Esprio
Get Esprio

Online IQ tests: how accurate are they, and what does IQ mean?

5 min read

Search for “IQ test” and you will find hundreds of free quizzes that give you a precise number in ten minutes. Some will tell you that you scored 138. It is tempting to take that number seriously. Before you do, it helps to understand what an IQ score is and how a real one is produced.

What an IQ score actually is

An IQ score is not a measure of raw brainpower like a weight on a scale. It is a rank. It tells you how your performance on a set of tasks compares with a large reference group of people your age.

Modern tests are built so that:

  • the average score in the reference group is 100;
  • the standard deviation is 15 points.

Because scores follow a bell curve, this gives a simple picture:

  • about 68% of people score between 85 and 115;
  • about 95% score between 70 and 130;
  • only about 2% score above 130.

So a score of 115 means “higher than about 84% of the reference group”, nothing more mystical than that. And a score of 145 from a ten-minute quiz is a claim that you outperform more than 99.8% of people. That should make you skeptical of the quiz, not proud.

How a real test is built: norming

The reference group is the key ingredient. Test publishers give the test to a large sample of people chosen to reflect the population by age, sex, education, region and other factors. This is called norming. Your raw answers are then converted into a score by comparing them with people in that sample who are your age.

Without a proper norming sample, a test has nothing reliable to compare you against. A raw score of “24 out of 30” means nothing until you know how comparable people did.

What tests like the WAIS measure

The best-known adult test is the Wechsler Adult Intelligence Scale (WAIS). Its fifth edition, the WAIS-5, was published by Pearson in 2024 for people aged 16 to 90. It is administered one-to-one by a trained professional, usually a psychologist, and the core subtests take about an hour.

It does not produce only one number. Alongside a Full Scale IQ, it reports separate index scores, including:

  • Verbal Comprehension: vocabulary, knowledge, verbal reasoning;
  • Visual Spatial: understanding and building visual patterns;
  • Fluid Reasoning: solving new problems and spotting rules;
  • Working Memory: holding and manipulating information;
  • Processing Speed: doing simple tasks quickly and accurately.

A professional also looks at the gaps between these scores, at how you approached the tasks, and at the margin of error. Reports usually give a range rather than a single exact figure, because no test is perfectly precise.

These tests are used for specific purposes, such as diagnosing learning difficulties or assessing cognitive changes after an injury. They are clinical tools, which is why access to them is restricted.

Why free online tests are rough estimates

Online tests can be fun and can give you a vague idea. But several things limit their accuracy.

1. Unknown or missing norms

Most free tests do not say who they were normed on, or whether they were normed at all. Some compare you only with other people who took the same quiz online. That group is self-selected: people who search for IQ tests are not a random slice of the population.

2. A narrow set of tasks

Many online tests use only one type of puzzle, often pattern matrices similar to Raven’s Progressive Matrices. These measure one aspect of reasoning well, but they do not cover vocabulary, memory or speed. One kind of puzzle gives a partial picture.

3. No controlled conditions

A clinical test happens in a quiet room, with standard instructions and timing. At home, you might be tired, interrupted, on your phone, or have seen similar puzzles before. Practice alone can raise scores on puzzle-type tests.

4. Business incentives

Some sites give flattering results, then ask you to pay for a “certificate” or a detailed report. A high score keeps people happy and paying. That is not a scientific design goal.

The Flynn effect: why norms go out of date

In the 20th century, average raw scores on IQ tests rose steadily in many countries, roughly three points per decade on tests like the Wechsler scales. This is called the Flynn effect, after the researcher James Flynn who documented it.

It means that norms age. If a test was normed decades ago, today’s test takers will look smarter than they are compared with the original sample. This is one reason serious tests are renormed regularly. Interestingly, the trend has slowed or reversed in some countries. A 2018 study of Norwegian military conscription data by Bernt Bratsberg and Ole Rogeberg found declining scores among men born after the mid-1970s.

How to read an online result

If you take an online test, here is a healthy way to treat the number:

  • See it as a rough estimate with a wide margin, not a fact about you.
  • Look at the type of tasks. A test with several kinds of puzzles tells you more than one with a single kind.
  • Don’t pay for certificates. They have no official value.
  • If you need a real assessment, for school, work or health reasons, see a qualified psychologist.

A test for fun

The IQ test in Esprio is built in that spirit. It takes about 15 minutes, mixes different kinds of questions, and gives you an estimate to satisfy your curiosity. It is not a clinical assessment, and it is not meant to replace one.

Also keep in mind that IQ is only one way of describing how people think. It says nothing about creativity, kindness, motivation or most of the things that shape a life. If you are curious about what training can and cannot change, our article on whether brain training games work looks at the research.

More guides