Reviewed 19 August 2026 by the Cognivy editorial team
This page is deliberately an orientation page. It explains what employers mean by logical reasoning, sets out the families you might be sent, works through an example from two of those families, and adds six free practice questions — four of them on analysing arguments — before handing you to the specific Cognivy guide that matches your assessment.
- Questions
- 10 to 50, depending on provider
- Time limit
- 6 to 15 minutes, depending on provider; Aon's AP Reasoning is instead timed per question, at 2 minutes 30 seconds each
- Answer format
- Multiple choice; Korn Ferry / Talent Q-style questions can use unusually large answer sets
- Material
- Verbal, numerical, symbolic or diagrammatic — the one label covers several distinct families
- Scoring
- Percentile against a comparison group. Providers do not publish pass marks
The skill
What is a logical reasoning test?
Logical reasoning describes any test where a consistent relationship has to be found and then applied. The material can be verbal, numerical, symbolic or visual, and the label sits on top of several genuinely different tasks.
Providers use the name in their own ways, and the label can point away from the task. Korn Ferry's Elements Logical analyses abstract information to identify the rule behind a pattern. Aon publishes AP Reasoning as its Logical Reasoning assessment, and its practice tasks are must-be-true premises and an ordering problem, which is deductive work. Criteria's CCAT reports a combined math and logic category. Read the published task, not the product name.
Because the name is shared and the tasks are not, preparing for logical reasoning in general is inefficient. Preparing for the family you were actually sent is not.
Searches for logical reasoning also return LSAT Logical Reasoning, a section of the United States law-school admission test with its own question types and its own preparation. Everything on this page is about employer aptitude assessment for job candidates.
Try it before anything else
Free logical reasoning practice questions
Six original questions: one deductive, one inductive, then four on argument analysis — finding the main conclusion, naming the flaw, and judging what strengthens or weakens a case. Commit to an answer before opening the solution — the solution shows the method and names why every wrong option is on the list.
Four tasks — J, K, L and M — must be scheduled across four consecutive days, one task per day.
K is scheduled at some point before L · M is scheduled on the last day · J is not scheduled on the first day
Which task must be scheduled on the first day?
Show the worked solution
Answer: B — K.
- M is fixed on day four, so days one to three hold J, K and L in some order.
- J cannot be on day one, which leaves K or L for that slot.
- L cannot be first, because K has to come before it and there is no earlier day available.
- Day one must therefore be K. J and L fill days two and three in either order, which is why the question asks only about the first day.
Why the other options are there. Option D is tempting because two of the four positions genuinely are undetermined. A question about one position does not require the whole arrangement to be fixed, and treating partial information as no information loses marks throughout deductive tests.
A number sequence of the kind used inside mixed cognitive ability tests.
4, 9, 19, 39, ?
Which number continues the sequence?
Show the worked solution
Answer: C — 79.
- The differences are 5, 10 and 20 — each one double the last, which points at a multiplicative rule rather than an additive one.
- Test it against every term: 4 × 2 + 1 = 9, then 9 × 2 + 1 = 19, then 19 × 2 + 1 = 39.
- The rule holds for all four given terms, so apply it once more: 39 × 2 + 1 = 79.
- Checking against the whole sequence rather than the last pair is what makes the answer safe.
Why the other options are there. Option A comes from adding 20 again — continuing the last difference instead of the pattern in the differences. It is the answer you reach by looking at one step rather than the whole sequence.
The first of four questions on analysing an argument rather than completing a pattern. An operations director circulates this case for a shift change.
“Northfield Logistics should move its drivers to four-day shifts. Accident rates at the firm rise sharply in the final two hours of a five-day week, and the three competitors that adopted four-day patterns last year all reported fewer insurance claims within six months. Fuel costs would fall slightly as well, since routes could be consolidated. Some managers worry about delivery backlogs, but scheduling software has handled larger changes before.”
Which option best states the main conclusion of the argument?
Show the worked solution
Answer: B — Northfield Logistics should move its drivers to four-day shifts.
- The main conclusion is the claim the rest of the passage exists to support, so test each sentence with because: the shift change is argued for because of the accident pattern, because of the competitors' results and because of the fuel saving.
