Deductive Reasoning Test: Question Types, Worked Examples, Scoring

A deductive reasoning test checks one narrow but revealing skill: can you take a set of premises, stay strictly inside the information given, and identify the conclusion that must be true. No outside knowledge, no plausible guesses, one correct answer. Consulting recruiters care about this because it is the same discipline that separates a clean case answer from a confident-sounding one built on a weak assumption. This guide covers what the test actually measures, the question types you will see with fully worked examples, the real formats and scoring used by SHL and consulting firms, and how to turn the skill into something that transfers into your case interviews.

What does a deductive reasoning test actually measure?

For a consulting applicant, a deductive reasoning test is less about sounding smart and more about proving your thinking stays controlled when the problem is tight. You are given a small set of statements, rules, or constraints, and your job is to figure out what follows from them and what does not. That sounds simple, but it exposes a common weakness: people answer the question they expected rather than the one in front of them.

The defining property is validity. A conclusion is valid only if it must be true whenever the premises are true. "Probably true," "usually true," and "true in the real world" all fail. The test deliberately strips out business knowledge so that prior experience cannot rescue weak logic, which is exactly why all the information you need is contained in the prompt.

That distinguishes it from an inductive reasoning test, which asks what is probably true from a pattern, like predicting the next shape in a sequence. Deduction is top-down and certain. Induction is bottom-up and probabilistic. Most consulting assessments test both, so it is worth understanding the difference between inductive and deductive reasoning before you sit one.

What question types appear on the test?

SHL, the largest test publisher behind these assessments, groups its Verify deductive questions into three families, with a fourth (grouping) appearing in some firm versions. Here is each type with a worked example.

Syllogisms

You get two or more premises and must judge which conclusion necessarily follows. The trap is accepting a conclusion that is merely consistent with the premises rather than forced by them.

Premise 1: All footballers are fit and healthy. Premise 2: All famous sports players are footballers. Which conclusion follows?

Work the chain. Every famous sports player is a footballer (premise 2), and every footballer is fit (premise 1). So the set "famous sports players" sits entirely inside "footballers," which sits entirely inside "fit and healthy." The valid conclusion is "All famous sports players are fit and healthy." Note what does not follow: "All fit people are footballers" reverses the arrow and is invalid, even though it sounds related.

A second example shows why drawing helps. "Some berries are red. Some grapes are green. All grapes are berries." Which conclusion must be true? Because all grapes are berries and some grapes are green, those green grapes are necessarily berries, so "some berries are green" must be true. "Some berries are red" plus "some grapes are green" tells you nothing about red grapes, so any conclusion linking red and grapes is a trap. A quick Euler diagram (overlapping circles) makes the forced overlap obvious in seconds.

Arrangements: ordering, scheduling, ranking and seating

You are given subjects and a set of conditions, then asked what must be true about their order or position. These are the most time-consuming, so sketch a grid.

Six cousins have different ages. Nanno is 19. Conditions place the others above and below her. Question: which statement about the ranking is false?

The method is mechanical. Draw a vertical ladder, anchor the fixed fact (Nanno = 19), then apply each condition to eliminate impossible positions. Whichever statement contradicts a forced position is the false one. Ranking, scheduling (placing items across Monday to Friday), and seating (positions around a table) all use the same draw-then-eliminate approach. The error candidates make is holding the constraints in their head instead of on paper.

Grouping

You sort subjects into subgroups under conditional rules. Example: a farmer must place six creatures into three areas where "a bee colony cannot stay with anything" and "a goat must stay with the hen." You work out which placements are forced. The animal with restrictions that bind in every valid scenario is the one whose position you can pin down, while the rest float.

Underlying assumptions

You identify the unstated premise an argument depends on. If an argument concludes a mentoring program worked because pass rates rose, the underlying assumption is that nothing else explains the rise. Spotting the hidden assumption is the same move you make in a case when you say, "this recommendation depends on demand staying flat, so I would want to test that."

What are the real formats and scoring?

