Abstract reasoning is the ability to spot rules, relationships, and patterns in shapes or symbols without relying on language, numbers, or facts you've learned. Matrix and pattern questions test this directly: you're shown a grid of shapes with one missing, and you have to work out which option logically completes it. No vocabulary, no cultural knowledge, no maths formulas — just pure pattern logic.
This is exactly why abstract reasoning shows up so heavily on IQ (intelligence quotient, a score that ranks reasoning ability against the general population) tests. It's considered one of the purest ways to measure raw thinking power, stripped of everything you might have picked up from school or life experience.
What abstract reasoning actually measures
Abstract reasoning taps into what psychologists call fluid intelligence — your ability to reason through brand-new problems on the spot, without relying on stored knowledge or prior experience. This is different from crystallized intelligence, which is the vocabulary, facts, and learned skills you've built up over years. If you want the full breakdown of how these two interact, fluid vs crystallized intelligence covers it in depth.
Matrix questions are the classic tool for measuring fluid intelligence because they can't really be "studied" for in the traditional sense. You can't memorise your way to a correct answer — you either see the rule governing the sequence, or you don't. That's also why these questions correlate so strongly with what researchers call the g factor (general intelligence, the common thread statisticians find running through performance on many different types of cognitive tasks). A person who's good at spotting patterns in shapes tends to also do well across verbal, numerical, and spatial tasks — not because the content overlaps, but because the same underlying reasoning ability is doing the work each time.
Why nonverbal formats matter
Language-based questions inevitably reward people who happen to have a rich vocabulary in that specific language, or who grew up in an education system that emphasised certain kinds of knowledge. Abstract reasoning sidesteps that problem almost entirely. A grid of triangles rotating and changing colour means roughly the same thing to a teenager in Tokyo, a retiree in Toronto, or a factory worker in Lagos. This is part of why the format is favoured in cross-cultural research and why it forms the backbone of one of the most referenced tests in the field, Raven's Progressive Matrices — a test built entirely around this principle since the 1930s.
How a matrix question is actually built
A typical matrix item presents a 3x3 grid of shapes with the bottom-right cell left blank. Your job is to identify the rule (or rules) governing how shapes change across each row and column, then pick the answer option that fits.
Test designers usually layer several types of rules into a single item, which is why difficulty can escalate quickly:
- Progression — a shape gets bigger, rotates further, or gains more sides as you move along a row.
- Distribution of three — three different shapes or colours appear once in each row and column, like a visual sudoku.
- Addition or subtraction — overlapping two cells produces the third (e.g., combine the shapes in columns one and two to get column three).
- Constant with variation — one feature stays fixed while another changes systematically.
- Figure rotation or reflection — the same shape appears flipped or rotated by a consistent angle each step.
Easier items usually rely on just one rule applied consistently. Harder items stack two or three rules on top of each other simultaneously, forcing you to track multiple relationships at once — which is where working memory (your capacity to hold and manipulate information in mind for a few seconds) starts to matter alongside pure pattern spotting.
A simplified example of the logic
Imagine a row where a single triangle becomes a double triangle, then a triple triangle. That's a simple progression rule: shapes increase by one each step. Now imagine the same row also rotates each shape 90 degrees and changes its colour on a fixed cycle. Suddenly you're tracking three independent variables at once. That layering — not any single rule — is what makes advanced matrix items genuinely hard.
Why these are called "Raven style" questions
Raven's Progressive Matrices, created by John C. Raven, is the test that popularised this exact format, and it remains one of the most cited nonverbal reasoning tools in psychology. "Raven style" has become shorthand in the testing world for any matrix-based, nonverbal pattern recognition test that follows this same blueprint: a grid, a missing piece, multiple-choice options, and increasing difficulty as you progress through the set.
The format has been adopted widely, including in commercial and workplace assessments, precisely because it's quick to administer, hard to game, and translates well across different countries and reading levels. If you're curious how this format stacks up against other testing approaches like WAIS or Cattell, types of IQ tests compared walks through the differences.
