Reviewed 19 August 2026 by the Cognivy editorial team
On a verbal reasoning test an answer can be right about the world and still wrong about the passage, and only the passage counts. This guide covers the response formats providers actually use, why true, false and cannot say is not universal, a method for reaching a verdict in under a minute — and six free questions to try it on.
- Questions
- 20 to 50, depending on provider
- Time limit
- 12 to 25 minutes, depending on provider
- Answer format
- True / false / cannot say, multiple-choice comprehension, verbal analogies, or word relationships
- Prior knowledge
- Not tested. Answers come from the passage alone, and outside knowledge is usually a liability
- Scoring
- Percentile against a comparison group. Providers do not publish pass marks
The skill
What is a verbal reasoning test?
A verbal reasoning test gives you written information — a policy extract, a report paragraph, a set of data sheets — and asks what it supports. Most questions are not vocabulary quizzes. They test whether you can hold the line between what a text states and what you assume alongside it.
The best-known form presents a passage followed by statements you mark true, false or cannot say. SHL Direct describes exactly that task, and Aon uses the same three-option response set on its verbal module. But it is one option among several: Pearson's DAT includes verbal analogies, Criteria's CCAT verbal section uses word relationships, and Saville separates verbal analysis from verbal comprehension.
Prior knowledge is not being tested and is usually a liability. If reaching your answer needed something the passage does not say, the answer is wrong even when it is true in the real world.
A verbal reasoning test usually asks one question: what does this passage support. A critical-thinking assessment such as Watson-Glaser splits that judgement into five separate sub-tests — inference, assumptions, deduction, interpretation and argument evaluation — each with its own decision rule. If your invitation names Watson-Glaser, the critical-thinking guide is the right starting point.
Try it before anything else
Free verbal reasoning practice questions
Six original questions covering the formats above: the three-verdict discipline first, then inference, assumptions and a detail question. Commit to an answer before opening the solution — the solution shows the reasoning and names why every wrong option is on the list.
A short extract from an internal expenses policy.
All expense claims submitted before the 25th of the month are paid in that month's payroll run. Claims submitted after the 25th are paid in the following month. Claims relating to international travel require line-manager approval before submission.
Statement: a domestic expense claim submitted on the 26th will be paid in the following month's payroll run.
Show the worked solution
Answer: A — True.
- The deciding sentence is the second one: claims submitted after the 25th are paid in the following month.
- The 26th is after the 25th, so the rule applies to this claim directly.
- The approval clause covers international travel and this claim is domestic, so it does not apply.
- Nothing had to be added to reach the verdict, which is what separates true from cannot say.
Why the other options are there. The third sentence exists to make you hesitate. An irrelevant condition is not a missing condition, and passages routinely include one to see whether you can tell the difference.
The same policy extract.
All expense claims submitted before the 25th of the month are paid in that month's payroll run. Claims submitted after the 25th are paid in the following month. Claims relating to international travel require line-manager approval before submission.
Statement: international travel claims are paid more slowly than domestic claims.
Show the worked solution
Answer: C — Cannot say.
- The passage sets payment timing by submission date alone, and applies that rule to all claims without distinction.
- It adds an approval step for international travel, but says nothing about how long approval takes.
- Concluding that approval causes a delay requires information the passage does not supply.
- It is not false either, because the passage does not deny it. Silence is cannot say.
Why the other options are there. This is the plausible-and-unsupported case. Approval almost certainly does slow things down in real life, which is precisely why the item works. Being right about the world is not the task.
A one-paragraph note from an office-operations update.
Standard office hours are 9.00am to 5.30pm, Monday to Friday. The Helsinki office is the exception: it closes at 4.00pm on Fridays. Meeting rooms in every office can be booked up to 14 days ahead on the intranet, and bookings that fall after a site's closing time are cancelled automatically.
Statement: the Helsinki office keeps standard hours on Fridays.
Show the worked solution
Answer: B — False.
- The statement names Helsinki and Fridays, so search for those words instead of reading end to end. They appear in sentence two.
- Sentence one defines standard hours as 9.00am to 5.30pm. Sentence two makes Helsinki the exception, closing at 4.00pm on Fridays.
- The passage asserts the opposite of the statement, and the contradiction can be quoted. That is the false test: a sentence that says the reverse.
- The booking rule plays no part. As in sample question 1, one clause exists purely to be irrelevant.
