Question 1

What is a good IQ score?

Accepted Answer

A 'good' IQ score depends entirely on the question being asked. Statistically, an IQ of 100 is the population median: half of all adults score above it, half below. So 100 is, in a literal sense, perfectly average — not a deficient result. Roughly 68% of the adult population scores between 85 and 115, defining the broad range that practitioners call 'average'. Scores above 115 are above average; scores above 130 are in the gifted range used by most school placement programs (the 98th percentile and above). For most everyday purposes, a score anywhere in the 90 to 120 range is consistent with reliable functioning across the cognitive demands of modern life. Scores below 70 may warrant clinical evaluation, especially if accompanied by adaptive-functioning concerns. The most important framing is that IQ is a population-relative measure — a single number says where you fall relative to other test-takers, not whether you are 'good enough' for any particular task or role.

Question 2

What is the average IQ?

Accepted Answer

By construction, the average IQ on any modern test is 100, with a standard deviation of 15. This is a definitional property of the test, not an empirical finding: when test publishers norm a new IQ battery, they convert raw scores to standard scores so that the mean of the norming sample equals 100 and the standard deviation equals 15. About 68% of adults score between 85 and 115, about 95% between 70 and 130, and about 99.7% between 55 and 145. Reports of 'national average IQ' that differ from 100 are typically the result of comparing one country's mean to another country's norming sample — the comparison is informative only if you know which sample is the reference. Within any given country normed sample, the mean is 100 by definition.

Question 3

What is a genius-level IQ?

Accepted Answer

There is no agreed-upon clinical or scientific definition of 'genius' that maps cleanly to a specific IQ cutoff. The popular folk definition typically places it somewhere between 140 and 160, but these numbers are essentially arbitrary. Lewis Terman's longitudinal study of high-IQ children, begun in the 1920s, used a cutoff of IQ 135 to identify 'gifted' subjects; many of his subjects went on to distinguished careers, but none won a Nobel Prize, while two children rejected from the study (William Shockley and Luis Alvarez) later did. The lesson is that exceptional cognitive ability is one input to exceptional achievement, but it is not sufficient and the threshold for 'genius-level' performance varies by domain. Mensa requires the 98th percentile (roughly IQ 130); the Triple Nine Society requires the 99.9th percentile (roughly IQ 146); the Mega Society requires the 99.9999th percentile. Each successive cutoff identifies a smaller and statistically less reliable group.

Question 4

What is the average IQ by age?

Accepted Answer

Modern IQ tests are age-normed, which means a 30-year-old and a 70-year-old who both score IQ 100 are not solving the same problems at the same level — they are both performing at the median for their age band. Raw cognitive performance changes substantially across the lifespan: fluid abilities (matrix reasoning, processing speed, working memory) peak in the late teens to mid-twenties and decline slowly thereafter, while crystallized abilities (vocabulary, general knowledge) rise through middle age and remain largely intact into old age. The age-norming convention means that 'average IQ' is always 100 within an age band, regardless of the actual cognitive performance level. If you compare your child's age-normed IQ to your own across decades, you are comparing relative position within different distributions, not raw cognitive ability.

Question 5

Can IQ change over time?

Accepted Answer

IQ scores can and do change. Test-retest reliability is high but not perfect — typical correlations between two administrations of the same test are 0.85 to 0.95. Within an individual, scores can shift by 5 to 10 IQ points between administrations due to measurement error alone, and by larger amounts due to changes in test conditions, motivation, sleep, or the test-taker's underlying cognition. Larger shifts in measured IQ across years of life are documented, especially in adolescence (when frontal-lobe development is still ongoing) and in old age (when cognitive decline begins to affect scores). The most actionable interpretation is that your IQ is reasonably stable across the medium term but is not a fixed property of your brain.

Question 6

Can IQ be improved with practice?

Accepted Answer

It depends on what you mean by 'improve IQ'. If you mean 'improve your score on a particular IQ test', the answer is yes — practice with similar items reliably produces 0.3 to 0.8 standard deviation improvements on subsequent administration. If you mean 'improve your underlying general cognitive ability', the evidence is much weaker. Working-memory training, n-back tasks, and commercial brain-training apps produce reliable gains on the trained tasks but transfer to general fluid intelligence is small and inconsistent in well-controlled studies. The most reliable cognitive interventions across studies are aerobic exercise (small to moderate gains across attention and executive function), adequate sleep (large effects on working memory and processing speed), and reading volume (substantial gains on crystallized intelligence over years).

