What does an analytical reasoning test actually measure?
An analytical reasoning test measures how well you take a body of
information apart, apply logic to it and reach a sound conclusion under
time pressure. It is less about a single skill than about a way of
thinking: reading a problem carefully, separating what matters from what
does not, and building an answer you can defend.
Because that way of thinking is useful almost everywhere, the test is one of
the most widely set. Employers use it to find candidates who will be quick
to learn, who stay calm when a problem is unfamiliar, and who can work
accurately against the clock rather than freezing or rushing.
The questions are usually timed at around a minute each, and the material is
chosen so that no specialist knowledge is needed. What is being tested is
not what you already know, but how cleanly you can reason from what you are
given.
What types of analytical reasoning test are there?
“Analytical reasoning” is an umbrella term, and an employer picks the type
that matches the skills they care about. The four you are most likely to
meet are verbal, non-verbal, inductive and deductive reasoning.
- Verbal reasoning gives you passages of text and asks you to judge what
they do and do not support, usually as true, false or cannot say.
- Non-verbal reasoning gives you graphs, tables and figures and asks you
to draw conclusions and spot patterns in the data.
- Inductive reasoning gives you a sequence or a set of cases and asks for
the most probable next step or the underlying rule.
- Deductive reasoning gives you stated facts and asks what must follow
from them with certainty.
If your invitation names the type, prepare for that one specifically. If it
does not, work across all four, because familiarity with the question style
is most of the battle when the clock is running.
How do you approach a verbal analytical question?
Decide each answer only on what the passage states or logically implies,
never on what you happen to know. The whole skill is separating fact from
inference, and the wrong options are written to blur that line.
Work in a fixed order. Read the statement you are being asked about first,
then read the passage looking specifically for the evidence that bears on it.
Choose:
- True only if the passage clearly supports the statement.
- False only if the passage clearly contradicts it.
- Cannot say whenever the passage neither supports nor contradicts it,
even if the statement seems likely from general knowledge.
“Cannot say” is the option candidates most often get wrong, because it feels
like a non-answer. It is frequently the correct one. If you find yourself
justifying “true” with a fact the passage never mentions, the answer is
almost certainly “cannot say”.
How do you approach non-verbal and data questions?
Read the question before you study the chart, so you only pull the figures
you actually need. A typical data stimulus contains far more information than
any single question uses, and time spent absorbing the rest is time lost.
A reliable order:
- Read the question stem and note the exact figure or comparison it
wants.
- Check the axis labels, units and any table headings. This is where the
trap usually sits, for example a rate quoted per week against a question
asked per year.
- Take only the values you need and do the arithmetic on those.
- Use a chart for shape and a table for precision. Charts show direction
and comparison but are rarely exact enough to answer from.
For a worked example: if a click cost 18p in April on one platform, and 14p
plus 6p across two others, the ratio of the single cost to the combined cost
is 20 to 18, which simplifies to 1 to 1.11. The method is always the same:
extract the right figures, then compute.
How do you work through a question with several constraints?
Write the constraints down, fix whatever they make certain, then use the rest
to eliminate. Constraint-heavy questions, where a set of rules pins down an
order or an arrangement, reward bookkeeping far more than cleverness.
A dependable routine:
- List every rule in short form before you look at the answer options, so
you are reasoning from a clean summary rather than re-reading prose.
- Apply the most restrictive rule first. The rule that allows the fewest
possibilities does the most work and often fixes a position outright.
- Combine rules to force further placements, noting each deduction as you
go so you do not lose track.
- Test each option against your rules and discard any that breaks even
one. With a well-constrained question, only one option survives.
The common mistake is holding the whole puzzle in your head. Under time
pressure that is where errors creep in, so trust the written notes over your
memory.
How do you tell a valid deduction from a likely inference?
Check whether the conclusion is forced or merely probable, because analytical
tests reward the two differently. A deductive item wants the answer that must
be true; an inductive item wants the answer that is most likely.
On a deductive question, a conclusion counts only if the premises
guarantee it. “Probably” is not good enough, and an option that goes one step
beyond what you were told is wrong however sensible it sounds.
On an inductive question, you are extrapolating from a pattern or a set of
examples, so you are choosing the best-supported answer rather than a certain
one. Look for the simplest rule that fits every case you have been shown, then
apply it.
Naming which kind of question you are on, before you answer, stops the most
common slip: demanding certainty where only likelihood is available, or
settling for “likely” where the question needed proof.
How are analytical reasoning tests scored?
Most analytical tests are scored comparatively, not against a fixed pass
mark. Your raw score is compared with a normative group, often other
candidates or previous applicants, and reported as a percentile that shows
how you did relative to them.
That has two consequences for how you approach the test:
- Speed counts as much as accuracy. Many tests are built so that few
candidates finish, so an unanswered question costs you ground against people
who attempted it.
- Guessing usually pays. Negative marking is uncommon, so if time is
running out, eliminate what you can and commit to an answer. Check your
instructions for the occasional exception.
Which employers and roles use analytical reasoning tests?
Analytical reasoning tests are common wherever a role involves untangling
information and making decisions from it, which is why they appear across
management, finance, engineering and data analysis, and heavily in science,
technology and IT.
Employers choose the version that fits the job. A role built on written
judgement tends to draw a verbal test; a data-heavy or technical role tends
to draw non-verbal, inductive or deductive formats. Large recruiters often
license these from a test publisher, sometimes rebranded with their own name
and pass thresholds.
Because the underlying skills, spotting patterns, weighing evidence and
solving problems quickly, are valued almost everywhere, an analytical test can
turn up in graduate schemes well outside the obviously technical sectors. If
the invitation names a publisher or a format, let that steer your practice.
What is the best way to prepare?
Practise every format under the clock, then spend real time reviewing what
you got wrong. Working questions untimed builds accuracy but not the speed the
test is actually measuring.
A sequence that works:
- Take one full test untimed first, reading every worked solution. The
test on this page is built for that. The aim is to learn how each question
type behaves, not to score well.
- Move to timed practice at about a minute a question, and expect an
early dip while you adjust.
- Review wrong answers and lucky guesses alike. A guess that landed is a
gap you have not closed.
- Weight your practice towards the format your employer uses, once you
know it, and towards your own weakest question type.
