Most online writing about IQ falls into one of two camps. One side treats the score as destiny: a fixed verdict about your worth, your earning potential, your future. The other side dismisses it as junk science, a relic of early 20th century pseudoscience that should be retired.
Both are wrong. The truth is more interesting, and a lot more useful.
This guide is what we wish someone had handed us before we ever sat down to take an IQ test. It walks through what the number actually measures, what it reliably predicts, what it categorically does not, and how to interpret the score you get with the precision it deserves.
The 100-point scale, explained
The number 100 isn’t magic. It’s a calibration choice.
When IQ tests were standardized in the early 20th century, psychometricians made a design decision: fix the population mean at 100, with a standard deviation of 15 points. Whatever the test, however many questions it contains, however the scoring rubric works under the hood, the final score is mapped onto that same scale. Mean of 100. Standard deviation of 15.
Everything else flows from that.
This means an IQ of 115 isn’t a fixed quantity of "smartness." It means you scored one standard deviation above the population average for your age group. An IQ of 130 means two standard deviations above average, putting you in roughly the top 2.5 percent of test-takers. The score is fundamentally a position on a bell curve, not a measurement on an absolute scale.
An IQ score is a relative measurement. You aren’t being compared against an absolute standard of intelligence. You’re being compared against a normed sample of other people who took the same test.
What IQ tests actually measure
A modern IQ test isn’t measuring a single property called "intelligence." It samples performance across several distinct cognitive domains, then aggregates them into a composite score.
The five domains assessed by most contemporary tests:
- Logical reasoning. Drawing valid conclusions from given premises. Conditional logic, syllogisms, deduction.
- Pattern recognition. Identifying the rule that governs a sequence or matrix of shapes. The Raven’s Progressive Matrices test is the canonical format.
- Working memory. Holding information in mind across multiple steps without writing anything down.
- Spatial reasoning. Mental rotation, paper folding, visualizing 3D objects from 2D diagrams.
- Verbal and numerical fluency. Word analogies, antonyms, number sequences, arithmetic patterns.
Charles Spearman’s 1904 paper documented something that has been replicated for more than a century: scores on these distinct domains correlate with each other. Score well on one, you tend to score well on the others. Spearman called the underlying common factor g, for general intelligence.
Whether g is a single biological reality or a statistical artifact of how the tests are constructed remains contested. What is not contested is the empirical pattern: cognitive abilities cluster.
Score ranges and what they mean
The categories below are a common interpretive convention, not a natural law. The cuts are arbitrary points on a continuous distribution. Someone scoring 114 and someone scoring 115 differ by less than measurement error.
| Score range | Conventional label | Population |
|---|---|---|
| Below 70 | Below average | ~2.5% |
| 70 to 84 | Low average | ~13.5% |
| 85 to 114 | Average | ~68% |
| 115 to 129 | High average | ~13.5% |
| 130 and above | Gifted range | ~2.5% |
Two takeaways. First, most people score in the average band, by definition. The labels can sound dramatic, but the distribution is mathematical: 68 percent of all test-takers fall within one standard deviation of the mean. That isn’t a finding; it’s how the test was designed. Second, the boundaries between bands are interpretive. They’re convenient buckets, not biological transitions.
The five-band table above is a simplification. The clinical Wechsler scheme used in most modern assessments has seven bands with sharper labels (Very Superior, Superior, High Average, etc.) — if you want the full mapping, see our IQ classification ranges guide. For a direct answer to whether your specific score is "good" by various practical standards, see what is a good IQ score?.
What your score predicts
The honest answer: less than people think, and more than skeptics admit.
Income
Strenze’s 2007 meta-analysis of 35 longitudinal studies found the IQ-income correlation is approximately r = 0.30. That is real but modest. It means cognitive ability accounts for roughly 9 percent of the variance in income. Conscientiousness, the personality trait sometimes called discipline, predicts income at a comparable level. So does years of education.
An r of 0.30 sounds abstract. Here is the same finding made tangible: if you randomly draw two adults and check who has both the higher IQ and the higher income, you’ll be right about 60% of the time. Pure chance would be 50%. So IQ tilts the dice; it does not load them. The other 40% of cases — where the lower-IQ person earns more — is where most of life actually happens.
Job performance
Schmidt and Hunter’s 1998 meta-analysis, the most widely cited paper in this area, found cognitive ability is one of the strongest single predictors of job performance, with correlations rising as job complexity increases. For complex jobs (executive, scientific, professional), r is around 0.50. For simple, structured jobs, the correlation drops considerably and structured interviews predict performance better.
