IQ Score Distribution and Percentiles: How to Read a Score on the Mean-100, SD-15 Scale

An IQ score is a standard score built to sit on a normal distribution with a mean of 100 and a standard deviation of 15. That is why 100 marks the middle of your age group (50th percentile), 115 sits at roughly the top 16 percent, and 130 at roughly the top 2 percent. This guide covers how to read that distribution, what online tests can and cannot tell you, and what a Soundary score means.

Why IQ has a mean of 100 and an SD of 15

A raw score, the number of items you got right, means nothing by itself. The test is first given to a representative sample of your age group (the norm group), and your position in that distribution is converted to a standard score on the normal curve. Modern tests use this deviation IQ instead of the early Stanford-Binet ratio of mental age to chronological age, and the Wechsler scales made mean 100, SD 15 the convention.

On a normal curve, about 68 percent of people fall within one standard deviation of the mean and about 95 percent within two. In IQ terms, roughly two-thirds score between 85 and 115, and about 19 in 20 between 70 and 130. Scores of 130 and above, or below 70, are therefore rare, each covering around 2 percent of the population.

Percentiles by score band

The table below pairs the descriptive categories commonly used with Wechsler-type tests with approximate percentiles under a normal distribution. A percentile is the share of people who scored below you. The numbers will not match a test manual to the decimal, so treat them as a rough guide.

Two things are worth remembering. First, the same 15-point gap moves the percentile a lot in the crowded middle of the curve and very little at the tails. Second, every standardized test has a standard error of measurement, so your true ability is a range, not a point. Even a well-built full-scale test carries a 95 percent confidence interval of roughly plus or minus 5 points, which means a 112 and a 117 are hard to tell apart statistically.

What an online IQ test can and cannot tell you

The biggest limitation of an online test is the norm. A mean of 100 only means something when the test was standardized on a representative sample, but people who seek out a test on the web skew in age, education, and motivation; the items are public; and time and setting are uncontrolled. Most online IQ figures are therefore closer to your position among that site's users than to your IQ in the population at large.

That does not make it useless. Figural reasoning items like Raven's matrices depend little on language, compare relatively fairly across cultures, and correlate strongly with fluid intelligence. A test built from items like these, repeated under the same conditions, is perfectly good for seeing your pattern of stronger and weaker areas and how you vary with your condition. For decisions at school, at work, or in health care, though, an intelligence test has to be an individually administered one given by a trained examiner under standardized procedures.

The Flynn effect: norms go stale

Across the twentieth century, raw scores on intelligence tests rose steadily in many countries, by roughly 2 to 3 points per decade. This is the Flynn effect (Flynn, 1987; Trahan and colleagues, 2014). Nutrition, schooling, and growing familiarity with abstract thinking are among the proposed causes, with no single settled answer. The practical implication is that stale norms inflate scores: scored against norms from twenty years ago, the same person can come out several points higher than they should, which is why standardized tests are periodically renormed.

How the Soundary cognitive battery is built and scored

In the Korean edition the battery has five sections: verbal reasoning (12 items, 5 minutes), numerical reasoning (12 items, 7 minutes), 3-by-3 matrix reasoning (10 items, 7 minutes), forward and backward digit recall (working memory), and a 60-second symbol comparison (processing speed). Other language editions drop the verbal section and run four. The composite is a weighted average on a 0 to 100 scale: in Korean, verbal, numerical, and matrix count 25 percent each and memory and speed 12.5 percent each; in other editions, numerical and matrix count one-third each and memory and speed one-sixth each.

The IQ conversion on the result screen is a reference value that maps this composite onto the mean-100, SD-15 scale. Because the standardization sample is still being collected, it currently uses provisional norms and is designed to switch automatically to the real distribution of users once enough results have accumulated for the test. Read it as an estimate of your relative position within this test expressed on a familiar scale, not as a value equivalent to the IQ from a standardized test such as the Wechsler.

• 🧩Cognitive ability (IQ)Verbal, numeric, figural, memory and speed in one battery.Take the test →
• 🔄Mental rotationSame shape or mirrored? Rotate it in your head.Take the test →
• 🃏Rule discoveryCrack the hidden sorting rule as it shifts.Take the test →

Why the same person gets different scores

Even on standardized tests, retest scores drift by a few points. Measurement error, your condition (sleep, fatigue, anxiety), practice effects from growing familiar with the item format, and different strategies under time pressure all mix together. Online tests add device and setting variables on top, so they swing more. That is why the average of several runs under similar conditions, together with your profile across domains, is far more useful than pinning a single score down as your IQ.

Relative standing in intelligence is fairly stable through adulthood, but processing speed and working memory are sensitive to how you are doing that day. If your composite dropped a few points, the sensible first guess is that the speed or memory section wobbled, not that your intelligence fell. Pairing it with the multiple intelligences questionnaire, a 24-item self-report on eight intelligence tendencies, also lets you compare measured ability with what you prefer.

• 🌟Multiple Intelligences Test — Your Map of 8 StrengthsMaps which of eight intelligences — linguistic, logical, spatial, musical, bodily, interpersonal, intrapersonal, naturalist — you use most naturally, and which to grow, in 24 questions. A self-report written by Soundary with reference to Howard Gardner's theory, not a test of ability; the result also points to cognitive tasks that measure each area for real.Take the test →

In short, an IQ is your relative position in a norm group expressed on a mean-100, SD-15 scale, and it is only as accurate as the norms and the testing conditions behind it. Used with its limits in mind, an online test is a good tool for seeing your profile and your trend; when a definitive judgment is needed, a standardized individual test is the answer.