The Watson Glaser test asks you to judge arguments, not do maths. It has five question types, a tight clock, and a scoring pattern you can learn. Here is how each section works and where candidates lose marks.
Most candidates meet the Watson Glaser through a graduate scheme in law, consulting, or a risk and compliance function at a bank. It arrives as a timed online test, usually before the first interview, and the score is reported as a percentile against a norm group.
The format is unusual enough that strong students often underperform on the first attempt. They read for general meaning, answer from intuition, and miss the precise wording that separates one option from another. This guide explains each question type, shows how the traps work, and gives you a practice routine. If the test sits before a live assessment day, our assessment centre tips guide covers that next stage.
What the Watson Glaser is actually measuring
Critical thinking, in the test's terms, is the ability to evaluate reasoning rather than to agree with a conclusion. That means noticing when a statement goes beyond the evidence, when an argument depends on an unstated premise, and when two conclusions are both defensible but only one follows from the passage.
The test is not a personality quiz and not a finance test. A candidate with no commercial background can score highly if they read carefully. A candidate with strong opinions can score badly if they import those opinions into the answer.
That is the first habit to break: answer only from the passage in front of you.
Inference: what probably follows
Inference questions give you a short statement and ask whether a proposed conclusion is true, probably true, insufficient data, probably false, or false.
The key word is probably. You are not looking for certainty. You are judging whether the evidence makes the conclusion more likely than not.
A worked example:
Passage: Four of the five analysts on the team worked late every night this week. The team submitted its report on Friday morning.
Proposed inference: At least one analyst did not work late every night this week.
This is true, because four of five is stated. A weaker inference such as "the team was under pressure" is probably true, since the evidence points that way without proving it. An inference such as "the report was late" is insufficient data, because the passage says nothing about the deadline.
The common error is choosing "probably true" when the passage already proves the statement. Read the quantifier carefully.
Recognition of assumptions: the unstated premise
Assumption questions ask whether an argument depends on a stated assumption. An assumption is something the argument takes for granted, without saying it.
Argument: The bank should hire more graduates because its analyst retention rate is falling.
Assumption: Hiring graduates will improve retention.
That assumption is necessary for the argument to work. If hiring graduates made retention worse, the recommendation would not follow. If the assumption is not necessary, the answer is that the argument does not depend on it.
The test is whether the argument collapses without the assumption. If it survives, the assumption is not required.
Deduction: what must follow
Deduction questions give you premises and ask whether a conclusion follows with certainty. These are closer to formal logic than the other sections.
Premises: All members of the deal team have passed the modelling test. Priya is a member of the deal team.
Conclusion: Priya has passed the modelling test.
This follows. Now change the second premise to "Priya has passed the modelling test" and the same conclusion no longer follows, because other people may also have passed it.
Watch for reversed relationships and for conclusions that swap "all" with "some". The wording is doing the work.
Interpretation: what the evidence supports
Interpretation questions give you a short scenario and ask whether a conclusion is follows, does not follow, or insufficient data.
The distinction from inference is subtle. Interpretation asks whether the conclusion is a reasonable reading of the data presented, not whether it is probably true in the world. Stay close to what the numbers or statements actually show.
A useful habit is to restate the conclusion in your own words, then ask: would a careful reader of this passage agree, disagree, or need more information?
Evaluation of arguments: strong or weak
This section gives you a question, two arguments, and asks which argument is stronger. The content can be political, commercial, or everyday. Your own view is irrelevant.
A strong argument is relevant, uses evidence rather than assertion, and addresses the question directly. A weak argument may be emotional, circular, or about a side issue.
Ask three questions of each argument:
- Does it answer the question asked?
- Does it give a reason, or just restate the conclusion?
- Would the reason still matter if the opposite conclusion were true?
That third question catches arguments that sound persuasive but prove nothing.
Timing and scoring
The sections are short and the clock is tight. Most candidates cannot finish comfortably, which is intentional: the test wants to see judgement under pressure as well as accuracy.
A workable pace is around 45 to 50 seconds per question, leaving a minute at the end for review. If a passage is long, read the question first so you know what you are looking for.
Scores are norm-referenced, which means your percentile depends on the comparison group. A raw score that looks modest can still be a strong percentile if the norm group found the test hard. You will not usually see the breakdown, only whether you progressed.
A two-week practice routine
- Day 1: Learn the five question types and the exact meaning of each answer option.
- Day 2: Do 20 inference items and review every mistake.
- Day 3: Do 20 assumption items, writing out the unstated premise for each.
- Day 4: Do 20 deduction items with a diagram or shorthand notation.
- Day 5: Do 20 interpretation items and justify each answer from the passage.
- Day 6: Do 20 argument evaluation items and rank the arguments before checking.
- Day 7: Take a full timed set and record your score.
- Week two: Alternate full timed sets with targeted practice on your weakest type.
Our verbal reasoning test guide covers the true, false, and cannot say format that overlaps with the deduction and interpretation sections. If your employer uses a broader critical thinking test rather than the branded Watson Glaser, the same five skills apply, and the psychometric test investment banking guide explains how the formats sit together.
How this fits the rest of your application
The Watson Glaser is usually one stage in a longer process. After the test comes a video interview or assessment centre, and the same precision that helps here also helps there. Our HireVue interview questions guide covers the recorded stage, and the psychometric test investment banking guide maps the other tests you may meet.
If you are applying to several firms at once, keep the test dates and formats in one place. Different employers use different providers, and practising the wrong format is a common waste of a week.
What to do after reading this
Before your next critical thinking test, learn the five question types, then do one timed set and review every error. Focus on the type where your accuracy drops, not the one you enjoy.
If assessment season is stacking up, Finbound's study board ranks practice by the applications and deadlines you are tracking, so the test you have next week comes first. Start for free to add your first three applications.