The Watson-Glaser test, section by section

The Watson-Glaser Critical Thinking Appraisal, from Pearson TalentLens, is 40 questions in 30 to 40 minutes on short written passages. Its five subtests are drawing inferences, recognizing assumptions, deducing, interpreting and evaluating arguments, and you answer from the information given, not what you know. The Profile Report gives a percentile against a norm group, overall and for three areas: Recognize Assumptions, Evaluate Arguments and Draw Conclusions. For logic under a clock, Soreno, an AI voice interview practice app for consulting and finance candidates, has a free brain teaser trainer with no sign-up for the interviews that often follow the test.

Updated

  • 40 questions, 30 to 40 minutes. That is 45 to 60 seconds a question, reading included.
  • Five subtests. Inferences, assumptions, deduction, interpretation and arguments.
  • Only the passage counts. Each subtest judges what follows from the information given.
  • Scored as percentiles. Against a norm group, overall and for three areas.
  • Free practice, unscored. TalentLens's practice test is in English and Dutch and gives advice, not a score.

What is the Watson-Glaser test?

A critical thinking test that TalentLens describes as measuring the ability to analyze arguments, evaluate information and make objective decisions. Employers use it in recruitment and development for graduates, professionals, supervisors, managers and executives, and TalentLens says university selection teams use it too.

Format

Short written passages with a statement to judge.

Questions
40
Time
30 to 40 minutes
Languages
US and UK English, French, Dutch, Spanish and more

Delivery

Online, from an item bank, with instant results for the employer.

Setting
Unsupervised or supervised
Questions
Item-banked, so tests differ
Fairness
Difficulty evened out in scoring

Reports

Three reports the employer can use.

Profile
Percentiles, overall and by area
Development
A learning plan
Interview
A structured interview with questions
The five subtestsTalentLens definitions

TalentLens calls these the building blocks of critical thinking, with questions of varying difficulty and format. The definition is TalentLens's, the tip after it is ours.

  1. 1

    Drawing inferences

    Rating the probability that an inference is true, based on the information given. Our tip: it's a scale of likelihood, so outside knowledge and a hunch about causes both pull you off.

  2. 2

    Recognizing assumptions

    Identifying unstated assumptions or presuppositions under a statement. Our tip: ask whether the statement would still make sense if the assumption were false. If not, it was assumed.

  3. 3

    Deducing

    Deciding whether a conclusion follows logically from the information. Our tip: it must follow every time. A rule that says all X are Y says nothing about what else is Y.

  4. 4

    Interpreting

    Weighing the evidence and deciding if a generalization or conclusion from the data is warranted. Our tip: check the group and the claim. A finding about some people rarely covers everyone.

  5. 5

    Evaluating arguments

    Judging the strength and relevance of arguments on a question. Our tip: strong means relevant and important to the question, not that you agree.

    TalentLens: Watson-Glaser
Telling the three conclusion subtests apartOurs

Inferences, deducing and interpreting all ask about a conclusion, and they read alike. This is how we separate them, by the standard of proof each asks for.

  1. Inferences

    How likely, from these facts?

    A degree of probability. The trap is reading a cause into two things that happened together.

  2. Deducing

    Must it follow?

    Strict logic. The trap is reversing a rule: all big deals are reviewed does not mean every reviewed deal is big.

  3. Interpreting

    Does the data support this claim?

    Warranted or not by the evidence given. The trap is a claim wider than the data, a new group or a cause the study never tested.

Pacing, computed from the formatOurs

40 questions in 30 or 40 minutes. Check your invite for which.

  1. Per question

    45 to 60 seconds

    That includes reading the passage, so read it once with care before you look at the statement.

  2. Checkpoints

    Look at the clock three times

    At questions 10, 20 and 30. If you are behind, move faster on the next passage rather than rushing the last one.

  3. Where the clock should be (our checkpoints)

    Where the clock should be (our checkpoints)
    Question30-minute test40-minute test
    107.5 min10 min
    2015 min20 min
    3022.5 min30 min
How the score is reportedThree reports

TalentLens's FAQ lists three reports. Your result is read against other people, not as a raw mark.

  1. Profile

    Percentiles against a norm group

    An overall percentile plus one for each of Recognize Assumptions, Evaluate Arguments and Draw Conclusions. Number Correct, T-score, STANINE and STEN sit in its technical section. Employers can choose from a wide range of norm groups, or map a company norm for a fee.

  2. Fairness

    Tests differ, scoring evens it out

    Each item is coded by difficulty. Tests may differ slightly in difficulty overall, which TalentLens says the scoring accounts for. Passages are coded by theme, such as Business, Science and Education, and every candidate gets enough business passages.

  3. Development

    Strength, area to deepen, development opportunity

    TalentLens's sample Development Report, a French-language sample, compares each area with people in your profession and rates it as a strength to build on (top of the group), an area to deepen (the middle) or a development opportunity (the bottom).

    TalentLens: sample Development Report (French)
  4. Interview

    Questions on your thinking

    The Interview Report helps an interviewer run a structured critical thinking behavioral interview, with sample questions. Expect to be asked how you test assumptions.

From TalentLens: Watson-Glaser Critical Thinking Appraisal, TalentLens: free critical thinking practice test and TalentLens: Watson-Glaser III sample Development Report (French).

What does the Watson-Glaser measure?

TalentLens defines critical thinking as understanding a situation from several perspectives while separating facts from opinions and assumptions. The Profile Report scores it in three areas.