Why the other options are there. Cannot say catches candidates who have learned to distrust confident verdicts and started using it as the safe default. It is only correct when the passage is silent, and this passage answers the question directly. True is the skim reader's answer: the exception sits in the middle of the passage, exactly where skimming misses it.
An extract from an engineering team's quarterly report.
After the platform team moved its weekly release from Friday afternoon to Thursday morning, the following quarter produced fewer out-of-hours incident calls than any of the previous three quarters. The move had been proposed by the on-call engineers, who argued that Friday releases left faults to surface over the weekend, when fewer staff were available to respond.
Which conclusion does the passage best support?
Show the worked solution
Answer: C — Out-of-hours incident calls fell in the quarter after the release day moved.
- Rank the options by what the passage forces, not by what it invites you to believe.
- The passage states two facts: the release day moved, and the following quarter produced fewer out-of-hours calls than any of the previous three. Option C restates the second fact and adds nothing.
- Option A goes further than the passage. Events in sequence are not events in a causal chain, and nothing in the text rules out other explanations for the fall.
- Options B and D reach outside the record entirely: the engineers' argument is reported, not confirmed, and staffing after the change is never mentioned.
Why the other options are there. A is the reflex answer, because a fall that follows a change reads as a fall the change produced. The passage records the sequence and stops there. B pays the engineers a compliment the passage never issues: their argument is quoted, not vindicated. D is assembled from the weekend-staffing clause, which describes the problem they argued about, not anything measured afterwards.
A decision recorded in an internal communications update.
Attendance at the optional Friday briefing has halved since it moved from the canteen to a video call. Interest in the briefing's content has clearly declined, so from next month the briefing will be replaced by a monthly summary email. Staff who want detail can request the full notes from the communications team.
The argument concludes that interest in the briefing's content has declined. Which assumption does it rely on?
Show the worked solution
Answer: B — The fall in attendance is not explained by the move to a video call itself.
- Separate the parts. Evidence: attendance halved after the briefing moved online. Conclusion, named in the question: interest in the content has declined.
- Two things changed at once — the format and the attendance. The conclusion blames the content, so the argument must be assuming the format change itself is not what emptied the call.
- Test each option by negating it. If the argument survives the negation, it was not relying on that assumption.
- Negate B — suppose the move to video is itself the explanation — and the halving no longer says anything about interest in the content. The argument falls, so B is the assumption.
Why the other options are there. A feels structural, but a change in content would still be consistent with the conclusion — worse content producing less interest is exactly what the conclusion describes. C fails on one word: the argument needs attendance to be evidence, not the best available measure of anything. D belongs to the recommendation about the email, which is a different claim from the one the question pins down.
An extract from an HR onboarding page.
The new starter checklist runs in three stages. IT accounts are created before the start date, and building access cards are issued on the first morning. Payroll registration closes on the 20th of each month; starters registered after that date receive their first salary the following month. Line managers complete the probation plan within two weeks of the start date.
According to the passage, when are building access cards issued?
Show the worked solution
Answer: D — On the new starter's first morning.
- Read the question first: it asks one thing, the timing attached to access cards. Search for that phrase; it appears once, in sentence two.
- The sentence carries two timings — IT accounts before the start date, access cards on the first morning. The question asks about the second.
- Match the option that restates the clause without stretching it: issued on the first morning.
- Confirm the wrong options come from elsewhere in the passage. In detail questions they usually do, and knowing that is what stops a familiar phrase being matched to the wrong object.
Why the other options are there. Every wrong option is a genuine timing from the passage attached to the wrong thing: A belongs to IT accounts, in the same sentence as the answer, B to payroll registration, C to the probation plan. Detail questions are placed to be answered quickly, and what they harvest is candidates matching a phrase they remember without checking what it modified.
A short verbal reasoning set at the real per-question pace, scored server-side like the full product, with a worked solution after every answer.
Start the timed free setDraws on your 20 free practice questions. No account needed.
What employers are looking at
What do verbal reasoning tests measure?
Employers use verbal reasoning to predict how someone handles written instructions, policy, contracts, briefs and reports: the material most office roles run on. Aon states that its verbal assessment measures the ability to analyse and understand written information and evaluate the truthfulness of statements on the basis of the information given.
Three distinct abilities are separated. Comprehension: did you understand the text. Inference discipline: can you tell a supported conclusion from a merely plausible one. Pace: can you do both at the speed the job implies.
