Average IQ Score Explained: What 100 Means

Try the free IQ-style visual reasoning test.

By MindScan Team · October 12, 2026

People search for average IQ for a simple reason: every free quiz and news headline seems to throw a number at you, and it is hard to know what that number actually means. This article explains what an average IQ score is on modern scales, why 100 is the conventional mean, how the normal distribution and a standard deviation of 15 work, and why different tests (and online quizzes) produce different figures.

This is educational reading, not clinical advice. MindScan’s own IQ-style reasoning test gives an estimate for insight. It is not a certified IQ score, not population-normed in the professional sense, and not a diagnosis. Informed by published research on cognition and neurodiversity, for educational insight, not diagnosis.

What does an average IQ mean?

On modern intelligence tests, an IQ of 100 means that the person’s performance matched the average (mean) of the age-matched norming sample used to standardise that test. It is not a percentage, a mark out of 100, or a claim that someone is “exactly 50% intelligent”. It is a relative score: how you did compared with the people who formed that test’s norms.

The American Psychological Association’s dictionary puts it plainly: the mean IQ is customarily 100, with slightly more than two thirds of all scores falling within plus or minus 15 points of the mean (usually one standard deviation). More than 95% of scores fall between 70 and 130 (APA Dictionary: IQ).

So when people say “average IQ”, they usually mean either:

Exactly 100 — the constructed mean of the scale, or
Roughly the middle band — commonly scores around 85–115 (within one standard deviation of the mean on a Wechsler-style scale).

Both uses are common. What matters is knowing which scale and which norms produced the number.

Why is the average IQ 100? (Deviation IQ, not a quotient)

Early “intelligence quotients” were literal ratios: mental age divided by chronological age, multiplied by 100. That worked poorly for adults and did not keep the same meaning across ages.

Modern scores are deviation IQs. The APA defines a deviation IQ as a standard score with a mean of 100 and a standard deviation specific to the test — usually 15 or 16 for intelligence tests (APA Dictionary: deviation IQ). The score reports how far a person sits from the mean of their comparison group, not a mental-age ratio.

Many widely used batteries — including Wechsler scales such as the WAIS–IV UK — use a Full Scale IQ (and index scores) standardised to a mean of 100 and a standard deviation of 15. Wikipedia’s overview of the intelligence quotient summarises the same convention: modern raw scores are transformed to a normal distribution with mean 100 and SD 15, so roughly two thirds of people score between 85 and 115, and about 2% each above 130 and below 70 (Intelligence quotient, Wikipedia).

100 is average by design. Publishers calibrate the scale so the mean of the norming sample is 100. If norms are updated and today’s sample scores higher on the old items, the new edition is re-centred so that “average” is still 100 on the new norms. That recalibration is one reason historical raw performance and today’s labelled IQ scores are not the same thing (see the Flynn effect below).

The normal curve and SD 15 in plain language

Picture a bell-shaped curve centred on 100. Most people cluster near the middle; fewer people sit far out on either side. With mean 100 and SD 15:

Distance from mean	Approximate IQ band	About how many people (normal curve)
Within 1 SD	85–115	Roughly 68%
Within 2 SD	70–130	Roughly 95%
Within 3 SD	55–145	Roughly 99.7%

Those percentages are properties of the normal distribution, which test publishers aim for when they scale scores. They are not mysterious properties of “intelligence itself”. APA’s dictionary figures line up with this picture: a bit more than two thirds within ±15 of the mean, and more than 95% between 70 and 130 (APA Dictionary: IQ).

A few practical implications:

Moving from 100 to 115 is one standard deviation above the mean — a clear step, but still within a common range.
Moving from 100 to 130 is two standard deviations above — much less common on a well-normed scale.
Small differences (for example 98 vs 102) are usually not meaningful on their own, especially when measurement error and different tests are taken into account.

IQ points are also not percentage points. Scoring 50 does not mean “half as able” as scoring 100. The scale is built for ranking relative to norms, not for counting units of ability like centimetres.

Why different tests (and online quizzes) give different numbers

Even among serious instruments, the same person can get different scores on different batteries, or on the same battery on different days. Reasons include:

Different content. Verbal comprehension, matrix reasoning, working memory and processing speed are related but not identical. A verbally heavy set and a non-verbal matrix set will not line up perfectly.
Different norms. A score of 100 means average for that test’s standardisation sample and edition. Older norms, different countries’ editions, and different age bands change what “average” maps onto in raw performance.
Measurement error. Professional manuals report confidence intervals around composite scores. A single point estimate is never as precise as a headline suggests.
Motivation, fatigue, anxiety and practice. Performance can shift session to session without anyone’s underlying ability flipping overnight.
Online quizzes that are not IQ tests. Many free web tools are entertainment or marketing. They may:
- use short, unstandardised item sets;
- lack a proper norming sample;
- mix speed, personality or trivia with reasoning;
- convert a raw count into a “score” that merely looks like a Wechsler IQ.

