text and one that merely sounds plausible to a casual reader.
Mastering the Five Specific Subtests
The test is divided into five distinct modules, each requiring you to apply a different analytical framework. You cannot use the same thought process across the entire assessment; you must pivot your strategy as you transition from one subtest to the next.
Inference
In this section, you are given a statement of facts followed by several proposed inferences. You must classify each inference into one of five categories: True, Probably True, Insufficient Data, Probably False, or False. An inference is True only if the facts leave no room for any other interpretation. It is Probably True if, based on the facts, it is highly likely but not absolutely certain. Insufficient Data means you cannot tell either way because the text lacks the relevant clues. Probably False means the inference is unlikely given the facts, while False means it directly contradicts the provided text.
Recognition of Assumptions
Here, you are presented with a statement followed by several suggested assumptions. Your job is to determine whether the speaker has made that assumption or not. An assumption is something taken for granted or presupposed from the outset. It is an unstated premise that must be true in order for the stated argument to make sense. If the statement says, "We must upgrade our software to increase productivity," the speaker assumes that the new software is capable of increasing productivity. If the statement can stand logically without the assumption, then the assumption is Not Made.
Deduction
In the Deduction subtest, you read a short passage containing premises, followed by suggested conclusions. You must decide if each conclusion "Follows" or "Does Not Follow" from the text. This is pure formal logic. If the premises are true, does the conclusion naturally and inevitably follow? Do not look for probability here. If there is any scenario where the premises are true but the conclusion could be false, then the conclusion does not follow. Watch out for invalid conversions, such as assuming that because "all corporate lawyers work long hours," anyone who works long hours must be a corporate lawyer.
Interpretation
This section requires you to judge whether a given conclusion follows beyond a reasonable doubt from the evidence provided. While similar to Deduction, Interpretation allows for a small degree of practical, real-world reasoning, but it remains highly restrictive. You must weigh the information presented and decide if the suggested conclusion is a logical consequence of that data. If the text states that 90% of students who attended a seminar improved their grades, a conclusion stating that the seminar was beneficial for the majority of attendees follows beyond a reasonable doubt.
Evaluation of Arguments
In this final section, you must determine whether proposed arguments are "Strong" or "Weak" in relation to a specific question. A strong argument must be both highly relevant to the core issue and deeply important. A weak argument is one that is trivial, not directly related to the question, or confuses a side issue with the main point. You must remain entirely objective here. An argument can be incredibly strong logically even if you personally disagree with the viewpoint it represents, or weak even if it aligns perfectly with your personal ethics.
Navigating the Extreme Time Pressures
The modern Watson Glaser test is a race against the clock. While exact limits depend on the specific version deployed by an employer, a very common format provides roughly 30 to 40 questions to be answered in 30 minutes. This leaves you with an average of less than 60 seconds per question. Because the passages are dense and the logical distinctions are subtle, it is incredibly easy to lose track of time while debating a single ambiguous inference.
To prevent running out of time on the final sections, you need a strict pacing routine. If you spend more than 75 seconds on a single question, make an educated guess based on your initial logical analysis and move on. The test does not penalize incorrect answers via negative marking, so leaving a blank space is a wasted opportunity. Furthermore, the questions are not necessarily arranged in order of difficulty; getting stuck on a hard Deduction question early on might prevent you from reaching straightforward Argument questions at the end of the test.
Why General Reading is Insufficient Prep
Many candidates assume that because they read complex texts for their university degrees or current jobs, they do not need targeted practice for the Watson Glaser. This is a dangerous mistake. Reading a law journal, an economic report, or a financial news article requires you to synthesize information, draw real-world connections, and consider broader contexts. The Watson Glaser requires you to do the exact opposite. It demands that you shrink your analytical field of vision exclusively to the provided text.
Practising the exact format of the assessment trains your brain to recognize the specific traps laid by test designers. For example, interactive practice via online systems like Intervyo allows you to experience the exact layout, countdown timers, and interface transitions you will encounter on the day of the assessment. Regular drills help build the mental stamina required to maintain absolute logical objectivity for 30 consecutive minutes without slipping back into habitual, real-world thinking patterns.