- Run the link the other way and it fails — the accident figures are not true because the firm should change its shifts — which confirms the direction of support.
- The final sentence is a rebuttal. It exists to clear an objection out of the recommendation's way, which is further evidence that the recommendation is the conclusion.
- Position proves nothing. Here the conclusion happens to come first; in a real test it sits as often in the middle of the passage, or at the end, or across two sentences.
Why the other options are there. Option A is a premise — it supports the conclusion and is supported by nothing. Option C turns the evidence about three named competitors into a claim about firms in general, which the passage never makes and which would in any case be support rather than the point. Option D overstates the rebuttal: the passage says scheduling software has handled larger changes, not that it will prevent backlogs.
A head of marketing presents this reasoning at a quarterly review.
“Since the marketing team introduced its new brand guidelines in March, website enquiries have risen by a fifth. The guidelines must therefore be responsible for the increase. The design agency that produced them should be retained for the packaging refresh, because whatever lifted enquiries once will lift sales in the same way.”
Which option identifies the most serious flaw in this argument?
Show the worked solution
Answer: C — it assumes that what followed the guidelines was caused by them.
- Find the conclusion first: the guidelines caused the increase, and on the strength of that the agency should be retained.
- Ask what the evidence actually establishes. The only support offered is sequence — enquiries rose after the guidelines appeared — and sequence cannot carry a causal claim while seasonal demand, a concurrent campaign or market growth go unexamined.
- The argument then stretches the assumed cause further, transferring an effect on enquiries to a different outcome, sales, in a different setting, packaging.
- Option C names the gap between evidence and conclusion, which is what a flaw question asks for: the reason the conclusion does not follow.
Why the other options are there. Option A fails because the one-fifth rise is presented as a measurement, not an opinion. Option B denies a premise the passage plainly supplies — a flaw is a reason the conclusion does not follow from the premises, never a premise you pretend is absent. Option D describes no contradiction: expecting the guidelines to lift sales as well as enquiries is unsupported, but the two claims are entirely consistent with each other.
The partners of a professional-services firm are debating a retention plan.
“Attrition among first-year analysts at Harlow & Wade is double the industry average. The firm's exit interviews cite workload as the main reason for leaving. The partners therefore plan to cap analyst hours at 55 a week, expecting first-year attrition to fall towards the industry average within two years.”
Which option, if true, most strengthens the partners' expectation?
Show the worked solution
Answer: D — a rival's identical cap was followed by attrition falling to the industry average.
- Map the chain the expectation rests on: workload is why analysts leave, so capping hours should cut attrition.
- A strengthener must make that chain more likely to hold. Read each option asking one thing — does this bear on the link between the cap and attrition, or does it merely say something agreeable about the firm?
- Option D is the same intervention followed by the predicted outcome at a comparable firm: direct evidence that the mechanism works, and the only option that touches the chain at all.
- Check it is not already assumed. The argument claims no precedent anywhere, so option D adds new support rather than restating a premise.
Why the other options are there. Option A is about pay, a lever the argument never uses, so it neither supports nor undermines the workload chain. Option B reports the people who stayed and says nothing about why leavers leave — it agrees with the mood of the plan without strengthening its logic. Option C weakens rather than strengthens: a leaver interviewed by the partners has a reason to give a safe, impersonal answer, which makes the workload diagnosis less reliable.
A supermarket group's retail director asks the board to approve a chain-wide rollout.
“Ridgeway Foods trialled self-checkout tills in ten of its 200 stores last quarter, and reported theft in those ten stores fell by 12%. Installing self-checkout across the whole chain will therefore cut reported theft company-wide, and the board should approve the rollout at Friday's meeting.”
Which option, if true, most weakens the director's argument?
Show the worked solution
Answer: B — the trial stores were also the only ones fitted with upgraded cameras that quarter.
- The argument moves from one observed change in ten stores to a company-wide prediction. Its weak point is anything else that could explain the fall in those stores.
- Option B supplies exactly that: a second change, made in the same stores over the same quarter, capable of accounting for the whole reduction.
- If the cameras did the work, rolling out self-checkout alone gives no reason to expect the fall to repeat, and the inference from trial to chain collapses.
- A weakener does not have to prove the conclusion false. It only has to remove the support the premises were lending it, and an unexamined alternative cause does precisely that.
