IQ Scores Explained: What They Mean

What an IQ Score Actually Is

An IQ score, short for intelligence quotient, is a standardized index of cognitive ability. It is not a raw count of correct answers and it is not a percentage. Instead, it is a relative measure: it expresses how your performance compares to the performance of other people in a carefully selected reference group, usually matched to your age.

Modern IQ scores are deviation scores. The scoring system takes your raw test performance, compares it against the distribution of scores in the norming sample, and converts it into a single number anchored to a fixed scale. On that scale, the population average is defined as 100. This is why an IQ of 100 always means "exactly average for your age group," regardless of which validated test you took or what year it was taken.

It is important to separate two ideas from the start. The first is the test performance itself, which reflects skills like reasoning, working memory, verbal comprehension, and processing speed. The second is the score, which is a statistical translation of that performance into a common language so results can be compared across people, tests, and time.

The Mean-100, SD-15 Framework

Two numbers define the modern IQ scale: the mean and the standard deviation. The mean is set to 100. The standard deviation, abbreviated SD, is set to 15 on most widely used tests such as the Wechsler scales and current versions of the Stanford-Binet.

Standard deviation measures how spread out scores are around the average. Because IQ uses an SD of 15, every 15 points represents one standard deviation away from the mean. A score of 115 is one SD above average. A score of 85 is one SD below. A score of 130 is two SD above, and 70 is two SD below.

This framework matters because it makes any IQ number instantly interpretable. Instead of asking "is 112 a lot," you can ask how many standard deviations it sits from the center, which then maps directly to a percentile. Note that not all tests historically used an SD of 15. A few older instruments, such as the original Cattell scale, used an SD of 16 or 24, which is why the same raw ability can produce a slightly different headline number on different tests. Always check which scale a score is reported on before comparing two numbers.

The Bell Curve and What It Predicts

IQ scores are designed to follow a normal distribution, commonly called the bell curve. In a normal distribution, most people cluster near the center and fewer people appear at the extremes, with the curve symmetric around the mean.

With a mean of 100 and an SD of 15, the bell curve produces predictable proportions of the population at each band:

• About 68 percent of people score between 85 and 115, within one SD of the mean.
• About 95 percent score between 70 and 130, within two SD.
• About 99.7 percent score between 55 and 145, within three SD.

This is why scores far from 100 are rare by design. Roughly 1 person in 44 reaches 130 or above, and a similar fraction falls at 70 or below. The further a score sits from 100, the fewer people share it, which is exactly what makes high and low scores statistically meaningful rather than arbitrary.

How IQ Scores Are Calculated and Normed

Producing an IQ score happens in stages. First, a test taker completes a battery of subtests measuring different cognitive domains, and each correct or scored response contributes to a raw score. Raw scores alone are meaningless across tests, because tests differ in length and difficulty.

To give raw scores meaning, test publishers run a process called norming. They administer the test to a large, representative standardization sample chosen to mirror the broader population by age, sex, region, and education. The distribution of raw scores in that sample becomes the reference against which everyone else is measured.

Each raw score is then converted into a standardized score. Conceptually, the test calculates how many standard deviations your raw performance sits from the sample mean, then rescales that to the IQ metric with mean 100 and SD 15. Because tests are age-normed, your score reflects how you did against people your own age, not against the entire population at once. This is why a 9-year-old and a 40-year-old can both score 110 while answering completely different questions.

Norms are not permanent. Publishers re-norm tests every decade or so to keep 100 anchored to current average performance, which connects directly to the Flynn effect discussed below.

IQ Score Ranges and Classifications

Psychologists group IQ scores into descriptive bands. These labels are conventions, not hard biological categories, and the exact cutoffs vary slightly between test manuals. The table below uses the common mean-100, SD-15 framework and lists approximate percentiles.

A few points keep these bands honest. The boundaries are not magic thresholds: a 129 and a 131 reflect essentially identical ability, even though they fall in different rows. Classification labels describe test performance, not a person's worth, creativity, or potential. And clinical diagnoses, such as intellectual disability, are never made on an IQ number alone; they also require evidence of impaired adaptive functioning in daily life.

