Deductive reasoning tests: what must follow

The skill

What is a deductive reasoning test?

A deductive reasoning test supplies a small, closed set of facts — premises, rules, constraints or a short scenario — and asks which conclusion necessarily follows from them. The information you are given is complete for the purpose of the question. Nothing outside it counts, including things you happen to know are true in the real world.

The defining move is refusal. Most wrong options on a deductive item are not false statements; they are statements that could be true but are not forced. A conclusion earns a tick only when there is no arrangement of the facts that makes the premises true and the conclusion false.

SHL describes its Verify Deductive Reasoning assessment as requiring candidates to evaluate arguments, analyse scenarios and draw logical conclusions, and as measuring the ability to complete scenarios using incomplete information. Aon's equivalent, which it publishes as AP Reasoning, is described as measuring the ability to apply given rules to logical problems in one or more steps.

Try it before anything else

Free deductive reasoning practice questions

Six original questions covering the shapes above, from a two-rule chain to a full allocation. Commit to an answer before opening the solution — the solution writes the model out and names why every wrong option is on the list.

What employers are looking at

What do deductive reasoning tests measure?

Employers use deductive items to predict how someone behaves when a decision has to rest on stated terms: a policy, a contract clause, an eligibility rule, a set of technical constraints. The question is whether you can work inside the boundary of what was actually specified rather than filling gaps with assumption.

Three abilities separate here. Rule handling: can you read a conditional in the direction it was written rather than its reverse. Combination: can you hold four or five constraints together and see what they jointly force. Restraint: can you reject an attractive conclusion because nothing supports it, under a clock that rewards decisiveness.

Reading your result

How deductive reasoning tests are scored

What the employer sees is rarely a raw count of correct answers. Providers convert your performance into a comparison against other candidates, and on the adaptive assessments in this family the difficulty of your questions carries information the count alone does not.

• No provider publishes a pass mark. Korn Ferry states plainly that it sets no pass or fail marks for its psychometric assessments and that employers choose their own benchmarks, and the same principle holds across this family. If a recruiter has disclosed a target, that is the only target worth working to.
• Results are read against a comparison group. A deductive result is normally reported against a comparison group rather than as a standalone mark. The same performance reads differently against different groups, which is why a result from one assessment says little about how you would stand in another.
• On an adaptive test, the raw count is not the score. SHL's Verify Deductive Reasoning uses computer adaptive technology, and Aon selects each AP Reasoning question from your response to the previous one. Question difficulty moves with your answers, so the number you answer does not by itself explain the result — and rising difficulty is normal behaviour rather than feedback.
• A Watson-Glaser deduction result sits inside a wider score. Deduction is one of five sections in the Watson-Glaser Critical Thinking Appraisal, alongside drawing inferences, recognising assumptions, interpreting and evaluating arguments. Performance on deduction lands inside a critical-thinking result, which is why an invitation naming Watson-Glaser is a reason to prepare all five sections rather than one.
• An unproctored sitting may be verified later. SHL designs Verify Deductive Reasoning for an unproctored environment with one sitting plus a verification test, so a result can be confirmed under supervision afterwards. The only preparation that survives that design is genuinely getting better at the reasoning.

Question formats

What do the questions look like?

Deductive assessments use a small number of recurring shapes. Recognising which one you are looking at decides the method before you read the options.

• Syllogisms and must-be-true conclusions. Two or three premises about categories — all, some, no, only — followed by candidate conclusions. Aon's published AP Reasoning practice tasks include exactly this shape, asking which statement must be true.
• Ordering and scheduling constraints. A set of items to be arranged in a sequence, on days, or in positions, with relational rules such as before, after and immediately after. Aon's practice material includes an ordering task of this kind.
• Conditional if-then rules. One or more rules of the form if A then B, chained together. The reliable question is which direction the rule runs, because if A then B tells you nothing about a case that lacks A.
• Grouping and allocation. Items assigned to teams, rooms or shifts under rules about which can and cannot share. Usually solved by placing the most restrictive rule first.
• Contradiction and impossibility. Rather than asking what follows, the item asks which arrangement cannot happen, or which pair of statements cannot both be true.
• Rule application in a code or grid. Some providers deliver deduction visually. Aon's gapChallenge presents a four-by-four or five-by-five grid where no shape repeats in any row or column, and asks which shape belongs in the marked square.

Launch coverage

Which providers use deductive reasoning tests?