This is where candidates lose time by preparing for the wrong test. The format depends on who set it.

A few details matter for prep. The SHL Verify G+ non-interactive test runs 18 multiple-choice questions in 20 minutes, while the interactive version uses drag-and-drop scheduling and ranking for 12 questions in 18 minutes. BCG's Cognitive Test, introduced in 2024, is a different beast: 80 questions in 30 minutes, around 22 seconds each, taken live over Zoom under camera supervision, with logical reasoning (including syllogisms and basic deductive logic) making up roughly a third of it. Bain's SOVA assessment embeds deductive logic inside its logical reasoning and business case sections rather than as a standalone block, with difficulty comparable to LSAT logic questions. These firm screens are part of the wider consulting interview process, and the full landscape of online screens is mapped in the consulting aptitude test overview.

Scoring is the part most people misread. You are not graded on raw percent correct. Your score is converted to a percentile against a norm group of comparable candidates. SHL reports a grade band: A is the 90th to 100th percentile, B is 70th to 89th, C is 30th to 69th. Being in the top 20% of the norm group is a typical pass, and competitive employers often want the 85th percentile or higher. Practically, this means accuracy under time pressure matters more than getting every question if it costs you the clock. On SHL there is usually no penalty for wrong answers, so an educated guess on a question you cannot finish beats leaving it blank.

How do you prepare for it?

The best preparation is specific, not abstract. Build the mechanics first, then the speed.

1. Drill one question type at a time until the method is automatic. Syllogisms get a Euler diagram. Arrangements get a grid or ladder. Grouping gets a table. Underlying-assumption questions get the "what has to be true for this conclusion to hold" test. A pen and paper are not optional; the candidates who hold constraints in their head are the ones who run out of time.
2. Separate the three buckets every time. For each prompt, write the conclusion that must follow, the one that could follow, and the one that cannot be supported yet. That single distinction is the muscle the whole test rewards, and it is the one most people skip.
3. Then add the clock. Once the method is reliable, practice timed so 22 to 67 seconds per question stops feeling rushed. Speed without method just makes you fast at being wrong.
4. Use the official sample first. Your SHL invitation email usually links to SHL Direct, the publisher's own practice site, so you train on the exact interface and question style you will face.

Bring the same habit into live case work. Before solving, restate the problem in plain language, identify what the prompt gives you and what it leaves open, and label any assumption as an assumption. When you practice case interview questions, notice how often answer quality depends on disciplined deduction rather than business trivia. If your logic holds in silence but breaks under social pressure, pair this work with behavioral interview practice so you can explain your reasoning while composed. You can drill the underlying separation-of-facts skill in Road to Offer's case practice and get structured feedback on where your logic drifts.

What mistakes cost candidates the most?

The biggest mistake is treating the topic as vocabulary rather than performance. Candidates read a definition, nod, and still over-assume on the next prompt. If your process does not change, the label does nothing.

A second trap is importing real-world knowledge. The test is deliberately abstract. "All footballers are fit" is true inside the test whether or not it matches reality, and adding your own facts is the fastest way to pick a wrong answer that feels right.

A third is confusing confidence with certainty. Interviewers, like the test, respect candidates who know the boundary of what the information supports. "This is the strongest conclusion the data allows" is a better answer than pretending to know more than the prompt gives you.

The last is practicing too broadly. Mixing random question types in every session blurs your improvement. Run focused reps: one session on syllogisms, one on arrangements, one on defending conclusions out loud. That keeps the work measurable and transferable, which also helps if you are a career changer compressing prep into a short window.

Sources

• SHL - Practice Tests and Assessments (checked June 18, 2026)
• SHL Deductive Reasoning Test format and scoring - Psychometric Success (checked June 18, 2026)
• Deductive Reasoning Tests: question types and examples - MConsultingPrep (checked June 18, 2026)
• Deductive Reasoning Tests: practice questions and examples - PracticeAptitudeTests (checked June 18, 2026)
• BCG Cognitive Test format - StrategyCase (checked June 18, 2026)