What does a deductive reasoning test actually measure?
A deductive reasoning test measures whether you can draw the conclusion
that must be true from a set of stated facts, and how quickly you can do
it. It is a test of valid reasoning, not of knowledge. You are given the
premises, and exactly one of the answer options follows from them with
certainty.
The questions are deliberately abstract. They do not rely on any
job-specific expertise or culturally specific knowledge, which is what lets
an employer use the same test across very different candidates. Everything
required to answer is in front of you, so bringing in an outside assumption
is usually how a wrong answer is chosen.
Employers use it because the skill transfers directly to work. Someone who
can take a policy, a rule or a contract clause and say precisely what does
and does not follow from it is doing the same thing a lawyer, an analyst or
an engineer does every day. It is judgement about what the evidence
actually supports.
How is deductive reasoning different from inductive reasoning?
Deductive reasoning works from general rules down to a certain conclusion,
while inductive reasoning works from specific examples up to a probable one.
That difference decides what counts as a correct answer, so it is worth
being clear about before you practise.
In a deductive question the conclusion is guaranteed by the premises. If
the premises are true and your logic is valid, the answer cannot be wrong.
You are looking for what must be true, never for what is merely likely.
In an inductive question, by contrast, you are shown a pattern or a set
of cases and asked for the most probable next step. The answer is a strong
inference rather than a certainty.
Many employers set both, sometimes in the same sitting, and the two are easy
to confuse under time pressure. If you catch yourself thinking “that is
probably right” on a deductive question, stop: probability is the wrong test.
Ask instead whether the option is forced by what you have been told.
What question types appear in a deductive reasoning test?
Most deductive tests are built from a small number of recurring formats, so
recognising them on sight saves time. The four you will meet most often are
syllogisms, conditional statements, logical elimination and arrangement
problems.
- Syllogisms give you two or more statements about groups (“all A are B”)
and ask which conclusion follows.
- Conditional statements are if-then rules, where the skill is knowing
which further statements the rule does and does not license.
- Logical elimination gives you a set of clues and asks you to rule
options out until one remains.
- Arrangement problems ask you to order or seat items under a list of
constraints, then answer a question about the valid arrangement.
Some publishers also present a short passage or a data table and ask whether
a statement is true, false or cannot be concluded. Almost every question is
multiple choice with one defensible answer, so the work is in proving that
answer rather than in recognising it.
How do you solve syllogism questions?
Chain the premises together and use only what is stated, nothing more. A
syllogism is solved by following the links between the groups, not by
checking whether the conclusion sounds sensible.
Take a worked example. All footballers are fit and healthy. All famous
sports players are footballers. Which conclusion follows? Work the chain:
famous sports players are footballers, and footballers are fit and healthy,
so all famous sports players are fit and healthy. That is the only conclusion
forced by the premises.
The tempting wrong answers are built from reversals and leaps the premises do
not support:
- “All footballers are famous sports people” reverses the second premise. Not
given.
- “All fit and healthy people are footballers” reverses the first premise. Not
given.
- “All football players are men” introduces a fact, gender, that never
appears. Not given.
The habit that catches all three: before you accept a conclusion, point to
the premises that force it. If you cannot, it is a distractor.
How do you handle conditional (if-then) statements?
Test each option against the rule, and remember that a conditional works in
one direction only. “If P then Q” tells you that P guarantees Q. It does not
tell you that Q guarantees P.
Two moves are always valid, and two always look valid but are not:
- Valid: given the rule and P, you may conclude Q.
- Valid: given the rule and “not Q”, you may conclude “not P” (the
contrapositive).
- Invalid: given the rule and Q, you may not conclude P (affirming the
consequent).
- Invalid: given the rule and “not P”, you may not conclude “not Q”
(denying the antecedent).
The two invalid moves are the ones examiners bank on, because in everyday
language people slide between them without noticing. When a question hinges
on a conditional, write the rule down as “P then Q” and check which of the
four situations the option describes before you answer.
How do you avoid the 'cannot be concluded' trap?
When the options include “cannot be concluded”, treat it as the correct
answer until the premises rule it out. Most candidates lose these marks by
filling gaps with sensible-but-unstated assumptions.
Consider a statistics example. This year 94% of students achieved A to C.
Last year 95% did. The school credits its new mentoring scheme.* Which
statement is true? The only supportable conclusion is that before last year,
fewer than 94% achieved A* to C, because results have been rising. Anything
about next year’s results, other schools, or every student’s grade goes
beyond the data and must be rejected.
Two rules keep you honest:
- Use only the figures and statements given. A cause the passage asserts
is not a cause you can rely on.
- A plausible statement is not a provable one. If you cannot trace a
statement back to the text, “cannot be concluded” is your answer.
How are deductive reasoning tests scored and timed?
Most employers score you against other candidates rather than against a
fixed pass mark. Your raw score is turned into a percentile, so scoring in
the 70th percentile means you did better than 70% of the comparison group,
and competitive schemes commonly sift around that level.
Timing is tight by design. You will usually have in the region of a minute
per question, and the tests are built so that few candidates comfortably
finish. That has two practical consequences:
- Do not stall on one hard item. A minute lost early is a question you
never reach at the end, and both cost the same mark.
- Guess rather than leave blanks. Negative marking is rare in graduate
deductive tests, so an eliminated-down guess beats an empty answer. Check
the instructions for the occasional exception.
Which employers and test publishers use deductive reasoning tests?
Deductive reasoning tests are standard in graduate and professional hiring
across law, finance, consulting, technology and the civil service. If a role
turns on interpreting rules and evidence, expect one.
Most employers license their tests from a small number of publishers, and the
format varies between them, which is why practising the right style matters:
- SHL produce some of the most widely used deductive tests in the UK. A
typical sitting runs around 18 minutes with 20 questions, split across
verbal statements, image-based items and a block of text to reason from.
- Kenexa, part of IBM, use a similar structure of roughly 20 timed
questions where you deduce the one correct option.
- Employer-specific tests are common at large recruiters, often built on a
publisher’s engine with the employer’s own branding and pass thresholds.
The invitation email usually names the publisher. Find out which one before
you prepare seriously, because it changes what good practice looks like.
What is the best way to prepare?
Practise under the real time limit, and spend as much effort reviewing
answers as taking tests. Untimed practice builds accuracy but not the speed
the test is really measuring.
A sequence that works:
- Take one full test untimed first, reading every worked solution. The
test on this page is built for exactly that. The goal is to learn the
question types, not to score well.
- Then work under the clock, at about a minute a question, and expect
your score to dip at first.
- Review the ones you got wrong and the ones you guessed right. A lucky
guess is a gap you have not found yet.
- Practise your publisher’s format once you know it, since SHL, Kenexa
and employer-built tests do not feel the same.
The bottleneck is almost never cleverness. It is reading the premises exactly,
using only what is stated, and refusing the assumption the question is baiting
you into.
