Inside an IQ Test: The Question Types You Will Actually Meet

Once you have seen a few assessments, the surprising thing is how small the underlying vocabulary is. Beneath the branding and the varying time limits sits a set of item formats that has barely changed in decades, because each one isolates a particular kind of reasoning with unusual efficiency. Knowing what those formats are and what each one is probing does not let you game a well-built test, but it does remove the disorientation that costs people points on the first few items.

Matrix Reasoning

This is the format that has come to define the whole field visually. You are shown a grid of abstract shapes, usually three by three, with the final cell missing, and you choose the option that completes the pattern. Nothing is written, nothing is culturally specific, and there is no knowledge to recall.

What makes matrices so useful to test designers is that difficulty scales cleanly. Easy items involve a single rule applied along one dimension. Hard items stack several rules at once: a shape rotates while a count increases while a fill pattern alternates, and the rules interact rather than sitting side by side. Because the items carry no language, they travel across countries and translations far better than verbal material, which is why they dominate free tests and international instruments alike.

The workable approach is to read across the rows first and establish what changes, then read down the columns and establish what changes there, then check the answer options against both rules rather than picking the one that simply looks right. Attractive wrong answers are deliberately constructed to satisfy one rule and violate the other.

Sequences and Series

A run of numbers, letters or symbols is presented and you supply what comes next. The number version tests whether you can identify an arithmetic relationship: differences, ratios, alternating operations, or two interleaved sequences masquerading as one. The letter version usually works on alphabet positions, so converting letters to their numeric positions turns an opaque puzzle into an ordinary arithmetic one.

Series items are where careless mistakes cluster, because a plausible rule that fits the first three terms and fails on the fourth is very easy to commit to. The discipline that helps is checking your candidate rule against every term given, not just the ones that suggested it in the first place.

Analogies

The classic form is stated as one thing is to another as a third thing is to what. Analogies can be verbal, numerical or visual, and all three probe the same underlying skill: extracting the relationship in the first pair and mapping it onto the second, ignoring surface resemblance.

Verbal analogies are unavoidably sensitive to vocabulary, which is why they behave differently from matrices. Someone who reads widely has a genuine advantage, and that advantage reflects accumulated knowledge rather than raw reasoning speed. The distinction between those two things is a substantial topic in its own right and is set out in the material on fluid and crystallised intelligence. The technique that works is to state the relationship explicitly in a short sentence before looking at the options, since the options are designed to pull you towards words that are merely associated rather than related in the same way.

Odd One Out

Five items are shown and one does not belong. This format looks simple and is quietly one of the most demanding, because you have to generate the classification rule yourself rather than verify one you have been handed.

The characteristic trap is that more than one defensible answer usually exists. In a well-constructed item, one rule accounts for four of the five cleanly while the competing rules account for only three, and the intended answer is the one supported by the tidiest classification. When two candidates seem equally good, the tie-breaker is usually which grouping requires the simpler description.

Spatial and Visual Reasoning

Spatial items ask you to manipulate objects mentally rather than physically. Several recurring versions appear across almost every instrument.

Mental rotation. Two shapes are shown and you decide whether they are the same object viewed from a different angle or genuine mirror images.
Paper folding. A sheet is folded, holes are punched, and you predict the pattern when it is unfolded.
Net folding. A flat net is shown and you identify which three-dimensional solid it produces.
Block design. In supervised sessions, physical blocks are arranged to reproduce a printed pattern under time pressure.
Figure completion. A visual pattern is missing a section and you select the piece that fits.

Spatial ability shows more individual variation than almost any other domain, and it is also the domain that improves most noticeably with practice, which is one reason heavy preparation distorts results in this area more than elsewhere. It is also the format most damaged by a small screen, a practical point worth remembering before starting any timed iq test on a phone rather than a proper display.

Verbal Reasoning and Vocabulary

Verbal sections cover synonyms and antonyms, sentence completion, comprehension of short passages, and logical statements of the form where two premises are given and you judge what necessarily follows. That last type separates people sharply, because the intuitive answer and the logically valid answer diverge more often than most of us would like to believe.

Vocabulary items sit slightly awkwardly in a reasoning assessment, since they measure accumulated exposure rather than inference. They survive because vocabulary correlates strongly with general ability across a population and is extremely quick to administer. Verbally weighted instruments therefore behave differently from pure matrix tests, which is a large part of why the same person can score noticeably differently on two respectable assessments, as the comparison in the piece on official and supervised assessments describes.

Working Memory and Processing Speed

Two domains that free tests almost always skip appear routinely in supervised sessions, and their absence is a significant part of why a quick online result is a partial picture.

Working memory items require you to hold material in mind while doing something to it: repeating digit sequences in reverse, sorting a jumbled list into order from memory, or tracking two streams of information at once. Processing speed items are deliberately easy in content and demanding in pace, such as matching symbols to numbers against a key for two minutes without errors. Neither is about cleverness in the everyday sense, and both are strongly affected by sleep, anxiety and attention, which is one of the clearest routes by which everyday habits show up in a score. That relationship is examined in the material on sleep, diet and cognitive performance.

How Difficulty Is Engineered

Items in a serious instrument are not written by intuition. Each candidate item is trialled on a sample, and two statistics decide its fate: the proportion of people who answer correctly, and how well performance on that item tracks performance on the test overall. An item everyone gets right carries no information. An item that strong performers fail more often than weak ones is broken and is discarded.

The surviving items are ordered so difficulty rises steadily, which is why the opening questions of a good test feel almost insultingly easy. That ramp is doing real work: it settles nerves, confirms you have understood the format, and reserves the discriminating power for the middle and upper range. A test whose difficulty stays flat throughout cannot separate anyone from anyone, and spotting that is one of the quickest ways to judge a free tool, as described in the piece on quick tests and instant results.

Why Real Items Are Kept Out of Circulation

People are often frustrated that they cannot find the actual questions from a major instrument. The restriction is not commercial protectiveness so much as basic measurement logic. An item measures reasoning only while it is novel. Once it has been seen and memorised it measures recall, and the developer has to replace it, restandardise and start again at considerable expense.

This is why publishers sell only to qualified purchasers, why supervised sessions collect materials at the end, and why every set of questions circulating freely online is either a practice item written for the purpose or a leak that has already stopped being useful. The same logic explains why deliberate cramming produces a hollow number, a point developed in the material on preparing for an aptitude assessment.

Getting the Most Honest Result

Familiarity with formats removes wasted effort. It does not, and should not, inflate a well-built score, because the tests are designed to measure how you handle unfamiliar problems rather than how many puzzle types you can name. The useful preparation is unglamorous: sleep properly, sit somewhere quiet, use a real screen, read the instructions carefully, do not agonise over any single item, and accept the result as a band rather than a point.

And keep the scope of the thing in view. Every format described here samples reasoning under artificial conditions with a clock running. None of them touches judgement, motivation, social intelligence or the capacity to keep working at something long after it stopped being interesting. Those matter more, and the material on assessment in childhood and on testing in the UK both come back to the same conclusion from different directions.