Deductive Reasoning Test: Syllogisms, Rules and Preparation

Most marks lost on a deductive reasoning test are not lost to hard logic. They are lost because the candidate answers from what they know rather than from what the premises actually supply. The defining property of the format is entailment: the conclusion has to follow from the stated information alone, and nothing you happen to know about the real world counts as evidence. A statement can be entirely true and still be marked wrong, because it was never entailed.

That same line separates deductive from inductive work. Inductive items hand you a pattern and ask what most likely comes next; the answer is the best supported extrapolation. Deductive items hand you rules and ask what must be the case; one option is forced and every other is wrong however sensible it sounds. Candidates who have only practised pattern work reason towards the plausible option rather than the forced one, and bleed marks without noticing the habit.

Practising for a real assessment? The Watson-Glaser Style Practice Pack runs these question types at full length — 5 timed tests, 150 questions, a worked explanation on every one.

The formats inside a deductive reasoning test

The label covers five mechanically different item types. They share the entailment rule and almost nothing else, so fluency in one buys little in another.

Syllogisms

Two premises and a proposed conclusion, and you mark whether it follows. The machinery is set membership, not plausibility. You are tracking what "all", "some", "no" and "some are not" do to overlapping categories. Two traps produce most errors. The first is the true-but-not-entailed conclusion: given "All auditors are employees" and "Some employees work remotely", the statement "Some auditors work remotely" is believable and is not entailed, because those two groups may be disjoint. The second is reading "some" as implying "not all".

Conditional reasoning items

A rule of the form "if P then Q" plus a fact, and you decide what follows. Four classical forms are tested, and naming them makes the invalid ones visible. Affirming the antecedent (modus ponens: P is true, therefore Q) is valid. Denying the consequent (modus tollens: Q is false, therefore P is false) is valid. Affirming the consequent (Q is true, therefore P) is invalid. Denying the antecedent (P is false, therefore not Q) is invalid. Those two invalid forms generate most wrong answers across the whole test type, because in both the conditional feels harmless run backwards. Mechanically they differ from syllogisms because the work is on a rule and its contrapositive, often chained, rather than on category overlap.

Seating and ordering arrangements

A constraint set covering people around a table, offices along a corridor or events in sequence, then one or more questions. The mechanical skill is elimination, not construction. Candidates lose time trying to build the one true arrangement; the efficient method is to take each option and kill it with a single violated constraint. Many such items have no unique arrangement at all, only a set of arrangements sharing the fact asked about, so building the layout is impossible by design.

Grouping and matching puzzles

Items are sorted into teams, shifts or committees, or matched to attributes. Position is irrelevant here, which separates them from ordering items; only co-membership and counting matter. The constraints that do the work are counting ones ("exactly two of the four must take the morning shift") and exclusion pairs ("R and T cannot share a group"). A grid of ticks and crosses resolves these far faster than prose reasoning.

Statement and conclusion items

A short passage, then conclusions you mark true, false or cannot say. The three-way response is the mechanical difference and the reason these score so poorly. "True" means the passage entails the conclusion. "False" means the passage entails its negation. "Cannot say" means neither, and it is the correct answer far more often than candidates expect. Almost every error here is a cannot-say item marked true because the conclusion sounded reasonable.

Scoring and timing, and what they mean for strategy

Two scoring models are in circulation, and they call for opposite behaviour. Under raw scoring you gain a mark per correct answer and lose nothing for a wrong one, so no question should be left blank. Under formula scoring, sometimes described as negative marking, a fraction of a mark is deducted for each wrong answer while unanswered items cost nothing. When errors cost you and blanks do not, the last thirty seconds are better spent on one item you can settle than on three you cannot. Partial elimination changes that: once two of five options are ruled out on solid grounds, an answer from the remaining three is worth submitting even under formula scoring. A pure guess is not.

Timing models vary as much. A fixed question count against a single clock lets you bank time on easy syllogisms and spend it on arrangement sets. An indefinite bank against a clock, where nobody finishes and the score is how far you get accurately, punishes perfectionism: long deliberation on one item costs you the two you never reached. Per-section timers stop you moving time between item types, so a slow start on the grouping section cannot be rescued later. Per-question timers remove skipping entirely, which makes a fast elimination habit the whole game. Adaptive tests select the next item from your running performance, so early items carry unusual weight and there is normally no way back to review.

The two facts to establish before you sit are therefore the scoring model and whether you can return to skipped items. Everything tactical follows from those.

A worked example

Premises:

1. If a shipment is flagged, it is inspected at the depot. 2. If a shipment is inspected at the depot, it is logged by the night team. 3. Shipment 47 was not logged by the night team.

Which conclusion follows?

A. Shipment 47 was not flagged. B. Shipment 47 was not inspected, but it may still have been flagged. C. Every shipment logged by the night team was flagged. D. Shipment 47 was inspected but not logged. E. Cannot say.

The answer is A. Work backwards through the chain. Premise 3 gives you "not logged". Premise 2 says inspection guarantees logging, so its contrapositive says that anything not logged was not inspected; shipment 47 was therefore not inspected. Premise 1 says flagging guarantees inspection, so its contrapositive says anything not inspected was not flagged. Shipment 47 was not flagged. That is modus tollens applied twice down a chain of two conditionals.

The wrong options are each built from one identifiable mistake. B applies the contrapositive of premise 2 correctly, then declines to apply the contrapositive of premise 1, treating the chain as breaking after one link. C is the converse error, affirming the consequent turned into a general rule: logging follows from inspection, but logged shipments need not be flagged ones. D contradicts premise 2 outright, and is chosen by those who read "inspected at the depot" as a free-standing fact rather than a term bound by a rule. E imports the cannot-say habit into a conditional item, where the chain does in fact close.

How to prepare for this type specifically

Learn the four conditional forms by name and drill them until affirming the consequent looks as wrong on the page as a spelling error. Write out the contrapositive of every rule you meet, because the valid inference in a chained conditional is almost always the one running backwards through the negations.

For syllogisms, practise the specific discipline of asking whether a conclusion is entailed rather than whether it is true. Take a plausible conclusion you marked as following and try to build any situation where the premises hold and it fails. If you can build one, it does not follow.

For arrangement and grouping sets, train elimination rather than construction. Time yourself on the single skill of taking a proposed layout and finding the constraint it violates, and build the habit of sketching a quick grid or number line before reading the options.

For statement and conclusion items, practise justifying every "true" by pointing at the words that entail it. If you cannot underline the entailing text, the answer is cannot say.

Finally, establish how your version is scored and timed before the day. Whether wrong answers are penalised decides what to do with the last three items you never properly read; whether the test is adaptive decides how much care the opening items deserve.

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.