KBIT-2 Test

Posted on October 5, 2026 by Mahesh

Scores & Scales

KBIT-2 Test: Subtests, Scores and What the Kaufman Brief Intelligence Test Measures

The KBIT-2 is a 20-minute IQ screen for ages 4 to 90 with three subtests: Verbal Knowledge, Riddles and Matrices. See how its scores work, what the 2022 KBIT-2 Revised changed, how closely it matches the WISC-V and KABC-II, and why it is a screen rather than a full evaluation.

By Leon Falkner, Data & Research AnalystPublished October 5, 2026 · 8 min readReviewed by Aarav Vardhan, Editor-in-Chief

Three cards about the KBIT-2. Three subtests, three scores: Verbal Knowledge and Riddles plus Matrices, for ages 4 to 90, in about 20 minutes, giving Verbal, Nonverbal and IQ Composite standard scores with a mean of 100 and a standard deviation of 15. Revised in 2022: the revision uses norms from 2021 data, while the first edition used data collected in 2002 to 2003. Agreement with longer tests: Pearson reports a correlation of .77 between the revised IQ Composite and the KABC-II NU, and an independent study of 101 children found .72 with the WISC-V but wide error in predicting individual scores.

The KBIT-2 (Kaufman Brief Intelligence Test, Second Edition) is a brief IQ screening test, given one-to-one by a trained examiner in about 20 minutes, for people aged 4 to 90. It has three subtests, Verbal Knowledge and Riddles for crystallized (verbal) ability and Matrices for fluid (nonverbal) ability, and it reports Verbal, Nonverbal and IQ Composite standard scores with a mean of 100 and a standard deviation of 15. Pearson published a revised edition, the KBIT-2 Revised, in 2022, with norms built from 2021 data. Its job is screening: Pearson’s own technical report says a longer test such as the WISC-V is the next step when screening results warrant it. This guide covers the subtests, what changed in the revision, how scores are reported, how well the KBIT-2 agrees with longer tests and who can give it.

What it is: a brief, individually administered intelligence screen; about 20 minutes; ages 4:0 to 90:0; qualification level B at Pearson.
Subtests and scores: Verbal Knowledge and Riddles feed the Verbal (crystallized) score; Matrices is the Nonverbal (fluid) score; all three make the IQ Composite.
Scale: standard scores with a mean of 100 and an SD of 15, plus percentile ranks by age.
The revision: KBIT-2 Revised (2022), norms from 2021 data; the first edition used data collected in 2002 to 2003.
Accuracy: Pearson reports r = .77 between the revised IQ Composite and the KABC-II NU; an independent study of 101 children found r = .72 with the WISC-V FSIQ but wide error for single scores.

What is the KBIT-2?

The KBIT-2 succeeded the original K-BIT and is published by Pearson, the same publisher as the Wechsler scales. It is meant for the situations where a long evaluation is not practical. Pearson’s technical report describes the demand directly: school psychologists face increasingly multifaceted evaluations, and in some settings practitioners routinely give a brief test like the KBIT-2 Revised to screen for cognitive issues, then a more comprehensive test, such as the KABC-II NU, the WISC-V or the Woodcock-Johnson IV Tests of Cognitive Abilities, if screening indicates more testing is warranted. For how brief tests compare on time, see our guide to how long an IQ test takes.

What are the KBIT-2 subtests?

The three KBIT-2 subtests
Subtest	What the examinee does	Feeds which score
Verbal Knowledge	Looks at pictures in a stimulus book and shows or says what fits a word or question, a vocabulary and general-knowledge task	Verbal (crystallized)
Riddles	Listens to a short riddle read aloud and answers with a word	Verbal (crystallized)
Matrices	Looks at a pattern or picture grid with a piece missing and picks the response that completes it	Nonverbal (fluid)

Pearson describes Verbal Knowledge and Riddles as assessing crystallized ability, the store of words and knowledge a person has accumulated, and Matrices as a measure of nonverbal fluid reasoning. It adds that gifted and talented centers use Matrices as a standalone measure of nonverbal fluid reasoning. Fluid reasoning is the same ability that working-memory training claims aim at; see our dual n-back guide for the evidence on that, our executive function vs IQ guide for which control abilities track it, and our fluid vs crystallized intelligence guide for the distinction. Matrices items work like the ones in Raven’s matrices.