Three of Cognivy's launch families assess deduction under a name you will see in an invitation. The tasks differ more than the names suggest.

• Verify Deductive Reasoning. SHL's fact sheet lists 20 minutes and a maximum of 18 questions, in a multiple-choice adaptive format designed for an unproctored setting, with one sitting plus a verification test. SHL states the assessment measures the ability to draw logical conclusions from information provided and to identify the strengths and weaknesses of arguments.
• AP Reasoning. Aon publishes AP Reasoning as its Logical Reasoning assessment, but the published practice tasks are deductive: premises followed by candidate conclusions, and an ordering problem with relational constraints. Aon lists 10 questions selected adaptively from your previous answer, with 2 minutes and 30 seconds allowed per question.
• Watson-Glaser Deduction. Deduction is one of five sections inside the Watson-Glaser Critical Thinking Appraisal, which Pearson describes as determining whether conclusions follow logically from the given information. Pearson lists the full appraisal at 40 items and 30 to 40 minutes timed, so the deduction section is one part of a longer assessment.

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, adaptivity and scoring

There is no single deductive reasoning test, so there is no single time limit or question count. These are the figures the providers publish for their own assessments.

• SHL Verify Deductive Reasoning: 20 minutes, up to 18 questions. SHL lists the question format as multiple choice and adaptive, states that the assessment uses computer adaptive technology, and designs it for an unproctored environment with a follow-up verification test.
• Aon AP Reasoning: 10 questions, 2 minutes 30 seconds each. Aon states that the questions are selected in an adaptive fashion according to your response on the previous question, and that the difficulty level may increase as you progress.
• Watson-Glaser deduction sits inside a longer appraisal. Pearson lists the Watson-Glaser Critical Thinking Appraisal at 40 items and 30 to 40 minutes timed across all five sections. It does not publish a separate limit for the deduction section alone.
• A per-question limit changes the right strategy. Where time is allocated per item rather than to the paper as a whole, banking time on easy questions does not help you on hard ones. Aon's own guidance for AP Reasoning is to work quickly and accurately within the limit for each question.

A repeatable approach

A method that survives the clock

Deduction rewards a written model over a mental one, because a model on paper stays fixed while you test each option against it. Five moves.

1. 1

   Write the rules in short notation. Turn each premise into a symbol before you read the options: U → S for every urgent file is supervised, L < M for L comes before M. Notation makes a contradiction visible instead of something you have to remember.

2. 2

   Mark the direction of every conditional. If A then B is not if B then A, and it is not if not A then not B. Draw the arrow. A reversed conditional is invisible in prose, because the reversed sentence reads as sensibly as the original one.

3. 3

   Place the most restrictive rule first. On an ordering or grouping item, the rule that fixes a position — must be last, immediately follows, cannot share — collapses the possibilities fastest. Rules that only forbid a pairing come after it.

4. 4

   Test each option against the model, not against your intuition. Ask one question per option: is there any arrangement consistent with all the rules in which this is false? If yes, the option is not forced, however reasonable it sounds.

5. 5

   Stop at the first option that must hold. On a must-be-true item exactly one option is forced. Once you have proved one, checking the rest is spending time you will want on the next question.

What slows progress

Common mistakes, and the fix for each

• Reversing a conditional. Fix: write every if-then as an arrow and read only in the direction of the arrow. If A then B tells you what happens when A is present and nothing at all about a case where A is absent.
• Treating possible as necessary. Fix: for each option, try to construct one arrangement that satisfies every rule and makes the option false. If you can build it, the option is not the answer, no matter how sensible it reads.
• Importing world knowledge. Fix: the premises are the universe. If an item says all analysts work Saturdays, that is true inside the item. Deductive tests are deliberately built with premises that would be odd in real life, because that is where imported assumptions show up.
• Holding the constraints in your head. Fix: write the model. Four constraints exceed what most people can juggle accurately while reading four options, and the cost of a scribbled line is a few seconds against a lost mark.
• Reading some as some but not all. Fix: in formal terms some means at least one, and it leaves open that the answer is all. A conclusion of the form some X are not Y does not follow from some X are Y.

In practice

How do you work through a syllogism without a diagram?

Turn each premise into an arrow before you look at the options. All A are B becomes A → B. No A is B becomes A → not B. Some A are B is different — it is not an arrow but a fact that at least one thing sits in both categories, so note it separately. Sample question 3 above reduces to exactly two arrows and one standalone fact, and the answer falls out of chaining the arrows head to tail.