Question 7

How accurate are online IQ tests?

Accepted Answer

Online IQ tests vary widely in quality. The best ones — typically those based on the ICAR public-domain catalog or the items released by the Open-Source Psychometrics Project — provide useful estimates of cognitive ability with documented psychometric properties. The standard error of measurement on a brief online screener is roughly ±7 IQ points, compared to ±3 for a clinically administered WAIS-IV. So a measured score of 115 on an online test means 'somewhere between about 101 and 129 with 95% confidence' — useful for self-curiosity but not precise enough for any high-stakes application. Tests that report exotic-sounding scores like 'IQ 168' from a 10-question quiz are not measuring what they claim to measure; ignore them. Reputable online tests will be transparent about their item sources, scoring approach, and measurement error.

Question 8

Why did I get different scores on different IQ tests?

Accepted Answer

Differences in scores across IQ tests are normal and expected. Different tests measure slightly different mixtures of cognitive abilities, use slightly different norming samples, and have slightly different measurement errors. The correlations between major adult IQ tests are typically 0.7 to 0.85 — high, but not perfect. A 10-point difference between two test scores is well within the normal range of measurement and instrument variability and does not indicate that one test is wrong. Larger discrepancies (20+ points) usually indicate that the tests are sampling different cognitive abilities — for example, a verbal-heavy test versus a fluid-reasoning-heavy test — or that one administration was affected by test conditions (fatigue, distraction, language difficulty).

Question 9

How long does a real IQ test take?

Accepted Answer

A clinically administered adult IQ test such as the WAIS-IV takes 60 to 90 minutes to complete, plus 30 to 60 minutes of scoring and report-writing by the clinician. Brief screening instruments such as the WASI-II take about 30 minutes. Online screening tests vary from 5 minutes (very brief, very imprecise) to 45 minutes (more thorough, still less precise than clinical tests). The 25-item screener on this site takes about 12 minutes for most test-takers. Length matters because IQ measurement precision improves with item count: a 10-item test cannot have the same standard error of measurement as a 60-item test, regardless of how good the items are.

Question 10

What is the difference between IQ and EQ?

Accepted Answer

IQ measures cognitive ability — reasoning, problem solving, working memory, processing speed. EQ (emotional intelligence) measures the ability to perceive, understand, and regulate emotions in oneself and others. The two constructs are statistically distinct: correlations between IQ and EQ measures typically run 0.1 to 0.3, low enough to confirm that they are not the same thing. EQ is moderately predictive of workplace outcomes that involve emotional labor (sales, healthcare, customer service) and adds some incremental validity above IQ for those roles. For roles that emphasize technical reasoning, IQ remains the better predictor. The popular framing that EQ is 'more important than IQ' is not supported by the empirical literature; both contribute, with the relative weight depending on the context.

Question 11

What IQ do I need to join Mensa?

Accepted Answer

Mensa requires a score at or above the 98th percentile on a battery of approved cognitive tests. On the WAIS-IV and Stanford-Binet 5, this corresponds to a full-scale IQ of approximately 130. On the Cattell IIIB (which uses a standard deviation of 24 instead of 15), the equivalent score is approximately 148. Mensa accepts scores from a long list of approved tests but does not accept results from online IQ tests, including this one. To qualify, you must take an in-person Mensa-administered test or submit prior results from a clinically administered test. About 1 in 50 adults qualifies by this standard.

Question 12

What is the highest IQ ever recorded?

Accepted Answer

Claims about 'the highest IQ ever' are essentially marketing rather than science. The reliability of any test at the extreme tails of the distribution is poor: the standard error of measurement at IQ 160 is much larger than at IQ 100, and most tests have a low ceiling that pins all extreme test-takers at the maximum reportable score. Marilyn vos Savant's reported IQ of 228, William James Sidis's reputed IQ of 250 to 300, and Christopher Langan's reported IQ of 195 are all based on tests with documented validity problems at the extremes. Modern psychometricians treat any IQ above about 160 as 'in the very superior range' rather than as a precise number. The Guinness Book dropped its 'highest IQ' category in 1990 specifically because the records could not be validated.

