Logical Reasoning Test: Deductive, Inductive and Critical Thinking

Logical reasoning tests are used to assess whether you can follow rules, identify relationships, and reach valid conclusions from limited information. They can look like statement logic, pattern puzzles, diagrams, matrices, sequences, or mixed reasoning questions. This guide explains the main formats, how to tell them apart, and how to practise effectively.

"Logical reasoning" is the least informative label in graduate assessment. It is not one test. Two candidates invited to a logical reasoning stage on the same morning, for different employers, can face papers with nothing in common: one gets a grid of rotating shapes and no words, the other gets blocks of text with premises and conclusions and no shapes. Both are correctly described as logical reasoning by the publisher who wrote them.

That matters because the two halves of the umbrella run in opposite directions. Inductive items ask you to move from particular instances to the general rule behind them. Deductive items hand you the rule and ask what necessarily follows. The mental operation is reversed, the meaning of a wrong answer changes, and the habits that make you fast at one can slow you down at the other. Working out which kind you are facing is the first real task.

Practising for a real assessment? The Abstract & Inductive Reasoning Practice Pack runs these question types at full length — 5 timed tests, 150 questions, a worked explanation on every one.

The formats inside this test type

Five distinct question formats appear under the logical reasoning heading. They share a name and very little else.

Inductive: infer the governing rule

You are shown a sequence or a set of figures that were all generated by the same hidden rule, and you extend it. Nothing tells you what the rule is; you recover it by comparing instances. A typical item gives five frames in order and asks for the sixth, where shape count increments, fill alternates and an arrow rotates by a fixed amount each frame, all at once. Mechanically, the answer is the best-supported inference, not a proven one. You are choosing the option consistent with every instance you were given, and a rule that explains four frames out of five is simply wrong, however elegant it looks. Most of these items are non-verbal, which is why they travel well across languages.

Deductive: work out what is entailed

You are given premises in plain language and asked which statement must be true. The rule is supplied; your job is to apply it exactly. The classic engine here is the conditional and its contrapositive: if every A is a B, and no B is a C, then no A is a C. Mechanically this is closed. Outside knowledge is not merely unnecessary, it is a trap, because the premises may describe a world that does not match yours. An option that is likely, sensible or true in real life still scores zero if the premises do not force it.

Diagrammatic or process: inputs through operators

A diagram shows an input travelling along a line through one or more labelled operators, each of which transforms it, and you either predict the output or work backwards to identify what an unknown operator does. The distinguishing mechanic is order dependence. Two operators applied in the opposite sequence usually give a different result, so the arithmetic of the transformation matters less than tracking state at each step. These items reward writing the intermediate states down rather than holding them in your head.

Abstract matrix: the missing cell

A grid, most often three by three, has one cell blank. The relationship runs across rows and down columns simultaneously, and the answer is the figure satisfying both. Unlike a linear sequence, a matrix gives you two independent lines of evidence for the same rule, so you can derive a candidate from the rows and verify it against the columns before looking at the options at all.

Verbal logic: syllogisms and conditionals

Text-only items covering syllogistic forms, conditional statements, and ordering puzzles where a handful of constraints fix an arrangement. The mechanic that catches people is quantifier handling. "All", "some", "no" and "only" behave differently under reversal, and "some A are B" never licenses "some A are not B".

Scoring and timing, and what they mean for strategy

Two scoring models are in circulation and they demand opposite behaviour. Under raw scoring, you get a mark for each correct answer and nothing is taken away for a wrong one, so every blank is a wasted chance and you should answer everything. Under formula scoring, marks are deducted for wrong answers and nothing is deducted for unanswered ones. Where formula scoring is used, guessing at the buzzer lowers your score. Filling the remaining boxes in the last few seconds is worse than leaving them empty, and the only guessing that survives the arithmetic is guessing after you have eliminated options, because narrowing five choices to two changes the balance completely.

Timing models vary just as widely. A fixed question count against a fixed clock means the paper is finishable and pace can be planned backwards from the total. An indefinite bank against a clock means nobody is expected to reach the end, the score is items correct within the window, and hesitating over one stubborn matrix costs you two easy ones further down. Per-section timers ring-fence each block, so time saved on the verbal section cannot be carried into the abstract one. Per-question timers remove the choice entirely: you cannot bank time, cannot return to a skipped item, and must commit before the clock expires. Adaptive versions adjust difficulty to your running performance and weight accuracy early, which makes rushing the opening questions expensive.

Read the instruction screen for two things: whether wrong answers carry a penalty, and whether you can revisit skipped items. Those two facts determine your whole approach and are usually stated plainly.

A worked example

Premises: every technician holding a Level 3 certificate is cleared for night shifts. No one cleared for night shifts is assigned to the calibration bay. Priya is a technician assigned to the calibration bay.

Which of the following must be true?

A. Priya does not hold a Level 3 certificate. B. Priya holds a Level 3 certificate but was refused clearance. C. Everyone in the calibration bay has been refused a Level 3 certificate. D. Some technicians holding Level 3 certificates are assigned to the calibration bay. E. Priya will be cleared for night shifts once she completes her certification.

The answer is A. Chain the first two premises: holding Level 3 guarantees clearance, and clearance rules out the calibration bay, so holding Level 3 rules out the calibration bay. Priya is in the calibration bay, therefore she cannot hold the certificate. That is the contrapositive, and it is forced.

The wrong options are each built from a specific misapplied rule. B treats clearance as a discretionary decision that could be granted or withheld, importing an assumption the premises never make; the premises make clearance automatic for Level 3 holders. C generalises from one instance to a whole group, an inductive move in a deductive item, and it also swaps "does not hold" for "was refused", which is not the same claim. D contradicts the chain outright and appeals to anyone who read the second premise as "some are not" rather than "none are". E adds a timeline the premises say nothing about, converting a static constraint into a prediction.

How to prepare for this type specifically

Before anything else, find out which format you are sitting. The invitation email and the practice link usually name the underlying test even when the stage is labelled generically, and knowing whether it is non-verbal or text-based changes what is worth practising.

Then practise the two directions separately, because they train different reflexes. For inductive and matrix work, drill the habit of listing the dimensions that can vary, such as count, rotation, shading, size, position and symmetry, and checking each one across every frame before you look at the options. Deriving the rule first and predicting the answer beats scanning five near-identical pictures.

For deductive and verbal logic, drill quantifiers and conditionals until reversal errors stop feeling plausible. Write each premise as a short arrow statement, add the contrapositive underneath, and only then read the options. Most wrong answers in this format are converse errors, and they look convincing precisely because they are close to something true.

For diagrammatic items, practise recording intermediate states rather than tracking them mentally, and practise working backwards from a known output, since identifying an unknown operator is a separate skill from applying a known one.

Finally, establish the scoring model in advance. Whether wrong answers cost marks decides whether to commit on the last three items or leave them blank, and that decision is worth more than another hour of untargeted drilling.

Frequently asked questions

TestSolve is independent and not affiliated with any test provider or employer named on this page. All product names and trademarks belong to their respective owners.