How Is IQ Calculated? The IQ Formula Explained

How was the original ratio IQ calculated?

Early Binet-derived scoring used the idea of mental age. An examiner estimated the age level at which a child’s pattern of answers was typical, then used:

Ratio IQ = (mental age ÷ chronological age) × 100

For illustration, a 10-year-old performing like the average 12-year-old would receive 120 under that formula. William Stern helped popularize the quotient terminology, and the Stanford-Binet made ratio IQ familiar in the United States.

The calculation has an obvious problem: mental age does not continue upward in a simple linear way for adults. A 40-year-old cannot sensibly receive a mental age of 48 from the same child scale and then be compared with an eight-year-old’s ratio. Ratio IQ also made the same number represent different levels of rarity at different ages. The name survived; the formula did not.

Related

The History of IQ Testing: From Binet to Modern Scales

A clear timeline of IQ testing, from Galton and the Binet-Simon scale to Stanford-Binet, Wechsler tests, and modern computer-based assessments.

How does a professional test turn answers into IQ?

The process has several stages, not one magic equation.

1. Administer standardized tasks. The examiner follows exact instructions and timing for verbal, visual-spatial, fluid-reasoning, working-memory, and processing-speed subtests, depending on the battery.
2. Score the raw responses. Correct answers, completion time, and sometimes partial-credit rules are converted into subtest raw scores.
3. Apply age-based norms. The raw score is matched to people in the examinee’s age band from the publisher’s representative norm sample.
4. Convert to scaled scores. Subtests may use a mean of 10 and standard deviation of 3, while index scores and Full-Scale IQ commonly use a mean of 100 and standard deviation of 15.
5. Combine correlated subtests. The publisher’s scoring model weights the required subtests to produce index scores and, when valid, a Full-Scale IQ.
6. Report uncertainty and interpretation. A qualified examiner reports a confidence interval, percentile, and whether the composite is an appropriate summary of the profile.

This is why copying a raw score from a different test into a generic IQ chart is invalid. The item difficulty, age norms, reliability, and composite weights belong to the named instrument.

Why are age norms necessary?

Cognitive performance changes with development and aging, and different tasks have different age patterns. A raw total of 40 may be typical for one age and unusually high or low for another. Norming creates a fair comparison by asking, “How did this person perform relative to peers at the same age when the norms were collected?”

Publishers periodically refresh norms because populations, education, language exposure, and familiarity with testing change. A score from an old edition is not automatically equivalent to a score from a current edition unless the publisher has provided an appropriate equating or linking procedure.

Age correction does not remove every influence. Language history, hearing or vision, motor speed, health, sleep, anxiety, and the testing environment can affect performance. The norm answers one comparison question; it does not erase the conditions under which the answers were produced.

How are percentiles derived from IQ?

If the norm distribution is approximately normal, the z-score can be translated into a percentile using the cumulative distribution. For example, IQ 115 on a 100/15 scale is z = +1 and falls near the 84th percentile; IQ 130 is z = +2 and falls near the 98th percentile. The exact percentile may differ slightly because published norms are empirical and may use age-specific tables rather than a perfect bell curve.

Percentiles are often easier to understand than raw standard deviations, but they become uneven near the extremes. A two-point change around the mean may move only a few percentile points, while the same two-point change near 130 can move a much larger fraction of the small high-score group. That is one reason precise claims about tiny score differences are unwarranted.

What is the difference between an IQ formula and an online estimate?

An online estimate may use a short set of questions and a simple lookup table. That can be useful for curiosity, but it usually lacks the publisher’s representative norm sample, controlled administration, secure items, and reliability evidence. Some sites report an “IQ” after converting a percentage correct with an invented multiplier; that number is not a clinical deviation IQ.

Our own assessment is transparent about scope. The free attempt uses 30 questions across four cognitive areas, and the detailed result report is paid after the attempt. It can provide a quick benchmark, but it is not a diagnostic evaluation, a substitute for WAIS or Stanford-Binet administration, or official evidence for Mensa.

What does measurement error do to an IQ score?

Every test score is an estimate. Reliability research allows publishers to calculate a standard error of measurement (SEM), which describes the expected spread of observed scores around a person’s underlying level. A report may therefore show a 95% confidence interval rather than implying that 118 and 119 are meaningfully different.

The interval depends on the test’s reliability and the score scale. If a score is 115 with an SEM of 3, a simple illustrative interval might be about 109–121 at roughly 95% confidence. The actual calculation and confidence level should come from the instrument’s manual. Repeating a test too soon can also introduce practice effects, so a retest is not a guaranteed way to find a “truer” number.

Related

IQ Range and Scale for Adults

For adults, modern IQ scales center at 100 with a 15-point standard deviation: 85–115 is the broad average range, 70–84 is below average, and 116–130 is above average.

Can you calculate IQ from a percentage score?

Not responsibly without the test’s norming information. A percentage correct is a raw performance measure. To convert it, you would need the test’s item-level difficulty, age-specific reference distribution, scoring rules, and evidence that the sample represents the population you want to compare with. Even then, a composite score may use several subtests rather than one percentage.

The practical rule is simple: interpret a score only with the name and edition of the test, the age norms, the scale’s standard deviation, and the report’s confidence interval. Never compare a “130” from an unknown quiz directly with a 130 from a supervised clinical battery.

Frequently Asked Questions

Q: What is the modern IQ formula?

A: Modern IQ is usually a deviation score, often represented as IQ = 100 + 15 × z. The z-score expresses distance from the mean of the correct age-based norm group; publishers use detailed norm tables to implement the conversion.

Q: Is IQ calculated as a percentage correct?

A: No. Percent correct is a raw score, while IQ is a norm-referenced standard score that also depends on age, item difficulty, the test’s norms, and its composite scoring model.

Q: What was the old ratio IQ formula?

A: Ratio IQ was mental age divided by chronological age, multiplied by 100. It was useful for early child testing but became unsuitable for adults and was replaced by deviation IQ.

Q: Does IQ 130 mean 98 percent of answers were correct?

A: No. On a 100/15 scale, 130 is about two standard deviations above the norm mean and near the 98th percentile, which describes relative rank rather than answer accuracy.

Q: Why can two tests give different IQ scores?

A: They may use different tasks, norms, reliability, timing, and composite weights. Compare scores only within the same instrument and context, and use the confidence interval rather than treating one point as exact.

References

• APA Dictionary of Psychology: deviation IQ
• APA Dictionary of Psychology: IQ
• Z scores, standard scores, and composite test scores (PMC)
• Clinical assessment of brain disorders: standard scores and IQ (NCBI)

Last updated: July 19, 2026