Reading your result
How verbal reasoning tests are scored
What reaches the employer is rarely a raw count of correct verdicts. Providers convert results into a comparison against other people, which changes what a good score means and what is worth optimising for on the day.
- Results are percentiles, not marks out of ten. Verbal results are normally reported as a percentile against a comparison group. A 70th-percentile result means you did better than 70% of that group, whatever the raw count was.
- No provider publishes a pass mark. Korn Ferry states directly that it does not set one and that any benchmark belongs to the hiring organisation. If a recruiter has disclosed a target, that is the only target worth working to.
- The norm group shapes the number. The same performance scores differently against graduate applicants than against the general population. Comparing your result with a friend's from a different assessment, or a different employer's comparison group, says very little.
- A guessed verdict tells the employer very little. On a three-option format a guess is right a third of the time, which is why accuracy usually matters more than volume. Where the provider says completion is not expected — Aon lists 49 tasks in 12 minutes and says exactly that — unanswered questions are designed in rather than a failure.
- On an adaptive test, difficulty is not feedback. Korn Ferry states that all of its cognitive ability tests select the next question from your previous response, so a harder passage often follows a correct answer. Treat a difficult run as normal behaviour rather than as a verdict on how you are doing.
Question formats
What do the questions look like?
Response format matters more than passage topic. Find out which one your assessment uses before you practise.
- True, false, cannot say. A passage or data sheet followed by statements. SHL Direct describes this task directly, and Aon's scales verbal uses the same three answer options with exactly one correct answer per statement.
- Multiple-choice comprehension. A passage followed by questions with four or five options — main point, specific detail, or what the author implies. Common inside mixed batteries.
- Verbal analogies. Word pairs where you identify the relationship and apply it to a new pair. Pearson's DAT Verbal Analogies is the named example in Cognivy's launch scope.
- Word relationships and vocabulary. Synonyms, antonyms and best-fit words at speed. The CCAT's verbal category works this way, interleaved with the test's other question categories.
- Instruction-based verbal items. A set of written rules and cases to classify against them. Aon publishes a separate module for this, and it is not the same task as a passage test.
- Argument and evidence evaluation. Judging whether a conclusion is supported, or whether an argument is both relevant and important. In its five-part form this becomes critical thinking rather than general verbal reasoning.
A two-paragraph extract from an expenses policy states which claims are reimbursed, the deadline for submitting them, and one exception covering international travel. Four statements follow. Two are settled directly by a single sentence. One is contradicted. The fourth describes a case the policy simply never addresses, and that is the one that has to be settled by noticing an absence rather than by finding a sentence.
Launch coverage
Which providers use verbal reasoning tests?
Verbal assessments from different providers can look near-identical on screen and score you under different rules. Match the name in your invitation.
- Verify Verbal Ability. SHL Direct describes the classic task as a passage of text followed by one or more statements, answered true, false or cannot say. SHL also publishes reading comprehension and verbal ability under separate product names, so check which one you were sent.
- scales verbal. Multiple verbal data sheets with statements marked TRUE, FALSE or CANNOT SAY, one correct answer each. Aon lists the module at 12 minutes and 49 tasks, precedes it with untimed examples, and says candidates are not expected to answer everything.
- Talent Q Elements Verbal. Adaptive delivery with a 16-minute limit. Korn Ferry states that every one of its cognitive ability tests selects the next item from your previous response.
- DAT Verbal Analogies. A relationship format rather than a passage format, delivered with computerised adaptive testing. Pearson's Watson-Glaser sits alongside it as a five-part critical-thinking assessment with its own verdict rules.
- Verbal Analysis. Saville separates verbal analysis from verbal comprehension and usually delivers verbal inside a Swift battery. Swift Analysis Aptitude runs 18 to 24 minutes across several aptitudes.
- Logiks Verbal (Advanced). Available as a standalone ability or as one section of Logiks General. Talogy publishes the range at 12 to 25 minutes and 20 to 50 items depending on version.
- CCAT. Verbal items are interleaved with maths and spatial questions in one 50-question, 15-minute test, so the verbal skill needed there is speed on word relationships rather than passage analysis.
Provider names identify the assessment format. Cognivy is independent and is not affiliated with or endorsed by these providers.
What varies, and what does not
Timing, verdict rules and scoring
As with numerical, the honest answer to how long it takes is which one did they send you. These are the published figures for the launch providers.
