You open an IQ test and see a grid of shapes. Each row changes in some quiet, deliberate way: a line rotates, a dot moves, a shape gains a side. The bottom-right square is empty, and you have to pick the piece that completes the pattern.
Puzzles like this can feel like a party game, but they are built on more than a century of research. Here is what is actually happening when you take an IQ test, from the types of questions to how a raw number of correct answers becomes a score.
A short history of the IQ test
Modern intelligence testing began in early 20th-century France, when Alfred Binet and Théodore Simon developed a test to identify schoolchildren who needed extra support. Their work was later adapted in the United States into the Stanford-Binet scales.
In the mid-20th century, psychologist David Wechsler created tests that are still the most widely used clinical instruments today, including the Wechsler Adult Intelligence Scale (WAIS). Around the same period, John C. Raven developed Raven's Progressive Matrices, a test built entirely from visual pattern puzzles.
Different tests, different formats, but the same goal: to measure how well a person reasons, learns and solves problems.
What IQ tests are trying to measure
Psychologists noticed early on that people who do well on one type of mental task tend to do well on others. Charles Spearman described this shared factor as general intelligence, or g. Most IQ tests aim to capture this general ability, along with more specific skills beneath it.
Many tests also draw on the distinction between two broad kinds of ability:
- Fluid intelligence: reasoning through new problems you have not seen before.
- Crystallized intelligence: using knowledge and skills you have already learned.
Pattern puzzles lean heavily on the first. Vocabulary and general knowledge questions lean on the second. We go deeper into this split in fluid vs crystallized intelligence.
Inside the matrix puzzle
Matrix reasoning puzzles are probably the most recognizable IQ question type, and for good reason. They require very little language or cultural knowledge, so they focus on raw pattern-finding and abstract reasoning.
A typical matrix item works like this:
- You see a grid, often 3×3, with one cell missing.
- Each row and column follows one or more hidden rules.
- You infer those rules and choose the option that satisfies all of them.
The rules can stack up as items get harder. Common ones include:
- Progression: a shape grows, shrinks or rotates step by step.
- Addition or subtraction: elements from two cells combine to form the third.
- Distribution: each row contains one of three variants, in a different order.
- Constant features: one property stays fixed while others change.
Early items might use a single rule. Later items combine several at once, which puts heavy demands on working memory. You have to track multiple rules simultaneously and test each answer option against all of them.
Other question types you might see
Full clinical batteries like the WAIS use many subtests, grouped into index scores such as verbal comprehension, working memory and processing speed. Online tests are usually shorter and often focus on one or two of these areas.
| Question type | What it asks you to do | Main ability tapped |
|---|---|---|
| Matrix reasoning | Complete a visual pattern | Fluid reasoning |
| Number series | Find the rule in a sequence of numbers | Fluid and quantitative reasoning |
| Vocabulary | Define or match words | Crystallized knowledge |
| Analogies | Complete "A is to B as C is to ?" | Verbal and abstract reasoning |
| Digit span | Repeat sequences forward or backward | Working memory |
| Symbol coding | Match symbols quickly under time pressure | Processing speed |
From correct answers to an IQ score
Here is the part most people never see. Your raw score, the number of items you got right, means very little on its own. It only becomes an IQ score through norming.
Before a test is released, it is given to a large sample of people chosen to reflect the population. Their results define what typical performance looks like at each age. Your raw score is then compared with that reference group and converted to a scale where the average is 100 and the standard deviation is 15.
That is why:
- A 25-year-old and a 70-year-old with the same raw score may receive different IQ scores.
- Tests are re-normed periodically, partly because average performance has shifted over time, a pattern known as the Flynn effect.
- About 68% of people end up scoring between 85 and 115.
For a full walk-through of score ranges and percentiles, see what is a good IQ score?.
What makes a test reliable and valid
Psychologists judge tests on two main qualities.
Reliability is consistency. A reliable test gives similar results if you take it again under similar conditions. Well-built clinical tests are highly reliable, though every score still carries some margin of error.
Validity is whether the test actually measures what it claims to. For IQ tests, this is supported by evidence that scores relate to other measures of reasoning and to outcomes like academic performance.
Clinical tests vs online tests
It is worth being clear about the difference. A clinical assessment is administered one-on-one by a trained psychologist, takes considerable time, covers many abilities and is used to inform real decisions, such as educational support.
An online test is a convenient way to get an estimate of your reasoning ability. A well-designed one, built around solid item types like matrix puzzles, can give you a meaningful snapshot. It is not a diagnosis, and it should not be used as one.
The bottom line
IQ tests work by presenting carefully designed problems, often pattern puzzles that strip away language and knowledge, and then comparing your performance with a large reference group. Matrix reasoning is at the heart of many tests because it captures abstract reasoning so cleanly.
Want to see these puzzles from the inside? Try our free IQ test and work through a set of matrix-style questions yourself.
FAQ
Why do IQ tests use shapes instead of words?
Visual pattern puzzles rely less on language, schooling and cultural knowledge than word-based questions. That makes them a more direct way to assess abstract reasoning across a wide range of people.
Are IQ tests timed?
Many are, at least partly. Some subtests specifically measure processing speed, while others allow generous time limits so that reasoning, not speed, is the main factor.
Can practicing matrix puzzles improve my score?
Familiarity with the format usually raises scores somewhat on retests, a well-known practice effect. That gain mostly reflects comfort with the puzzles rather than a deep change in underlying ability.
