IQ Test Scores: What Your Number Actually Means

What does an IQ test score mean?

An IQ test score tells you roughly where you rank among people your age on the tasks that test uses. A 100 means you scored right in the middle; a 115 means you scored higher than about 84% of your age group; an 85 means higher than about 16%. For the full chart of ranges and percentiles, see the IQ scale.

The key word is roughly. A score is an estimate, not a reading off a dial. To use it well, you need to know four things: how much error it carries, what can push it up or down, which kind of test produced it, and what it leaves out.

Why is your IQ score really a range?

Because every test carries measurement error. If you could take the same test many times with no memory of the earlier attempts, your scores would scatter around a “true” value. The size of that scatter is called the standard error of measurement (SEM).

For full-scale IQ on major clinical tests like the Wechsler scales, the SEM is typically around 2 to 3 points, because those tests are long and highly reliable. Shorter tests, single subtests and index scores have larger errors.

Psychologists use the SEM to build a confidence interval, a range likely to contain your true score:

These are illustrative figures using a round SEM of 3. Real reports use the test manual’s own tables, which also adjust for a statistical effect called regression toward the mean, so the interval isn’t always perfectly centered on the score.

What this means in practice:

• Labels blur at the edges. A 109 (“average”) and a 111 (“high average”) are not meaningfully different results.
• Cutoffs need care. When a decision hinges on a line, such as 70 for intellectual disability or 130 for a gifted program, a responsible evaluator looks at the whole interval and at other evidence, not the single point.
• Small differences between people mean little. Two people 4 points apart could easily rank in the opposite order on another day.

What can push your score up or down?

Several things move a score without changing anything about your underlying ability:

1. Practice effects. Retaking the same or a similar test raises scores, sometimes by several points on the full scale and more on speed and novel-reasoning tasks. The effect is biggest when the retest comes soon after the first attempt, which is why clinicians usually wait a year or more and sometimes switch to a different test.
2. Outdated norms. Because of the Flynn effect (raw performance rose about 3 points per decade through much of the 20th century), a test normed long ago can produce inflated scores against today’s population. See average IQ for more on this.
3. State on the day. Sleep, illness, anxiety, medication and motivation all affect timed tasks.
4. Language and culture. Taking a test in a second language, or without familiarity with the testing format, can lower scores, especially on verbal sections.
5. Uneven profiles. If your index scores differ a lot (say, strong verbal comprehension and much weaker processing speed), the full-scale number averages over real differences and can describe you poorly. Clinicians often focus on the profile instead.

What do the index scores in a report mean?

A clinical report gives you more than one number. Alongside the full-scale IQ, modern Wechsler tests report index scores, each on the same mean 100, SD 15 scale:

• Verbal comprehension: word knowledge and verbal reasoning.
• Fluid or perceptual reasoning: solving new problems, often with visual patterns.
• Working memory: holding and manipulating information in mind.
• Processing speed: quick, accurate work on simple tasks.
• Visual-spatial (on some editions): mentally handling shapes and space.

Each index has its own, usually larger, margin of error than the full-scale score, because it rests on fewer items. Differences between indexes are common in the general population, so a gap only becomes clinically interesting when it’s large and fits other evidence. A psychologist interprets that pattern; a single number can’t.

Online IQ tests vs clinical IQ tests

The biggest single factor in what a number means is which kind of test produced it.

A few warning signs on online tests:

• Everyone scores high. If your friends all got 125 or above, the test is flattering, not measuring.
• No information about norms. A test that won’t say who it was calibrated against can’t tell you where you rank.
• Payment to see the score. The business model rewards a satisfying number.

The free Mentarium IQ practice test is deliberately honest about this. It gives you 30 original questions across the item types IQ tests use, with explanations, so you can practice and learn how the questions work. It isn’t normed like a clinical test and doesn’t give you a clinical IQ.

How stable is an IQ score?

In adults, IQ ranks are fairly stable over years: people who score high relative to their peers tend to keep scoring high. In young children, scores are much less stable, so an early-childhood result is a weak forecast of an adult score.

Stable doesn’t mean fixed. Individual scores retested years apart commonly move by several points, and schooling is one of the better-documented influences on test performance. Large changes in adulthood usually point to something specific: a brain injury, an illness, or, often, a difference in the tests used.

What should you do with your score?

Treat it as one piece of information with an error bar, and use it for what it’s good at.

• If it came from a clinical assessment, read the whole report, not the headline number. The index profile and the examiner’s notes usually say more than the full-scale IQ.
• If it came from an online test, don’t build anything on it. It may be in the right neighborhood or nowhere near.
• If you’re comparing yourself with others, stop. Small gaps are noise, and big ones say little about how a life will go.
• If you want to think better, a score won’t help. Skills will. The reasoning habits in critical thinking and the mental models library can be practiced; your IQ mostly can’t.

The Mentarium take

The most important thing to know about your IQ score is that it’s a range pretending to be a point. Once you add the error bar, most of the drama goes away: 112 versus 117 is not a story, and an online 140 is a story about the website, not about you.

People tend to grip a precise-looking number harder than it deserves, which is overconfidence applied to measurement. The fix is to remember that the map is not the territory. An IQ score is a useful map of a narrow slice of your thinking, drawn with a known margin of error. Read it that way and it’s helpful. Read it as your identity and it’s misleading.