British Mensa is explicit that its free Online Workout is for entertainment, is not an IQ test, and will not qualify anyone for membership. That honesty is a useful benchmark for any free tool that flashes a big number.

Research item banks such as the International Cognitive Ability Resource (ICAR) are valuable for science because they are public-domain and documented (Condon & Revelle, 2014). Using ICAR-style item types still does not automatically make a website’s output a certified Full Scale IQ. Norming, administration conditions and psychometrics matter.

A brief note on the Flynn effect

Across much of the twentieth century, raw scores on intelligence tests rose from generation to generation — the Flynn effect. When tests are re-normed, the average labelled IQ is reset to 100, so the rise shows up as people scoring higher on older editions than on newer ones.

A meta-analysis by Trahan and colleagues estimated about 2.3 standard-score points per decade overall, and closer to 3 points per decade for modern Stanford–Binet and Wechsler comparisons (Trahan et al., 2014). A broader century-scale meta-analysis by Pietschnig and Voracek likewise documented global gains that vary by domain and have slowed in more recent decades (Pietschnig & Voracek, 2015). Wikipedia’s summary figure of roughly three IQ points per decade on average is in the same ballpark (Intelligence quotient, Wikipedia).

Why it matters for “average IQ”:

Average is relative to norms. “100” on a 1990s edition and “100” on a current edition are not identical statements about raw task performance.
Gains do not prove that every generation is simply “smarter” in every everyday sense; they show that performance on these tasks has shifted, for environmental and other reasons that researchers still debate.
Some countries show slowing or reversal in recent cohorts. That is another reason league-table style comparisons of “national IQ” are a poor use of the concept — and why this article does not publish country rankings.

What a free online “IQ” score can and can’t tell you

It can (sometimes):

give a rough sense of how you find certain puzzle types (matrices, series, spatial rotations);
be a low-stakes way to practise reasoning under mild time pressure;
spark curiosity about how professional assessments work.

It cannot, if it is not a supervised, properly normed instrument:

replace a Full Scale IQ from a test such as the WAIS–IV UK administered by a qualified practitioner;
diagnose intellectual disability, giftedness, learning difficulty or any clinical condition;
support high-stakes decisions (education placement, employment disputes, clinical diagnosis, Mensa membership) on its own;
be treated as precise to the nearest point.

If you want an official route in the UK for high-ability membership testing, British Mensa describes supervised options (in person or proctored online) and states that membership requires a score in the top 2%, via their supervised tests or accepted prior evidence (Test Your IQ – British Mensa). For wider cognitive or clinical assessment, look for an HCPC-registered practitioner psychologist (HCPC Check the Register) and the BPS Find a psychologist directories. Speak to your GP when health or learning concerns are involved.

Try MindScan’s IQ-style reasoning estimate (honest disclaimer)

If you want a structured, low-friction practice run after reading this, try MindScan’s free IQ-style reasoning test. You will get an instant estimated reasoning score for educational insight. It is not a certified IQ score, not a clinical assessment, and not population-normed like a Wechsler Full Scale IQ.

Prefer unscored practice first? Work through our IQ-style sample questions with answers. For a broader cognitive self-assessment (traits and strengths, still educational and not a diagnosis), see the free 20-question test or MindScan Arcade.

FAQ

What is the average IQ score?

On modern deviation-IQ scales, the average (mean) is set at 100. Roughly two thirds of people score between about 85 and 115 when the standard deviation is 15 (APA Dictionary: IQ; Intelligence quotient, Wikipedia).

What does an IQ of 100 mean?

It means performance at the mean of the age-matched norming sample for that test edition — “average” on that scale, not a universal percentage of human ability.

Is 90 or 110 still “average”?

Publishers use different label bands, but both 90 and 110 sit near the centre of a mean-100, SD-15 scale. Many people treat the wider 85–115 band as the everyday “average range”. Exact band names belong in a professional report, not as casual self-labels from a free quiz.

Why did my online score disagree with another site?

Different items, different (or missing) norms, different scoring formulas, and measurement error. Treat free numbers as entertainment or rough practice feedback unless the provider documents a proper standardisation sample and administration rules.

When should I seek a professional assessment?

When decisions or concerns are serious — suspected learning difficulty, cognitive change, educational planning that requires formal evidence, or interest in supervised high-ability testing. In the UK, start with your GP where appropriate, verify practitioners on the HCPC Register, and use the BPS Find a psychologist tools. For Mensa membership pathways, see British Mensa.

Bottom line

Average IQ is not a mystery number invented by websites. On contemporary scales it is 100 by construction, with most people falling within about 15 points either side on a normal curve. The useful questions are always: which test, which norms, how was it administered, and what decision (if any) are you trying to make? Free online tools can be interesting practice. Certified scores and clinical conclusions need qualified professionals and properly normed instruments.

Sources

Start Free Assessment