The 5 Types of Reasoning Every IQ Test Measures (and Why)
Pattern recognition, logic, numbers, verbal, and spatial reasoning: the five question types behind almost every IQ test, explained with real examples.
If you’ve ever sat down to take an IQ test, you’ve probably noticed something: the questions jump around a lot. One minute you’re staring at a grid of shapes trying to figure out what goes in the missing box. The next you’re untangling a sentence full of “if this, then that” logic. Then suddenly there’s a number sequence, or a word you haven’t thought about since high school, or a shape that needs to be rotated in your head.
That’s not random. Almost every well-constructed IQ test, including our own free IQ test, is built around a handful of core reasoning categories that researchers have settled on over the better part of a century. Different tests package them differently and use different names, but the underlying skills tend to fall into five broad buckets: pattern recognition, logical reasoning, numerical reasoning, verbal reasoning, and spatial reasoning.
Knowing what each one is actually testing (and why test designers bother including it) makes the whole experience less mysterious. It also helps you understand your own results better, because a score is really five different stories rolled into one number. Let’s go through them one at a time.
Pattern recognition (matrix and sequence reasoning)
This is probably the category people picture first when they think “IQ test.” You’re shown a grid, usually three rows by three columns, filled with shapes, lines, or symbols that change according to some rule. One cell is missing, and you have to figure out what belongs there based on the pattern.
The most famous real-world example of this format is Raven’s Progressive Matrices, developed by psychologist John C. Raven in the 1930s. It’s still referenced constantly in psychology and testing literature today, largely because it’s designed to rely as little as possible on language or specific cultural knowledge. You’re not being asked to know a fact. You’re being asked to notice how shapes rotate, combine, or multiply across a sequence, and then apply that rule to fill in a gap.
Here’s a simple, made-up example just to show the flavor of it: imagine a row of shapes going circle, circle-with-one-dot, circle-with-two-dots. What comes next? A circle with three dots, obviously, once you see that the rule is “add one dot each step.” Real test items are more visually complex and often combine two or three rules at once (rotation plus color change plus a shifting number of elements, say), but the underlying skill is the same: spot the rule, then extend it.
This category is often treated as one of the purer measures of what psychologists sometimes call fluid intelligence, meaning the ability to reason through something brand new rather than recall something you already learned. That’s part of why it shows up so consistently across different tests and different countries’ testing traditions. If you want to feel what this is actually like rather than just read about it, it’s worth trying a pattern recognition question yourself in our free test.
Logical reasoning (deduction and conditional thinking)
Where pattern recognition is mostly visual, logical reasoning is mostly verbal or symbolic. These questions ask you to work through statements and figure out what must be true, what could be true, and what’s actually a trap.
Two structures show up again and again. The first is the classic syllogism: you’re given two premises and asked what conclusion follows. Something like “All squids have eight arms. This creature has eight arms. Is it necessarily a squid?” (It isn’t, of course, since plenty of things have eight arms without being squids. That’s the point of the exercise.)
The second structure is conditional, “if-then” reasoning. You’re given a rule like “if it rains, the game is cancelled,” and then a scenario, and asked to figure out what can be validly concluded. This is where one of the most well-known logical traps in the whole field shows up: affirming the consequent. If you’re told “if it rains, the game is cancelled,” and then someone tells you “the game was cancelled,” it’s tempting to conclude “so it must have rained.” But that doesn’t actually follow. The game could have been cancelled for a dozen other reasons (a scheduling conflict, a power outage, whatever). Good logical reasoning questions are specifically designed to see whether you fall for that trap or catch it.
This category matters because it isolates something slightly different from pattern recognition. It’s less about spotting a visual rule and more about tracking the structure of an argument carefully enough not to get fooled by something that sounds right but isn’t. Our free IQ test includes logic questions built around exactly this kind of conditional and syllogistic reasoning, so it’s a good place to check your own instincts against a real if-then trap.
Numerical reasoning (sequences, arithmetic patterns, and math riddles)
Numerical reasoning gets lumped in with “math ability” a lot, but it’s really testing something narrower. Nobody’s checking whether you remember the quadratic formula. They’re checking whether you can spot a pattern when the pattern happens to be made of numbers instead of shapes or words.
Number sequence questions are the most common format: you’re given a string of numbers and asked what comes next. Something simple, like 2, 4, 8, 16, ___, where the rule is “double the previous number” and the answer is 32. Real test items get more layered than that (alternating rules, two interleaved sequences, or a rule based on the position of the number rather than the value before it), but the basic move is identical: find the rule, then extend it.
There’s also a subcategory that leans more toward lateral thinking than pure arithmetic. These are riddle-style problems where the “obvious” math approach leads you astray and you have to reframe the question. A simple original example: “A farmer has 17 sheep, and all but 9 die. How many are left?” If you rush and calculate 17 minus 9, you’ll get 8, which is wrong. The answer is 9, because “all but 9” means 9 survive. It’s less about calculation and more about reading carefully and resisting the urge to jump straight to arithmetic.
This category tends to correlate with comfort around numbers, but it’s really measuring flexible thinking under a numerical disguise. If sequences like these are your thing (or your weak spot), our free test’s numerical section will give you a quick read on where you land, and it’s worth trying alongside the other four categories rather than in isolation.