Common misconceptions about pattern questions
"They're just an IQ test for shapes." Not quite — abstract reasoning is one component that feeds into an overall IQ score, but it's typically combined with verbal, numerical, and sometimes spatial subtests. Matrix reasoning specifically targets fluid intelligence, while a full test also captures crystallized knowledge and processing speed.
"You either have the talent or you don't." Practice does help you get familiar with the recurring rule types listed above, which can modestly improve your speed and confidence. But it won't dramatically change your underlying pattern-spotting ability the way memorising vocabulary can inflate a verbal score. For a realistic take on what training can and can't do, see can you improve your IQ.
"Harder just means more shapes." Actually, difficulty in well-designed matrix tests usually comes from stacking rules, not from adding visual clutter. A cluttered-looking grid with one clear rule can be easier than a clean-looking grid hiding three overlapping rules.
How matrix scores compare to the population
Like most cognitive test formats, abstract reasoning performance is scored against a norm (a reference group of people who've already taken the test, used as the comparison baseline) and typically converted onto the familiar IQ scale, where the average is 100 and roughly two-thirds of people score between 85 and 115. This follows the same bell curve shape used across the IQ bell curve, and your raw number of correct matrix items gets translated into a percentile (the share of people your score outranks) so it means something relative to others, not just as an isolated tally.
| Matrix score band | Rough percentile | What it suggests |
|---|---|---|
| Well below average | Bottom 10% | Difficulty with multi-rule patterns, may benefit from untimed practice |
| Below average | 10th–25th | Handles single-rule progressions, struggles with layered rules |
| Average | 25th–75th | Solves most standard items within typical time limits |
| Above average | 75th–90th | Comfortable with 2–3 stacked rules per item |
| Well above average | Top 10% | Rapidly decodes complex, multi-variable grids |
Key takeaways
- Abstract reasoning measures fluid intelligence — the ability to solve new problems using logic alone, without relying on learned facts.
- Matrix questions work by hiding a rule (or several layered rules) inside a grid of shapes, and you have to infer it from context.
- Common rule types include progression, distribution of three, addition/subtraction of shapes, and rotation.
- The format is called "Raven style" after Raven's Progressive Matrices, the test that made this design widely known.
- Nonverbal formats reduce the influence of language and cultural background, making them a fairer comparison across different people.
- Difficulty typically comes from stacking multiple rules in one item, not just adding visual complexity.
Abstract reasoning questions are a genuinely useful window into how you think through unfamiliar problems, but like any single test format, they're only one piece of a bigger picture. A test like iqmetria's is meant for self-knowledge and curiosity — it's orientative, not a clinical or medical diagnosis, and it won't tell you everything about your mind. Treat it as a fun, honest snapshot, not a final verdict.
FAQ
What is an example of abstract reasoning?+
A classic example is a matrix question: a 3x3 grid of shapes with one missing, where you must figure out the rule linking rows and columns (like shapes rotating or growing) and pick the option that completes the pattern logically.
Is matrix reasoning the same as IQ?+
No. Matrix reasoning is one component that feeds into an overall IQ score, specifically targeting fluid intelligence. A full IQ test usually also includes verbal, numerical, and processing-speed sections.
Can you improve at Raven style questions?+
You can get faster and more confident by learning the common rule types (progression, rotation, distribution of three), but this mostly improves familiarity and test-taking speed rather than dramatically changing your underlying reasoning ability.
Why do abstract reasoning tests use shapes instead of words?+
Shapes and patterns remove the advantage that comes from vocabulary, education, or cultural background, making the test fairer to compare across different people and languages.
What skill does abstract reasoning measure?+
It measures fluid intelligence — your ability to spot rules and solve new, unfamiliar problems using logic alone, rather than relying on facts or knowledge you've already learned.