Question 13

What does it mean to have a low IQ?

Accepted Answer

A 'low' IQ score is one significantly below the population mean — typically defined as below 70 (the 2nd percentile, two standard deviations below the mean). In clinical practice, an IQ below 70 is one of several criteria used in the diagnosis of intellectual disability, but it is not sufficient on its own. The DSM-5 and ICD-11 both require evidence of impairment in adaptive functioning across conceptual, social, and practical domains, in addition to a low cognitive score. A single online screener cannot diagnose anything. People with measured IQs in the 70s and 80s — the borderline and low-average bands — typically function adequately in everyday life with appropriate education, support, and practice.

Question 14

Does a high IQ guarantee success?

Accepted Answer

A high IQ does not guarantee success, and a moderate IQ does not preclude it. The empirical correlation between IQ and life outcomes (academic achievement, income, occupational success) is reliable but modest — typically r = 0.3 to 0.5. This means IQ accounts for 9 to 25% of the variance in these outcomes, with the remaining 75 to 91% reflecting personality, motivation, opportunity, family background, and chance. The personality trait conscientiousness predicts long-term academic and occupational success roughly as well as IQ does, and the two predictors are largely independent. The most reliable summary is that high IQ removes some obstacles but does not create achievement on its own; sustained effort, good fortune, and supportive context matter at least as much.

Question 15

Does IQ test preparation actually work?

Accepted Answer

Yes — and this is one of the central problems for the construct validity of IQ testing. Practice with sample items from any major IQ test produces measurable score gains on a subsequent administration. Effect sizes vary by item type but typically run 0.3 to 0.8 standard deviations of improvement (4 to 12 IQ points). The gains are largest on the trained item types and on closely related items; transfer to other cognitive measures is more modest. For high-stakes IQ tests (school placement, Mensa qualification), test publishers attempt to control coaching effects by using alternate forms with rotated item content and by requiring waiting periods between administrations. For online screeners, no such controls exist, and practice effects are large enough to make repeated test-taking on the same instrument produce inflated scores.

Question 16

Does IQ predict creativity?

Accepted Answer

The relationship between IQ and creativity is one of the most-studied questions in the cognitive ability literature. The empirical pattern is the 'threshold hypothesis': below an IQ of approximately 120, IQ and creativity correlate positively; above 120, the correlation drops to near zero. This means that very high IQ is not strongly associated with very high creativity. Creativity also depends substantially on personality traits (openness to experience), motivation, domain expertise, and the social context in which creative work is produced. Highly creative individuals are typically intelligent — they need enough cognitive ability to master the domain in which they work — but among the cognitively capable, additional IQ points do not predict additional creativity.

Question 17

How does IQ predict academic performance?

Accepted Answer

IQ is one of the strongest single predictors of academic performance, with correlations of approximately 0.5 with school grades, 0.7 with standardized test scores, and 0.5 with years of schooling completed. The correlation is stronger for grades in subjects that emphasize abstract reasoning (mathematics, science) and weaker for subjects that emphasize execution (art, physical education). The correlation also strengthens at higher levels of academic difficulty: IQ predicts grades in graduate school more strongly than grades in elementary school. Importantly, IQ accounts for less than half the variance in academic performance even at the strongest correlations; conscientiousness, motivation, schooling quality, and family support account for substantial additional variance.

Question 18

Does IQ predict job performance?

Accepted Answer

Cognitive ability predicts job performance across virtually all occupations, with operational validities (Hunter & Schmidt meta-analyses) of approximately 0.5 for complex jobs and 0.2 for simple jobs. This makes IQ one of the strongest single predictors of job performance available to employers, comparable to structured interviews and substantially stronger than unstructured interviews, education credentials, or experience. The correlation is partly mediated by job-specific learning: cognitively able workers acquire job-relevant knowledge faster, which then translates into better performance. Cognitive testing for employment purposes is legal but heavily regulated in many jurisdictions, requiring documented job-relatedness and validation studies.

Question 19

How is a child IQ test different from an adult one?