Why the other options are there. Option A attacks the plan on cost, which the conclusion never mentions — an objection to a different argument. Option C reports what rivals have done with no outcome attached, so it neither supports nor undermines the theft claim. Option D is a different measure again: unchanged customer satisfaction is compatible with theft falling for any reason at all.
A short logical 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 logical reasoning tests measure?
Across the whole family, employers are measuring one core thing: whether you can extract a rule from limited information and apply it without carrying assumptions into it.
What differs is the surface. A deductive item tests whether you can refuse a conclusion that is merely possible. An inductive or abstract item tests whether you can form a hypothesis from examples and check it. A mixed test such as the CCAT adds one more demand: switching between them quickly, without letting the previous rule leak into the next question.
Reading your result
How logical reasoning tests are scored
A logical reasoning result is rarely a count of correct answers. Providers convert it into a comparison against other people, and two features of this family — adaptive delivery and mixed formats — change what the number means.
- Results are percentiles, not marks out of ten. Results across this family are reported as a percentile against a comparison group. Criteria, for example, sends the employer a raw score and a percentile for the CCAT, and it is the comparison, not the raw count, that gets benchmarked.
- No provider publishes a pass mark. Korn Ferry states directly that it does not set them, and the benchmark belongs to the hiring organisation. If a recruiter has disclosed a target, that is the only target worth working to.
- On an adaptive test, difficulty is not feedback. Korn Ferry states that all of its cognitive ability tests are adaptive, and Aon's AP Reasoning selects its 10 questions adaptively. Question difficulty moves with your answers, so a harder question often follows a correct one, and the number of questions completed does not by itself explain the result.
- A mixed-test score hides the expensive subtype. On the CCAT, logic questions are one of three categories inside a single result, so the overall score cannot tell you whether logic, verbal or spatial questions cost the marks. Tracking accuracy by subtype in practice is what recovers that information.
- Unanswered questions are part of the design. Aon tells candidates across its range that they are not expected to answer every task, and Criteria reports that fewer than 1% of candidates answer all 50 CCAT questions. Pacing to accuracy is the strategy, not a compromise.
Question formats
The families a logical reasoning test can turn out to be
Work out which of these your items look like. Each family has its own Cognivy guide, linked at the end of this page.
- Deductive reasoning. Rules or premises given, and one conclusion that must follow. Syllogisms, ordering and scheduling constraints, and conditional if-then logic.
- Inductive reasoning. A rule inferred from examples and applied to an unseen case, usually as a next-in-series or odd-one-out figure. SHL and Aon both sell products under this name.
- Abstract reasoning. The broad non-verbal family: sequences, analogies, classifications and sets of shapes with no verbal or numerical content.
- Matrix reasoning. A grid with one cell missing, solved from the row and column rules together. Matrigma, Raven's Adaptive and Aon's gapChallenge all use this shape.
- Diagrammatic reasoning. Symbols and operators applied in order to transform an input into an output, or to recover an input from an output.
- Mixed cognitive ability tests. One test containing several of the above plus verbal and numerical items, scored as a single result. Criteria's CCAT is the clearest example in Cognivy's launch scope.
An invitation reading logical reasoning test, 15 minutes could be Korn Ferry's Elements Logical, which is adaptive and abstract; it could be an Aon module asking for a logical conclusion from a set of items; or it could be the CCAT, which is 50 mixed questions with no calculator. Same three words, three different preparations. Find a product name before you start.
Launch coverage
Which providers name a logical reasoning test?
Three of Cognivy's launch families name a logical assessment directly. Others deliver the same underlying skills under a different label.
- Talent Q Elements Logical. Korn Ferry lists Elements Logical at 15 minutes and states that all of its cognitive ability tests are adaptive, selecting each question from your previous answer. The task is abstract: identify the rule behind a pattern.
- AP Reasoning. Aon publishes AP Reasoning as its Logical Reasoning assessment and describes it as measuring the ability to apply given rules to logical problems in one or more steps, over 10 adaptively selected questions with 2 minutes 30 seconds each. Its published practice tasks are a must-be-true syllogism and an ordering problem, so practise it on the deductive reasoning guide.