How it works
How the Watson Glaser is scored
The Watson Glaser test operates on a standardized psychometric scoring model. In most modern applications, your raw score (the total number of correct answers) is converted into a percentile rank relative to a specific norm group. This norm group is typically composed of individuals with a similar profile, such as UK training contract applicants, US law school graduates, or global executive candidates. This means that getting 28 out of 35 correct might be a passing score against a general graduate norm group, but a failing score when measured against a hyper-competitive magic circle or Wall Street cohort.
Some advanced online testing platforms use an Item Response Theory (IRT) model, which can be adaptive. In an adaptive test, the difficulty of the next question is adjusted based on whether you answered the previous question correctly. Under an IRT framework, your final score is determined not just by how many questions you got right, but by the difficulty level of those questions. This prevents candidates from gaming the system by simply rushing through hard questions to collect easy points elsewhere.
Employers do not see a simple pass or fail mark; they receive a detailed report. This report includes your overall percentile rank alongside sub-scores for each of the core critical thinking domains. If a firm requires high performance in risk management, they may look closely at your Deduction and Inference scores. The exact pass mark is entirely at the discretion of the employer and changes from year to year based on applicant volume, but elite firms commonly set their cut-offs around the 70th to 80th percentile.
Anti-cheating protocols on modern test platforms are extensive. When taking the test at home, you are often monitored via webcam, microphone, and browser-locking software to prevent you from opening other tabs or using external tools. The system also tracks your behavioral biometrics, such as your response latency (the exact time you spend on each question). If a candidate answers a highly complex deduction logic puzzle in two seconds, the system flags the attempt for manual review due to anomalous behavior patterns.
How to prepare
- 01
Study the formal definitions of each subtest
Memorize the exact criteria for a true inference versus a probable inference, and learn what constitutes a formal logical assumption.
- 02
Take a non-timed diagnostic test
Work through a small set of practice questions slowly, focusing entirely on accuracy and understanding why incorrect options are wrong.
- 03
Transition to timed practice sessions
Use a simulated testing environment to build your pacing skills until you can naturally average 50 to 60 seconds per question.
- 04
Review every mistake systematically
Keep a log of your incorrect answers to identify whether you are consistently struggling with a specific subtest or falling for outside-knowledge traps.
A preparation timeline
Two weeks before
Take a diagnostic test to identify your weakest subtests and review formal logic rules.
One week before
Run full-length, timed practice tests to build your pacing habits and focus heavily on eliminating external knowledge.
The day before
Rest your mind, review your personal log of common logical traps, and check your testing hardware and internet connection.
During the test
Track your time closely, treat every text as an absolute truth, and do not leave any question blank.
How candidates approached it
Anonymised accounts of how recent applicants prepared, what they experienced, and how it turned out.
Commercial Law / UK Market / Pass
Experience. I applied for a highly competitive training contract at a City firm in London. On my first attempt at the Watson Glaser without preparation, I scored in the 45th percentile because I used my university law knowledge to evaluate the arguments. For my next application, I practiced isolating the text entirely and memorized the exact rules for the Recognition of Assumptions module.
Outcome. My score jumped to the 85th percentile, and I advanced straight to the assessment centre.
Investment Banking / US Market / Fail
Experience. I was navigating the summer-analyst pipeline for an investment bank in New York. I assumed my strong math and finance background would carry me through the Watson Glaser, so I did not use any targeted practice platforms beforehand. I got bogged down in a complex Deduction section, panicked about the clock, and ended up guessing the last eight questions blindly.
Outcome. I missed the firm's strict percentile cut-off and was automatically rejected from the process.
Questions to practise
A bank of adjacent questions candidates run into. Drill each one in the exact format firms use.
- Is an inference considered probably true if it is supported by common sense but not explicitly stated in the text?
- What is the structural difference between a deduction question and an interpretation question?
- How can I tell if an argument is weak because it is irrelevant or weak because it is unimportant?
- Does the Watson Glaser test utilize negative marking for incorrect answers?
- Can an assumption be explicitly stated within the main text of the question?
- What is the average percentile cut-off score for magic circle law firms in the UK?
- How does the Item Response Theory scoring model affect my strategy on difficult questions?
- What should I do if a passage directly contradicts an established scientific fact that I know to be true?
- How do absolute qualifiers like all or none alter the validity of a deduction conclusion?
- What is the most effective way to manage my time if I get stuck on an inference question?
This answer is general guidance for orientation, not a guarantee. Test formats, timings and employer cut-offs change, so verify the details on the provider or employer site before you apply. Last updated July 1, 2026.