- Published time limits vary widely. Aon lists scales verbal at 12 minutes for 49 tasks. Korn Ferry lists Elements Verbal at 16 minutes. Criteria's CCAT is 15 minutes for 50 mixed questions.
- Task volume can exceed the time on purpose. Forty-nine tasks in twelve minutes is not an accident. Aon tells candidates directly that they are not expected to answer all the questions and should work quickly and accurately instead.
- Adaptive verbal tests exist. Korn Ferry states that all of its cognitive ability tests are adaptive. An adaptive test may follow a correct answer with a harder item, but you cannot read your score from how hard the passages feel, so treat a harder run as normal behaviour rather than as feedback.
- Cannot say is not offered everywhere. Analogy, vocabulary and comprehension formats have no cannot-say option at all. Practising three-verdict logic for a four-option comprehension test wastes the run-up.
A repeatable approach
A method for reaching a verdict
One method covers the three-verdict formats. Comprehension and analogy items need a variant, noted in the last step.
- 1
Read the statement first. You are looking for one specific thing in the passage. Knowing what it is turns reading into searching.
- 2
Find the sentence that decides it. Almost every true or false verdict rests on one sentence. Locate it and keep your finger on it before deciding.
- 3
Ask what would have to be added. If you needed to supply anything at all — a definition, a likelihood, a piece of general knowledge — the verdict is cannot say.
- 4
Separate contradicted from unmentioned. False means the passage says the opposite. Cannot say means the passage is silent. These are different findings, and providers score them differently.
- 5
For analogies, name the relationship before reading the options. State it as a sentence — a fork is the tool used for eating — then reject every option that does not preserve it in the same direction.
What slows progress
Common mistakes, and the fix for each
- Importing what you already know. Fix: after each answer, name the sentence that decided it. If you cannot point at one, you inferred rather than read.
- Treating probably true as true. Fix: apply one test. Could this statement be false while every sentence in the passage stays true? If it could, the verdict is cannot say.
- Confusing cannot say with false. Fix: false needs a contradiction you can quote. Cannot say needs an absence you can confirm. Check which one you actually have before marking.
- Reading the passage end to end first. Fix: read the statement first and search. Full reading is the right move only on comprehension items that ask for the main point.
- Practising the wrong response format. Fix: confirm whether your assessment uses three verdicts, four-option comprehension or analogies before you build the habit. They reward different reading.
You have seen the method. Employers set the full battery.Every provider, every track, a worked solution on every question.
Get full accessIn practice
How do you tell false from cannot say?
Apply two different tests. False needs a contradiction you can quote: a sentence in the passage that asserts the opposite of the statement, the way sample question 3's passage gives Helsinki a 4.00pm Friday close while the statement claims standard hours. Cannot say needs an absence you can confirm: the passage never addresses the point, the way sample question 2's policy adds an approval step but never mentions how long approval takes. One verdict rests on a sentence you can point at, the other on a search that comes back empty.
Most errors run in a predictable direction. Statements that sound likely get marked true and statements that sound unlikely get marked false, when the passage is actually silent either way. The repair is one question: could this statement be false while every sentence in the passage stays true? If it could, nothing in the passage forces it, and the verdict is cannot say however plausible the statement feels.
The opposite failure exists too. Candidates burned by importing outside knowledge start treating cannot say as the safe default, and three-verdict tests include directly contradicted statements precisely to catch that. False and cannot say are different findings — the passage says the opposite versus the passage says nothing — and providers score them as different findings, so check which one you actually hold before marking either.
In practice
How do you find the sentence that decides a verdict?
Read the statement before the passage, and let it tell you what to search for. A statement names its own subject — a city, a deadline, a payment method — and scanning for that word is faster than reading every sentence in order. On multi-sheet formats like Aon's scales verbal, where 49 tasks share 12 minutes, knowing which data sheet to open is most of the time saved.
When you reach the candidate sentence, slow down for the small words. All, only, never, unless, except, may and must can each flip a verdict on their own, and exceptions are where test writers put the answer: sample question 3 turns entirely on the sentence that names one. Keep your finger on the deciding sentence, decide once, and move — re-reading a passage to confirm a verdict you have already reached costs more than it recovers.
The same search discipline wins on detail questions, where the traps are made from the passage itself. Every wrong option in sample question 6 is a real timing attached to the wrong object, which is why matching a familiar phrase is not the same as finding the sentence. Point at the clause that answers the question, check what it actually modifies, then commit.