Percentiles: The Clearest Way to Read a Score

A percentile tells you the percentage of people who scored at or below a given IQ. It is often the most intuitive way to communicate a result because it answers the real question: how do I compare to everyone else?

Useful anchor points on the mean-100, SD-15 scale:

• IQ 100 is the 50th percentile. Half score below, half above.
• IQ 110 is roughly the 75th percentile.
• IQ 115 is roughly the 84th percentile.
• IQ 120 is roughly the 91st percentile.
• IQ 130 is roughly the 98th percentile.
• IQ 85 is roughly the 16th percentile.

Percentiles also correct a common misreading. Being in the 90th percentile does not mean you are "90 percent intelligent." It means you outperformed about 90 out of 100 test takers. Percentiles are rankings, not quantities of intelligence, and they compress unevenly: the gap in ability between the 50th and 60th percentile is smaller in score points than the gap between the 97th and 99th, because the curve thins out at the edges.

What Counts as a Good, Average, or High IQ

Average IQ is, by definition, the range around 100. Scores from 90 to 109 are classified as average and describe roughly half of all people. There is nothing deficient about an average score; it simply means typical cognitive performance for your age.

A "good" IQ score is usually understood as one above the average band. Scores from 110 upward are high average to superior and indicate stronger measured reasoning and problem-solving relative to peers. There is no official cutoff for "good," because the answer depends on context, such as academic demands or a specific role.

A high IQ generally refers to the superior and gifted ranges. Scores of 120 and above are commonly called superior, and 130 and above is the conventional threshold for giftedness and the entry point for many high-IQ societies, placing a person in roughly the top 2 percent. Scores above 145 are sometimes labeled highly or profoundly gifted and become statistically very rare.

The honest caveat: a higher number correlates with certain academic and analytical outcomes, but it does not guarantee success, wisdom, or happiness. IQ captures one slice of cognition and leaves out creativity, emotional intelligence, motivation, and practical skill.

Reliability, Confidence Intervals, and the Flynn Effect

No IQ score is a single exact point. Every test has measurement error, so results are most accurately reported as a confidence interval. A well-constructed test might report a score of 118 with a 95 percent confidence interval of roughly 112 to 124, meaning the true score very likely falls within that band. This is why reading too much into a few points, or treating a category boundary as decisive, is a mistake.

Reliability describes how consistently a test produces the same result for the same person over time. Validity describes whether it actually measures the cognitive ability it claims to. Professionally developed tests are engineered for high reliability and validity, but no instrument is perfect, and factors like fatigue, anxiety, illness, or unfamiliarity with the format can move a score on any given day.

The Flynn effect adds a longer-term wrinkle. Average raw performance on IQ tests rose substantially across the twentieth century in many countries, likely driven by better nutrition, education, and environmental complexity. Because 100 is re-anchored to the current population each time a test is re-normed, this rise is invisible in your headline number, but it is the reason an outdated test can inflate scores: it is comparing you to people from an earlier, lower-scoring norm group.

Online Scores vs Clinical Scores

Not all IQ scores are produced the same way, and the source of a score determines how much weight it deserves.

A clinical IQ score comes from a comprehensive, individually administered test such as the Wechsler Adult Intelligence Scale or the Stanford-Binet, given one-on-one by a trained professional under standardized conditions. These tests are normed on large samples, cover multiple cognitive domains, and report confidence intervals. They are the appropriate tool for diagnosis, school placement, and any high-stakes decision.

Online IQ tests vary enormously. A well-designed online assessment can give a reasonable, useful estimate of where you fall on the curve and is an excellent low-pressure way to explore your reasoning skills. But online tests are typically shorter, self-administered, and normed less rigorously, so they should be read as approximations rather than official, clinical results.

Treat an online score as a directional snapshot. If you want a precise, defensible number for a formal purpose, that requires a proctored, professionally administered test. For curiosity, self-tracking, and practice, a solid online test is a sensible and convenient starting point.

Frequently asked questions