  • Recognize Assumptions

    Spotting what a statement takes for granted.

    Show what they listen for
    What interviewers listen for
    Our read: the gap between what is said and what has to be true for it to make sense.
  • Evaluate Arguments

    Judging whether an argument is strong and relevant to the question.

    Show what they listen for
    What interviewers listen for
    Our read: weak arguments are off topic, trivial or lean on feelings rather than reasons.
  • Draw Conclusions

    Covers the inference, deduction and interpretation subtests: what follows, and how firmly, from the evidence.

    Show what they listen for
    What interviewers listen for
    Our read: no wider than the data, and no outside knowledge.

8 practice items for the Watson-Glaser test

All written by our team, with invented companies. None are TalentLens items, and the answer choices are ours. Work from the words in each prompt only.

One per subtest5 items

Answer, then check the reason.

  1. Inference. Larkmoor Bank moved its 120 loan officers to new software in Q3. Average loan turnaround fell from 14 to 11 days that quarter. Inference: the software caused the faster turnaround. True, probably true, not enough information, probably false or false?

    Our answer and why

    Not enough information. The two happened in the same quarter, but nothing in the passage rules out other causes, such as staffing or the loan mix.

  2. Assumption. "We must cap associates at 55 billed hours a week so first-year attrition stops rising." Is it assumed that working over 55 hours contributes to attrition?

    Our answer and why

    Yes, it is assumed. If long hours had no link to attrition, capping them to stop attrition would make no sense.

  3. Deduction. Every cross-border merger above $500 million is reviewed by the firm's antitrust group. Last month the group reviewed the Helios-Orion deal. Does it follow that Helios-Orion is a cross-border merger above $500 million?

    Our answer and why

    No. The rule says what the group must review, not that it reviews nothing else. Reversing the rule is the classic trap.

  4. Interpretation. A survey of 400 partners found 68% of those using contract-review software billed more hours than the firm median. Conclusion: the software makes any lawyer bill more. Warranted or not?

    Our answer and why

    Not warranted. It is one survey of partners showing an association. It says nothing about cause or about associates.

  5. Argument. Should banks require analysts in the office five days a week? "No, because commuting is tiring for people who like to sleep in." Strong or weak?

    Our answer and why

    Weak. It is on topic but rests on a minor preference, not on output, training, clients or retention.

Interview follow-ups3 items

The kind of questions a critical thinking interview asks.

  1. Tell me about a recommendation you made when a key assumption hadn't been checked.

    Show the sample answer

    Sample answer. Example answer written by our team, the applicant is invented. On a lending internship, a borrower's forecast assumed fuel costs would fall 6% after new trucks arrived, but the new trucks only replaced short-haul vans. I flagged it, reran the model with flat long-haul costs, and coverage dropped below our floor. We asked for route data and added a tighter covenant before the credit committee approved it.

    Name the assumption, how you tested it, what changed in your answer and the result.

  2. Tell me about a time you disagreed with your manager.

    Our answer and why

    The evidence you brought, how you raised it and what was decided. Show you weighed their argument too. Practice it, see a model answer

  3. Tell me about a decision you made with incomplete information.

    Our answer and why

    What you knew, what you didn't, how you sized the risk and what you'd check next time. Practice it, see a model answer

Prep plan and FAQ

A one-week plan for the Watson-Glaser

Ours, built around what TalentLens publishes.

  1. Day 1

    Take TalentLens's free practice test

    In English or Dutch. It shows what each subtest looks like. TalentLens says it can't score it, but you get a report with advice on preparing.

    From TalentLens: practice test.

  2. Days 2 to 4

    One subtest a day

    Do our five items below, then write two of your own per subtest from news articles. Say why each answer is right in one sentence.

  3. Day 5

    Timed run

    40 questions at 45 seconds each, with checkpoints at questions 10, 20 and 30.

  4. Test day

    Read only what is on the screen

    Read each subtest's instructions. Answer from the passage alone, even when you know the industry.

  5. After

    Get ready for the interview

    If the employer uses the Interview Report, expect questions on how you test assumptions and weigh evidence. Practice two stories out loud.

After the Watson-Glaser, the interview

Practice behavioral answers and cases out loud with Soreno's AI interviewer, which asks follow-up questions and scores every answer, quoting your own words.

Questions people ask

How many questions are on the Watson-Glaser test?

TalentLens gives the format as 40 questions in 30 to 40 minutes. That works out to 45 to 60 seconds a question by our count.

What are the five sections of the Watson-Glaser test?

Drawing inferences, recognizing assumptions, deducing, interpreting and evaluating arguments. The Profile Report groups them into Recognize Assumptions, Evaluate Arguments and Draw Conclusions.

How is the Watson-Glaser scored?

As percentiles against a norm group, overall and for the three areas. Number Correct, T-score, STANINE and STEN appear in the report's technical section.

Is every Watson-Glaser test the same?

No. Questions come from an item bank. TalentLens says each item is coded by difficulty and small differences between tests are accounted for in scoring.

Is there a free official Watson-Glaser practice test?

Yes, on TalentLens's site in English and Dutch. TalentLens says it can't score it meaningfully, but you get a report with advice on preparing for the live test.

Who uses the Watson-Glaser test?

TalentLens lists it for graduates, professionals, supervisors, managers and executives, and lists Law and Banking & Finance among the industries on its site. University selection teams use it too.

Keep going