- CCAT. Logic questions are one of three categories inside a single test: 50 questions in 15 minutes, no calculator, interleaved with verbal and spatial items. Criteria reports that fewer than 1% of candidates answer all 50.
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 and scoring across the family
Because the family is broad, so is the timing. These are the published figures for the assessments that carry the logical label.
- Korn Ferry Elements Logical: 15 minutes, adaptive. Each question carries its own time limit, and the next item is selected from your last answer. An adaptive test may follow a correct answer with a harder item, but you cannot read your score from how hard the questions feel, so treat rising difficulty as normal behaviour rather than as feedback.
- Criteria CCAT: 50 questions in 15 minutes, no calculator. That is roughly 18 seconds a question on average, with logic items interleaved among verbal and spatial ones.
- Aon modules are short by design. Aon lists AP Reasoning at 10 questions with 2 minutes 30 seconds each, scales clx at 6 minutes and switchChallenge at 6 minutes, and tells candidates across its range that they are not expected to answer every task.
- No provider publishes a pass mark. Korn Ferry states directly that it does not set them. Criteria reports a raw score and a percentile to the employer, who chooses the benchmark for the role.
A repeatable approach
A method that works across the family
Whatever the surface, the same five moves apply. The first one does most of the work.
- 1
Identify the subtype before choosing a method. Deductive, inductive, abstract, diagrammatic or mixed. Each subtype has its own acceptance rule, so a method chosen for the wrong subtype produces a confident answer to a different question.
- 2
State the rule in words. Whether it is a premise, a pattern or an operator, write it as a short sentence before you use it. A rule you cannot state is a rule you cannot check.
- 3
Separate must from could. On deductive items, reject anything merely consistent with the premises. On inductive items, prefer the simplest rule that covers every example rather than the most interesting one.
- 4
Eliminate rather than derive. Where a full solution is slow, test the options against the rule instead. On a mixed timed test this is often the only viable route.
- 5
Reset between questions. On a mixed test the rule changes every item. Deliberately dropping the previous rule stops it being carried into a question that does not use it, which is exactly what a format like the CCAT interleaves its items to catch.
What slows progress
Common mistakes, and the fix for each
- Using one favourite method everywhere. Fix: name the subtype first. Pattern-hunting on a deductive item and rule-hunting on a syllogism both fail in the same way.
- Carrying a rule between questions. Fix: on mixed tests, take a deliberate half-second reset. The rule that just worked is almost certainly not the next rule.
- Deriving when eliminating is faster. Fix: at roughly 18 seconds a question on a CCAT-style test, four options are a finite set you can test against the rule, while solving from first principles has no fixed end point.
- Treating possible as necessary. Fix: on deductive items, ask whether the conclusion could be false while every premise stays true. If it could, it does not follow.
- Reading only the total score. Fix: track accuracy by subtype in practice. A single mixed-test score hides the one category that is actually costing you the marks.
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 find the main conclusion of an argument fast?
Look for the claim the rest of the passage exists to support, not the last sentence. Test writers routinely put the conclusion first or bury it mid-passage, then close on a supporting detail or a concession, so position is the least reliable signal there is.
The dependable check is the because test. Take a candidate sentence and ask whether the other sentences exist to answer why you should believe it. In sample question 3 above, the recommendation to move to four-day shifts is argued for because of the accident figures and because of the competitors' results — read it with because between the claims and the structure sorts itself out. Run the link the other way and it turns to nonsense, which is how you know the direction of support.
Indicator words help, but only secondarily: therefore, so and which is why point at conclusions, while since, because and given that mark support. When a passage contains two conclusion-shaped sentences, the main conclusion is the one the other also supports — an intermediate conclusion passes the because test in one direction only.
In practice
What makes an option strengthen an argument rather than merely agree with it?
A strengthener has to make the conclusion more likely to be true given the premises — it works on the link between evidence and conclusion, not on the general mood of the passage. An option can be consistent with the argument, flattering to it, even true, and still strengthen nothing, because it leaves the inference exactly as exposed as it was.
Sample question 5 turns on the distinction. That most analysts would recommend the firm agrees with the spirit of the plan but never touches the chain the partners rely on — workload drives leaving, so a cap will cut attrition. The rival firm's result does touch it: the same intervention was followed by the predicted outcome, which is evidence the mechanism works.