Verbal reasoning (vocabulary and analogies)
Verbal reasoning covers two closely related skills: how deep your vocabulary runs, and how well you can map relationships between concepts. The vocabulary side is fairly self-explanatory: you’re asked to define a word, pick a synonym, or spot the odd one out among a list of related terms.
The analogy side is more interesting and a bit trickier. You’re given a relationship like “hot is to cold as” and asked to complete it with a matching pair, such as “up is to down.” The skill being tested isn’t just knowing what the words mean individually, but recognizing the type of relationship between them (opposites, in that case) and finding another pair that shares the same structure. A simple original example: “book is to library as painting is to ___.” The relationship is “item to the place it’s typically housed or displayed,” so “museum” or “gallery” fits.
Here’s a fair caveat worth being upfront about: verbal reasoning performance is influenced more heavily by reading habits and exposure to a particular language than some of the other categories on this list. Someone who reads widely and has spent years immersed in a language’s idioms and less common vocabulary is likely to have an edge on this section that has less to do with raw reasoning ability and more to do with accumulated exposure. That doesn’t make the category meaningless (the analogy-mapping skill in particular is still a genuine reasoning task), but it’s a reasonable caveat to keep in mind when comparing verbal scores to, say, matrix reasoning scores, which are designed to lean on language much less. If you want to see how your own vocabulary and analogy instincts hold up, our free IQ test covers all five categories, including this one.
Spatial reasoning (mental rotation and 3D visualization)
Spatial reasoning asks you to manipulate shapes in your head rather than on paper. The classic format is mental rotation: you’re shown an object and asked to identify which of several other images is the same object rotated in space, versus which ones are actually mirror-image reflections dressed up to look like rotations.
A related format asks you to imagine a flat, folded-out 2D shape (think of a cardboard box pattern before it’s folded up) and figure out what it would look like once it’s folded into a 3D object, or which face would end up opposite which. Both formats are really testing the same underlying skill: can you hold a shape in your mind’s eye and transform it accurately without physically moving anything?
The foundational research here goes back to Roger Shepard and Jacqueline Metzler’s mental rotation studies from the 1970s, which are still commonly cited as the starting point for how psychologists think about this kind of visualization ability. Their work showed that the time it takes someone to mentally rotate an object tends to scale with how far the object has to be rotated, which is a pretty elegant way of demonstrating that people really are doing something like rotation in their heads rather than just guessing.
Spatial reasoning is often discussed in relation to fields like engineering, architecture, and other STEM disciplines, where visualizing how parts fit together in three dimensions is a daily requirement. It’s also frequently cited as one of the more trainable categories on this list: unlike some other reasoning skills, spatial visualization tends to improve with deliberate practice, whether that’s through puzzles, certain video games, or hands-on work with 3D objects. If this sounds like your kind of challenge (or your kryptonite), see how you do on the spatial reasoning questions in our test.
Why these five categories, specifically?
There’s nothing magical about the number five. Some test batteries split things further (separating short-term memory or processing speed into their own categories, for instance), and some combine categories differently. But pattern recognition, logical reasoning, numerical reasoning, verbal reasoning, and spatial reasoning keep showing up across different testing traditions because together they cover a genuinely broad slice of reasoning: visual pattern-spotting, structured argument, quantitative pattern-spotting, language-based relationships, and 3D visualization.
No single category tells the whole story on its own. Someone can be excellent at spotting matrix patterns and merely average at verbal analogies, or vice versa. That’s exactly why a well-rounded test measures all five rather than leaning on just one, and it’s the reasoning behind how we built our own free test.
Frequently asked questions
Do all IQ tests use these same five categories?
Most well-known IQ tests draw on some version of these five reasoning types, though the exact labels and the number of subtests vary. Some tests add categories like working memory or processing speed as separate sections. The core idea of testing several distinct reasoning skills rather than just one, though, is close to universal.
Which category is hardest to improve with practice?
Pattern recognition and logical reasoning are often considered less responsive to short-term practice than something like spatial reasoning, since they’re closer to raw, on-the-spot problem-solving rather than a skill you can drill directly. That said, familiarity with the question format itself (knowing what to look for) can still help your performance even if it doesn’t change the underlying ability much.
Is verbal reasoning a fair measure of intelligence?
It’s a reasonable measure of one kind of reasoning: relationship-mapping between concepts. But because it leans more heavily on vocabulary built up through reading and language exposure than categories like matrix or spatial reasoning do, it’s worth treating a verbal score as somewhat separate from the more language-independent categories rather than as the single best stand-in for general reasoning ability.
What is Raven’s Progressive Matrices, and is it still used?
It’s a pattern recognition test developed by John C. Raven in the 1930s, built around grids of shapes with one missing piece. It’s still widely referenced in psychology today, largely because the visual, largely language-free format makes it a useful tool for comparing reasoning ability across people with different educational or language backgrounds.
Can I get better at spatial reasoning questions?
Many people find that spatial visualization improves with deliberate practice more readily than some other reasoning categories. Activities that involve mentally manipulating 3D objects, from certain puzzle games to hands-on spatial tasks, are commonly cited as helpful, though results will naturally vary from person to person.
How long does it take to cover all five categories in a test?
It depends on the test, but a well-built short-form test can sample all five categories in well under half an hour. Our free test is built specifically to touch on all five without dragging the experience out.