Educational attainment
Strong: r is approximately 0.50. IQ predicts how far someone goes in school more reliably than it predicts most other outcomes.
Life satisfaction
Once income is controlled for, the correlation between IQ and self-reported happiness or life satisfaction is close to zero. Smart people aren’t reliably happier than less-smart people.
What your score doesn’t predict
- Wisdom. IQ measures problem-solving in defined formats with a single right answer. Wisdom is the capacity to make sound judgments under ambiguity, with incomplete information and conflicting values. They are different cognitive systems and they correlate weakly — the Berlin Wisdom Paradigm studies (Baltes & Staudinger 2000) report correlations between IQ and wisdom-related performance in the r = 0.2–0.3 range.
- Social intelligence. The ability to read people, navigate group dynamics, and manage relationships is a separate construct that correlates only modestly with general IQ.
- Persistence. Angela Duckworth’s research on grit consistently finds that persistence is essentially uncorrelated with IQ but is one of the better predictors of long-term achievement.
- Creative output. Above an IQ of roughly 120, additional cognitive ability contributes very little to creative achievement (Simonton 2009). Past a threshold, what matters more is sustained engagement, domain knowledge, and risk tolerance.
- Moral character. No measurable relationship.
For most demanding cognitive work, what matters is reaching a threshold of capability, not maximizing the score. An IQ of 130 is enough for almost any profession. From there, what differentiates outcomes is the rest of the toolkit: discipline, curiosity, resilience, social competence.
The asymmetry most popular writing misses
Here is something the IQ-as-destiny camp gets right and the IQ-as-junk camp gets wrong: the score predicts much more on the low side than on the high side.
Above an IQ of roughly 120, additional cognitive ability stops sorting people much. Two physicians, one at IQ 125 and one at IQ 145, will not differ measurably in clinical judgment, patient outcomes, or career success on any large-sample comparison. The score has run out of resolution; what differentiates them is now grit, curiosity, communication, and the specific shape of their domain knowledge.
Below an IQ of roughly 85, by contrast, the score becomes a real constraint. Linda Gottfredson’s 1997 work on occupational complexity showed that jobs requiring rapid learning of novel procedures, juggling multiple instructions in working memory, or interpreting written material under time pressure become genuinely harder — not impossible, but markedly more effortful. Earl Hunt’s reviews of the same literature reach the same conclusion: IQ functions less like a meritocratic gradient and more like a floor on cognitive demand.
The practical consequence: most people fascinated by their own IQ score are between 100 and 130 — the band where the number matters least for life outcomes. The bands where it carries the most predictive weight (very low and very high) are the bands the average curious test-taker is statistically least likely to occupy.
This is not an argument against measuring your IQ. It is an argument for reading the result honestly: useful as a directional signal, weak as a tiebreaker, and most informative when something else — an unexplained struggle, a chronic mismatch with school or work — sent you looking for the number in the first place.
The Flynn Effect: why average IQ keeps rising
Here is the unsettling part: the population’s mean IQ keeps moving.
James Flynn documented in the 1980s that average IQ scores rose roughly 3 points per decade across the 20th century, in every country with longitudinal data. Trahan’s 2014 meta-analysis of 14 studies confirmed the pattern. An IQ of 110 today would have scored as roughly 125 in your grandfather’s time, on the same test.
Why does this happen? The leading explanations include better childhood nutrition, smaller family sizes, more years of formal schooling, increased exposure to abstract reasoning in everyday life (think: navigation apps, spreadsheets, complex media), and improvements in test-taking familiarity. No single explanation accounts for the whole effect.
What this means for you: your score is calibrated against the current population, not against a fixed standard. To stay current, IQ tests must be re-normed every decade or so. The number you receive is a position on today’s curve, not yesterday’s.
Online tests vs. clinical assessments
A clinical IQ assessment, such as the Wechsler Adult Intelligence Scale (WAIS-IV) or the Stanford-Binet (SB5), is administered in person by a licensed psychologist. It typically takes one to two hours, includes ten or more subtests, and is normed against thousands of proctored test-takers from a stratified national sample. The result is a clinically meaningful score, accurate to within a few points, that can be used in educational, occupational, or diagnostic contexts.
A 15-minute online IQ test is something different. The questions can be the same item types: matrix completion, number sequences, analogies, mental rotation. But the testing conditions, the norming sample, and the administration are not.