Then guard against the two moves that break syllogisms. An arrow reads in one direction only: all Leeds auditors hold the certificate does not put every certificate holder in Leeds, which is sample question 3's option A. And some means at least one, leaving open that the answer is all — a conclusion of the form some are not does not follow from some are, which is option B in the same question.

A diagram is never wrong, but under a clock the arrows are faster and they survive checking. If a conclusion needs an arrow read backwards, or a fact no premise supplies, it is not forced, however natural it sounds read aloud.

In practice

What does “must be true” actually exclude?

Everything that is merely consistent with the rules. Most wrong options on a deductive question are not false statements — they are statements the rules happily permit and never force. The working test is a counterexample: one arrangement that satisfies every rule and makes the option false is enough to discard it. In sample question 4, three of the four options hold in one valid seating and fail in the other, which is exactly the trap of proving an option possible and calling it necessary.

It also excludes anything imported from outside the premises. Real-world plausibility, a timescale the rules never mention, a count nobody supplied — sample question 1's wrong options add a two-day deadline and a second officer that appear nowhere in the rules. The premises are the whole universe of the question, and a conclusion that reaches beyond them fails, however sensible it would be in real life.

Note what the phrase does not exclude: a conclusion that feels too obvious. Candidates sometimes skip the direct chain because it seems too plain to be the point. On a must-be-true question the plain, forced conclusion is the point — and once one option is proved forced, the method says stop and move on.

In practice

How do you start an ordering or seating question?

Place the fixed rules first and the most restrictive relational rule second. A rule that pins a position — sits in chair 2, completed last — shrinks the problem immediately, and a locked adjacent pair like M immediately after J in sample question 2 collapses it further, because the pair moves as one block. Rules that merely forbid a pairing do their work last, as checks on whatever remains.

If the question asks about a single position, try elimination on that position before building any full layout. Sample question 4's opening chair falls in four short lines: R is fixed in chair 2, S is barred by rule, Q would need a chair to its left that does not exist, and P would push Q onto R's seat — leaving T and nobody else. No complete layout was needed to prove it.

When elimination does not settle the question, enumerate cases from the locked pair's possible positions and let each supposition fail against a named rule, the way sample question 2 walks J through second and third place before first survives. Write the cases down as you go — four constraints held mentally while reading four options is where ordering questions are lost.

A clear route

How to prepare in the days you have

Most candidates get a few days, not a study term. This order gets the most from a short run-up.

1. 1

   First, fix the quantifiers. Spend one untimed session on all, some, no and only, and on the difference between a conditional and its reverse. These are rules of meaning rather than habits of pace, so timed practice does not repair them.

2. 2

   Next, build the notation habit. Work untimed and write the model for every item, even the easy ones. The habit has to be automatic before it can survive a clock.

3. 3

   Then split by shape. Practise syllogisms and ordering problems in separate sets. They use different methods, and mixing them early hides which one you are actually weak at.

4. 4

   Add the provider's clock. If your invitation names SHL Verify Deductive Reasoning, rehearse against a whole-paper limit. If it names Aon AP Reasoning, rehearse against a per-question limit, because the pacing decisions are completely different.

5. 5

   On the last day, re-solve your misses. Go back only to items you got wrong and reconstruct the exact rule each wrong option broke. Naming the failure is what stops it repeating.

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 on a deductive test comes from deciding faster, not reading faster. Four things do most of the work.

• Read the question stem before the premises. Must be true, could be true and cannot be true are three different tasks with three different methods. Knowing which one you are doing before you read the rules saves a second pass.
• Eliminate on one counterexample. You do not need to prove an option false in general. One arrangement that satisfies the rules and breaks the option is enough to discard it, and it is usually quicker to find than a full proof.
• Keep one notation. Use the same symbols every time — arrows for conditionals, a left-to-right line for order, a slash for negation. A consistent layout lets you re-check your own working without redoing it.
• Know your pacing rule before you start. On a whole-paper limit, set an abandon point per item and honour it. On a per-question limit such as Aon's, there is nothing to bank, so use the full allowance rather than rushing an answer you cannot check.

Keep exploring

Related test types and how they differ

Deductive reasoning sits next to several assessments that share its vocabulary and reward different preparation.

Direct answers

Deductive reasoning test FAQs