In practice
When is a conclusion actually supported by the passage?
A conclusion is supported when the passage forces it, not when the world makes it sensible. In sample question 4, the record shows that out-of-hours calls fell after a release day moved: that the calls fell is entailed; that the move caused the fall is merely invited. Tests in this family reward the modest reading, and the strongest-sounding conclusion on the list is usually the one that quietly went beyond the text.
Arguments have the same structure in reverse: they lean on assumptions the text never states. The reliable test is negation, as in sample question 5 — take the candidate assumption, suppose its opposite, and see whether the argument still stands. If the argument survives, it never relied on that assumption; if it collapses, you have found the load-bearing gap.
This is the boundary where verbal reasoning meets critical thinking. A Watson-Glaser-style assessment splits exactly these judgements — inference, assumptions, deduction, interpretation and argument evaluation — into separate sub-tests with their own decision rules, so if your invitation names one, the critical-thinking guide picks up where these two questions leave off. The underlying discipline transfers whole: being right about the world is never the task, and what the text forces is the only thing being marked.
A clear route
How to prepare in the days you have
Accuracy first, then format, then pace. Adding speed before the verdict rules are secure only makes the same errors arrive sooner.
- 1
First, learn the verdict grammar. Work untimed on true, false and cannot say, and for every answer write down the sentence that decided it. If you cannot point to a sentence, the verdict came from somewhere the passage did not supply.
- 2
Next, attack your cannot-say errors. Sort your mistakes into contradicted, unmentioned and assumed. The third pile is the one to work on, because it is where outside knowledge crept in without being noticed.
- 3
Then switch to your actual format. Move to the passage length, response set and navigation named in your invitation. An analogy test and a data-sheet test need genuinely different practice.
- 4
Time it. Short sets at the real pace, then one full rehearsal. Track how long a verdict takes you, not only whether it was right.
- 5
On the last day, re-solve misses only. Return to the items you got wrong and nothing else. Reading new passages the night before rarely helps.
Cognivy uses your assessment date to choose the route rather than asking you to predict a study schedule. When you sit down to practise, you choose the session length that fits that day.
Pace
Getting faster without losing accuracy
Verbal speed comes from searching instead of reading, and from deciding without re-reading.
- Search, do not read. Statement first, then scan for the noun or figure it names. On data-sheet formats like Aon's, knowing which sheet to open is most of the time saved.
- Decide once. Re-reading a passage to confirm a verdict you have already reached costs more than it recovers. Once the deciding sentence is found, commit.
- Learn the qualifier words. All, only, never, unless, except, may and must can each flip a verdict on their own. Reading them deliberately is faster than reading the sentence twice.
- Do not chase every task. Where the provider says completion is not expected, pace to accuracy. Forty-nine tasks in twelve minutes is a design choice, not a target.
Direct answers
Verbal reasoning test FAQs
It means the passage does not contain enough information to decide the statement either way. It is not a hedge, and it is not the same as false. False requires the passage to contradict the statement; cannot say requires the passage to be silent on it.
No. SHL Direct and Aon both describe that three-option format, but Pearson's DAT Verbal Analogies uses word relationships, the CCAT's verbal category uses vocabulary and analogies, and many comprehension formats use four or five options with no cannot-say option at all.
It depends on the assessment. Aon lists scales verbal at 12 minutes for 49 tasks. Korn Ferry lists Elements Verbal at 16 minutes. Criteria's CCAT mixes verbal into 50 questions in 15 minutes. Follow the time stated in your invitation.
For passage-based formats, rarely, because the answer is in the text. For analogy and word-relationship formats such as the DAT and the CCAT verbal category, vocabulary genuinely helps, though naming the relationship still matters more than knowing every word.
There is no universal pass mark. Results are usually reported as a percentile against a comparison group, and Korn Ferry states that it does not set pass marks. Any benchmark comes from the employer.
They overlap but are scored differently. A verbal reasoning test usually asks what a passage supports. A critical-thinking assessment such as Watson-Glaser splits that judgement into five separate sub-tests, each with its own decision rule.
No. If reaching your answer required anything the passage does not state, the answer is wrong even when it is true in the real world.
No. Cognivy is independent and is not affiliated with or endorsed by any assessment provider. All practice material is original and written to teach the verdict rules, formats and pacing of a named style.