Here is what a well-designed online IQ test can give you:
- A directional read on your cognitive performance
- A breakdown across multiple domains, so you can see where you’re strongest
- A comparison against the standard IQ scale (mean 100, SD 15)
- A useful starting point for self-reflection
What it cannot give you:
- A score guaranteed accurate to within a few points
- A diagnosis or anything that holds up in a clinical or legal context
- A definitive verdict on your cognitive potential
The lesson: take the score as a starting point, not a verdict.
How to take an IQ test for the most accurate result
If you’re going to take an IQ test, online or otherwise, here’s how to give yourself the best read on your actual range:
- Sleep first. Sleep deprivation has been shown to reduce performance on standardized tests by 5 to 10 points (Killgore 2010). If you’re tired, do something else first.
- Eliminate distractions. Phone in another room. Door closed. Single sitting. The test you take while half-watching TV isn’t a measurement of your cognitive ability; it’s a measurement of your divided attention.
- Don’t look anything up. The score is only useful if it reflects what you can do without aids. Looking up a verbal analogy might bump your overall number, but you’ll have measured Google, not yourself.
- Take it once, not repeatedly. Practice effects inflate scores by 5 to 10 points on a second sitting. The first attempt is the most honest.
- If you do retake, space it out. Wait at least a few weeks. Average across attempts. Cognitive performance varies day to day; one bad night doesn’t define your range.
Frequently asked questions
Is a higher IQ always better?
Up to a point. Above an IQ of roughly 120, additional cognitive ability correlates weakly with most life outcomes. What you do with your cognitive baseline matters more than the baseline itself.
Can your IQ change?
Performance on IQ tests can shift by 5 to 10 points across a lifetime due to education, health, sleep, and practice effects. The underlying construct is more stable than the test score, but it is not fixed.
What is considered a genius IQ?
There is no formal cutoff. Mensa accepts the top 2 percent (an IQ of approximately 130 or above). Most modern IQ tests cap meaningfully around 160 because the norming sample becomes too thin to estimate reliably beyond that range.
Are IQ tests culturally biased?
Older versions had documented cultural biases. Modern tests, especially Raven’s Progressive Matrices, are designed to minimize cultural and language bias by relying on visual pattern recognition rather than vocabulary or general knowledge. Bias is an active research area, not a fully solved problem.
How accurate are online IQ tests?
A well-designed online IQ test can give you a directional read on your cognitive performance, particularly for pattern recognition and logical reasoning. It cannot match the precision of a clinical assessment administered in person by a licensed psychologist, which uses a normed sample of thousands of proctored test-takers and a more comprehensive battery of subtests.
Related reading
- Average IQ by Age — how cognitive scores are normed across the lifespan, and why fluid and crystallized intelligence move in opposite directions.
- Pattern Recognition Test — the cognitive domain most tightly correlated with general intelligence (r ≈ 0.7).
- Spatial Reasoning Test — the cognitive ability with the strongest known link to long-term STEM achievement.
- Logical Reasoning Test — deduction, syllogisms, and why even educated adults fail basic logic tasks.
- Numerical Reasoning Test — sequences, ratios, and the cognitive ability that predicts performance in finance and engineering.
- Verbal Reasoning Test — vocabulary, analogies, and the most trainable cognitive domain.
References
- Spearman, C. (1904). General intelligence, objectively determined and measured. American Journal of Psychology, 15, 201–293.
- Strenze, T. (2007). Intelligence and socioeconomic success: A meta-analytic review of longitudinal research. Intelligence, 35(5), 401–426.
- Schmidt, F. L., & Hunter, J. E. (1998). The validity and utility of selection methods in personnel psychology. Psychological Bulletin, 124(2), 262–274.
- Trahan, L. H., et al. (2014). The Flynn Effect: A meta-analysis. Psychological Bulletin, 140(5), 1332–1360.
- Simonton, D. K. (2009). Genius 101. Springer.
- Killgore, W. D. S. (2010). Effects of sleep deprivation on cognition. Progress in Brain Research, 185, 105–129.
- Duckworth, A. L., et al. (2007). Grit: Perseverance and passion for long-term goals. Journal of Personality and Social Psychology, 92(6), 1087–1101.
- Roberts, B. W., et al. (2007). The power of personality. Perspectives on Psychological Science, 2(4), 313–345.
- Gottfredson, L. S. (1997). Why g matters: The complexity of everyday life. Intelligence, 24(1), 79–132.
- Hunt, E. (2011). Human Intelligence. Cambridge University Press.
- Baltes, P. B., & Staudinger, U. M. (2000). Wisdom: A metaheuristic (pragmatic) to orchestrate mind and virtue toward excellence. American Psychologist, 55(1), 122–136.