The KBIT-2 Revised shares design with a longer test: its Verbal Knowledge and Riddles subtests were built as parallel forms of the same-named subtests in the KABC-II Normative Update. Pearson says that when both tests are needed, substituting the KBIT-2 Revised scores into the KABC-II NU saves about 10 to 20 minutes and avoids repeating the same task, which could distort scores through procedural learning, changed effort or regression to the mean.

What changed in the KBIT-2 Revised?

Pearson released the revision in 2022. Its own side-by-side sheet and FAQ list the changes:

New norms. The revision is based on 2021 data collection, with one norm for in-person and remote administration. The original KBIT-2 data were collected in 2002 and 2003.
Refreshed items. Outdated drawings were replaced on several Verbal Knowledge items, and items were updated for auditory issues, diversity and alternate correct responses, with the target concepts left unchanged.
Matrices. Pearson dropped an item that showed a book of matches because young children are not exposed to matches, after finding it could go with no effect on reliability, and added newly designed difficult items to extend the upper range.
Riddles. Modern answers were added to the scoring responses, five items were dropped for diversity and bias concerns and two difficult items were added, making the subtest shorter.
Scoring. Q-global scoring is now an option, and the scores come with 90% and 95% confidence intervals.
Materials. The Manual and Record Form changed, so the two editions’ materials are not interchangeable, but Pearson says the overall scoring and administration procedures stay familiar.

How are KBIT-2 scores reported?

The KBIT-2 reports three scores, Crystallized (Verbal), Fluid (Nonverbal) and IQ Composite, as standard scores with a mean of 100 and a standard deviation of 15, along with percentile ranks by age. The table below shows what those standard scores mean on the normal curve; the published percentile ranks come from Pearson’s own norm tables and can differ slightly from the smooth curve.

Standard scores and percentiles on a mean-100, SD-15 scale
Standard score	Percentile (normal curve)	Our guides
70	2.3%	See IQ classifications
85	15.9%	See the percentile calculator
100	50%	The average by definition
115	84.1%	See IQ classifications
130	97.7%	See gifted cut-off scores

Because the KBIT-2 is short, a single score carries more uncertainty than a full scale; Pearson reports confidence intervals for that reason. If you want to compare against a scale with a different standard deviation, our SD 15 vs SD 16 guide shows the conversion.

Your own number

Where would your own score land?

Take the IIF-certified assessment and get your score with the scale it was measured on, the percentile it corresponds to and the confidence range around it — the three figures most online tests leave out.

Find your IQ score now! →

Secure & encryptedInstant results10–20 minutes

How well does the KBIT-2 agree with longer IQ tests?

Pearson’s KBIT-2 Revised and KABC-II NU technical report (Table 1.1) correlates the two tests. The KBIT-2 Revised IQ Composite correlates .77 with the KABC-II NU Fluid-Crystallized Index and .67 with its Mental Processing Index. The Verbal score correlates .87 with the KABC-II Knowledge/Gc scale. The Matrices-based Nonverbal score correlates .61 with Pattern Reasoning and .52 with the Planning/Gf scale, lower than the verbal link, and the same-named Verbal Knowledge and Riddles subtests correlate .75 across the two tests. These are the publisher’s own numbers.

An independent look at the original KBIT-2 comes from Cheng and colleagues (2024), who reviewed records of 101 children with developmental disorders at a children’s hospital in Singapore. The KBIT-2 IQ Composite correlated r = .722 with WISC-V Full Scale IQ and r = .648 with WPPSI-IV Full Scale IQ. Regression analysis showed that KBIT-2 scores tended to underestimate or overestimate Wechsler scores, with the highest variability for the Nonverbal IQ against the WISC-V Fluid Reasoning Index (error limits −53.1 to 39.7). The authors concluded that the KBIT-2 has good concurrent validity and is a good screening tool in this group, and that more studies are needed before it is used for diagnostic decisions or special-education placement.

Squared, a correlation of .72 means the two tests share about 52% of their variance, and .77 about 59%. That is strong enough for screening and not strong enough to treat a KBIT-2 score as a stand-in for a full Wechsler result. Our guides to the WPPSI-IV, the WISC-V, the Stanford-Binet and the Woodcock-Johnson cover the longer instruments.

Who gives the KBIT-2, and can you take it online?