Accepted Answer

Children's IQ tests are structured similarly to adult tests but use age-appropriate items and are normed against same-age peers. The most widely used clinical instruments are the Wechsler Intelligence Scale for Children (WISC-V) for ages 6 through 16, the Wechsler Preschool and Primary Scale of Intelligence (WPPSI-IV) for ages 2 through 7, and the Stanford-Binet 5 (SB5), which spans ages 2 through 90. Childhood IQ is less stable than adult IQ — the test-retest correlation across long intervals is lower because childhood cognitive development is rapid and uneven. Most clinicians weight a childhood IQ result less heavily than an adult one when interpreting it.

Question 20

Is memory the same as IQ?

Accepted Answer

Memory and IQ are correlated but distinct. Working memory — the active manipulation of information over short timescales — is one of the four CHC index scores reported by the WAIS-IV and correlates with general intelligence at approximately r = 0.5 to 0.7. Long-term memory (episodic and semantic) correlates with IQ more weakly. Memory ability can be impaired in isolation by neurological events (stroke, traumatic brain injury, dementia) without significant impact on other cognitive abilities, and conversely high IQ does not protect against memory disorders. The CHC framework treats memory as a broad ability separable from g, with its own subtest profile and clinical significance.

Question 21

What is a nonverbal IQ test?

Accepted Answer

A nonverbal IQ test measures cognitive ability using items that minimize language content. The most common nonverbal items are matrix reasoning and spatial rotation tasks. Nonverbal tests are particularly useful for test-takers whose first language differs from the test language, for individuals with hearing impairments or language disorders, and for cross-cultural research where language confounds would otherwise bias results. Common nonverbal batteries include Raven's Progressive Matrices, the Test of Nonverbal Intelligence (TONI-4), the Naglieri Nonverbal Ability Test (NNAT), and the Comprehensive Test of Nonverbal Intelligence (CTONI-2). Nonverbal IQ correlates highly with general intelligence (r ≈ 0.7 to 0.8) but is not identical — it samples a narrower range of cognitive abilities than a comprehensive battery.

Question 22

What is a culture-fair IQ test?

Accepted Answer

A 'culture-fair' IQ test is one designed to minimize cultural and educational confounds, typically by using abstract symbols and minimal language. Raymond Cattell's Culture Fair Intelligence Test (CFIT), first published in 1949, is the prototype; Raven's Progressive Matrices are widely cited as a culture-fair instrument. The 'culture-fair' label is somewhat aspirational: even abstract symbol manipulation depends on schooling, exposure to abstract reasoning tasks, and familiarity with multiple-choice formats — all of which vary by culture. A more accurate framing is that culture-fair tests are 'culture-reduced' rather than truly free of cultural influence. They remain useful for cross-cultural research and for testing populations where verbal items would introduce larger biases.

Question 23

How much does a real IQ test cost?

Accepted Answer

A clinically administered adult IQ test in the United States typically costs between US$500 and US$2,500, depending on the practitioner, location, and whether the test is part of a broader neuropsychological evaluation. The cost covers the clinician's time (60 to 90 minutes for the test plus scoring and reporting), the per-use cost of proprietary test materials, and professional overhead. School-administered IQ tests for children are usually provided at no cost as part of an educational assessment. Online IQ tests range from free (this site, the Open-Source Psychometrics Project) to several hundred dollars for proctored online versions of clinical instruments.

Question 24

Can I retake an IQ test?

Accepted Answer

You can retake an IQ test, but the second administration will overestimate your true ability due to practice effects. Effect sizes for repeat administration are typically 0.3 to 0.5 standard deviations on most batteries, larger if the test-taker explicitly studied items between administrations. Test publishers recommend a minimum interval of 6 to 12 months between retest administrations to allow practice effects to dissipate, and clinicians often switch to an alternate form (e.g., WAIS-IV → Stanford-Binet 5) for retests to minimize the practice effect. For online screeners, you can retake at any time, but you should expect inflated scores on subsequent administrations.

Question 25

How often should I take an IQ test?