So test each option against the argument's gap, not its topic. Ask what the argument needs to be true that it has not shown — a representative sample, a working causal mechanism, the absence of an alternative explanation — and choose the option that supplies it. The same discipline in reverse finds weakeners: sample question 6 turns on an alternative cause the trial never ruled out.
In practice
How do you spot the flaw in an argument under time pressure?
Learn the shortlist rather than starting from nothing each time. Most flawed arguments in employer tests fail in a small number of repeatable ways: treating sequence as cause, generalising from an unrepresentative sample, assuming what worked in one setting transfers to another, attacking the person rather than the claim, and confusing a necessary condition with a sufficient one.
Then read the conclusion first and ask what would have to be true for the premises to establish it. The gap between the two is the flaw, and the correct option is the one that names that gap. Sample question 4 measures a rise in enquiries after March and concludes the guidelines caused it — the missing step, ruling out every other explanation for the rise, is exactly what the right option points at.
Be wary of options that criticise something true of the argument but not wrong with it. An argument can rest on one trial, use round numbers or predict the future without any of those being its flaw. The flaw is the reason the conclusion does not follow from the premises, and an option that would leave the conclusion just as believable if repaired is not naming it.
A clear route
How to prepare in the days you have
The most valuable thing you can do first is stop preparing for logical reasoning in general.
- 1
Find the product name. Check your invitation, the landing page and the instructions shown before the timed section. Elements Logical, an Aon scales module and the CCAT need different practice.
- 2
Open the guide that matches it. Deductive, inductive, abstract, matrix or diagrammatic. Each has its own method and its own common traps, and each is linked at the end of this page.
- 3
If it is genuinely mixed, practise switching. For a CCAT-style test, practise mixed sets rather than one subtype at a time, because changing method from one item to the next is part of the task.
- 4
Build pace with elimination. Practise reaching an answer through the options rather than from first principles, and time yourself doing it.
- 5
Track accuracy by subtype. One overall score will not tell you where to spend the last day. A per-subtype breakdown will.
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
Speed in this family is mostly about not wasting time on the wrong method.
- Classify in the first three seconds. A technique applied well to the wrong item type still produces the wrong answer, so name the type before you choose a method.
- Use the options early. On both deductive and pattern items, one violated feature is enough to reject an option, while confirming one requires the whole rule.
- Write the rule down. A three-word note keeps a constraint available while you work on the next one, which is what ordering and scheduling items demand when several constraints interact.
- Accept unanswered questions. On a test where fewer than 1% of candidates finish, pacing to accuracy is the strategy rather than a compromise.
Direct answers
Logical reasoning test FAQs
A test where you find a consistent relationship in the information given and apply it. It is a family name rather than one assessment: deductive, inductive, abstract, matrix and diagrammatic tests are all sold under it, and some providers use it for a mixed cognitive ability test.
A deductive item gives you rules and asks what must follow, so anything merely possible is wrong. An inductive item gives you examples and asks you to infer the rule that best explains them, then apply it to an unseen case. Deduction moves from rule to case; induction moves from cases to rule.
Not quite. Abstract reasoning is one of the families that sits under the logical label, and it specifically uses non-verbal shape material. A logical reasoning test can equally be verbal, numerical or symbolic.
It depends on the product. Korn Ferry lists Elements Logical at 15 minutes with adaptive delivery. Criteria's CCAT is 50 mixed questions in 15 minutes with no calculator. Aon's modules are shorter still, with AP Reasoning listed at 10 questions of 2 minutes 30 seconds each.
No. LSAT Logical Reasoning is a section of a United States law-school admission test with its own question types and scoring. This guide covers employer aptitude assessment, which is a different set of tests and a different preparation.
There is no universal pass mark. Korn Ferry states that it does not set them, and Criteria reports a raw score and percentile that the employer benchmarks. Ask the recruiter whether a target has been disclosed.
Look for a product name rather than a skill name: Elements Logical, an Aon scales or challenge module, or the CCAT. The product name tells you the item type, the timing and the response format, none of which the phrase logical reasoning test tells you on its own.
No. Cognivy is independent and is not affiliated with or endorsed by any assessment provider. Every practice item is original material written to teach the relevant method and pacing.