Accepted Answer

There is no standard recommendation for how often a person 'should' take an IQ test. For self-knowledge purposes, once is enough — a single result with documented standard error of measurement provides as much information as you can use. Repeated testing on the same instrument produces inflated scores due to practice effects. For research or clinical follow-up purposes, IQ is typically retested at intervals of 1 to 5 years, with longer intervals preferred to minimize practice effects. After a neurological event (stroke, traumatic brain injury), follow-up testing at 3 months, 6 months, and 12 months is common to track recovery.

Question 26

What is a gifted IQ in children?

Accepted Answer

Gifted designation in school placement programs typically requires an IQ at or above the 95th to 98th percentile, depending on jurisdiction. The most common single threshold is IQ 130 (98th percentile) on a clinically administered battery such as the WISC-V or Stanford-Binet 5. Some programs use multiple criteria — IQ score plus academic performance plus teacher recommendation — to identify gifted students; others use single-score cutoffs. Gifted designation typically qualifies students for differentiated programming (acceleration, enrichment, pull-out programs), but the specific opportunities vary widely by school district. Twice-exceptional students — those who are both gifted and have a learning disability or attention disorder — are often missed by single-criterion screening because the disability masks the giftedness on group-administered tests.

Question 27

When is a low IQ a clinical concern?

Accepted Answer

A measured IQ below 70 is one of three criteria used in the clinical diagnosis of intellectual disability under the DSM-5. The other two criteria are deficits in adaptive functioning (the practical skills needed to live independently) and onset during the developmental period. All three must be present for a diagnosis. A low IQ score in isolation, without adaptive-functioning concerns, does not warrant a diagnosis. The standard error of measurement on clinical IQ tests is approximately 5 points, so a measured score of 70 has a 95% confidence interval of about 60 to 80 — meaning that the diagnostic threshold is approximate, not absolute. Anyone with a low screener score who has functional concerns should pursue a formal evaluation by a licensed clinician.

Question 28

Are IQ tests biased?

Accepted Answer

IQ tests have been the subject of substantial bias research over the past century. The empirical findings are nuanced. Predictive bias — the question of whether the same IQ score predicts different outcomes for different demographic groups — has been extensively studied and is generally found to be small or absent. The same IQ score predicts similar outcomes (academic achievement, job performance) across major demographic groups in well-designed validity studies. Construct bias — the question of whether the test measures the same underlying construct equivalently across groups — is more contested, particularly for verbal items that depend on cultural and linguistic familiarity. Modern test designers actively work to minimize content that depends on group-specific knowledge. The persistent fact that group means on IQ tests differ across some populations is a finding that requires careful interpretation and is the subject of ongoing research and debate.

Question 29

How much of IQ is genetic?

Accepted Answer

Twin and adoption studies estimate the heritability of adult IQ at approximately 0.5 to 0.8 in the populations studied. This means 50 to 80% of the variance in adult IQ within a population is statistically attributable to genetic differences. Importantly, heritability is a population-level statistic, not an individual one — it tells you about variance within a population, not about what proportion of any individual's IQ is 'due to genes'. Heritability also depends on the environmental range of the studied population: in highly equal environments, more residual variance is genetic; in highly unequal environments, more residual variance is environmental. Genome-wide association studies (GWAS) have identified thousands of variants that contribute to IQ, each with very small individual effects. Polygenic scores currently account for 10 to 20% of variance in IQ in the populations studied — useful for research, currently of very limited individual predictive power.

Question 30

Can I improve my IQ score quickly?

Accepted Answer

There is no reliable way to substantially improve your underlying cognitive ability quickly. Score improvements on a particular IQ test from practice and coaching are real and can be substantial (4 to 12 IQ points), but these reflect changes in the test rather than changes in the underlying ability. Acute factors that can move your screener score by 5 to 10 points within hours: getting adequate sleep, eating, hydrating, ensuring a quiet test environment, and using your habitual caffeine. Short-term improvements in working memory and executive function follow from regular aerobic exercise (effects develop over weeks). The largest documented short-term cognitive enhancer is recovery from sleep deprivation: a single night's full sleep can recover roughly 1 standard deviation of working memory in chronically sleep-deprived test-takers.

Questions about IQ tests, scores, and cognitive ability

About IQ Scores & Bands

General Questions about IQ

Improvement & Brain Training

Test Format & Accuracy