It is a professional test, not an online quiz. Pearson lists the KBIT-2 Revised at qualification level B, one level below the level C that Wechsler scales need (see how an official IQ test works), and it is administered by a trained examiner using a stimulus book. Pearson’s comparison sheet lists administration in person and remotely for the revised edition, with digital stimulus books and the manual on Q-global, and the product page lists a Spanish version and telepractice guidance. A page or PDF online that calls itself the KBIT is not the test, because the items and norms are the publisher’s. If what you want is a score on the IQ scale from home, our IQ test is built for that.

When is a brief test enough?

For screening, to find out whether a longer evaluation is worth the time, a brief test does its job, and that is how Pearson positions it. For decisions with consequences, such as a school placement or a disability determination, the longer instruments are the usual next step, and the Singapore authors above say more research is needed before the KBIT-2 supports such decisions. Gifted programs are a nuance: Matrices alone is used by some centers as a nonverbal measure, but a cut-off score on any single test has an interval around it; our gifted cut-off scores guide and guide to IQ testing for children explain how those decisions are made.

Quick answers

What is the KBIT-2 test? The Kaufman Brief Intelligence Test, Second Edition: a brief, individually administered intelligence screen from Pearson.
What does the KBIT-2 measure? Verbal (crystallized) ability through Verbal Knowledge and Riddles, and nonverbal (fluid) ability through Matrices, combined into an IQ Composite.
How long does the KBIT-2 take? About 20 minutes, according to Pearson.
What ages is the KBIT-2 for? 4 years 0 months to 90 years 0 months.
What is the KBIT-2 Revised? A 2022 update with norms from 2021 data, refreshed items, new hard items in Matrices and Riddles, and Q-global scoring.
What scale does the KBIT-2 use? Standard scores with a mean of 100 and a standard deviation of 15.
Is the KBIT-2 as accurate as the WISC-V? It correlated .72 with WISC-V Full Scale IQ in one independent study, with wide error for single scores; it is a screen, not a replacement.
Can I take the KBIT-2 online? No. A trained examiner gives it; Pearson sells it at qualification level B.

Sources

Pearson Assessments: Kaufman Brief Intelligence Test, Second Edition Revised (KBIT-2 Revised), product page: age range, scores, qualification level, completion time and norms.
Pearson: KBIT-2 Revised comparison sheet (what changed, side by side), 2022.
Pearson: KBIT-2 Revised FAQs (publication, compatibility, scoring and administration).
Pearson: KBIT-2 Revised and KABC-II NU Technical Report (correlations in Table 1.1, the substitution studies and the screening rationale).
Cheng, A.S.M. et al. (2024): Concurrent validity of intelligence assessments in children with developmental disabilities in an Asian setting: comparison of the Kaufman brief intelligence test – Second edition with the Wechsler Intelligence Scales, Pediatrics & Neonatology, 65(4), 341-347.

Share this article

Know someone who keeps seeing these numbers quoted without the scale they were measured on? Send it to them.

Keep reading

All articles →

Taking a Test

Professional IQ Tests

When a psychologist measures intelligence, they reach for one of a small number of standardised instruments. This guide covers what the Wechsler scales, Stanford-Binet, Woodcock-Johnson and Raven’s Progressive Matrices each measure, who each one is for, and where an online test genuinely fits.

August 27, 20266 min read

Taking a Test

How Long Does an IQ Test Take?

An official IQ test usually takes 45 to 90 minutes of testing: 45 for the WAIS-5, 60 to 90 for the WAIS-IV, 12 for the Wonderlic. See the time for every major test, how many questions to expect, why two people take different amounts of time and what a full evaluation adds.

September 30, 202610 min read

Scores & Scales

WPPSI-IV Test

The WPPSI-IV test covers ages 2:6 to 7:7 in two bands, with 15 subtests, five index scores and a Full Scale IQ. See what each subtest asks, how a published sample report reads, how the score labels work, and what a preschool score can and cannot tell parents.

September 26, 202610 min read

Can a Concussion Lower Your IQ?

Posted on October 5, 2026 by Mahesh

Research & Evidence

Can a Concussion Lower Your IQ? What the Meta-Analyses Show

Can a concussion lower your IQ? Meta-analyses of mild and severe head injury show scores dip in the first days, recover within weeks for most people and stay lower after moderate or severe injury. See the effect sizes in IQ points, why clinic samples look worse and why a baseline is rarely available.

By Leon Falkner, Data & Research AnalystPublished October 5, 2026 · 8 min readReviewed by Aarav Vardhan, Editor-in-Chief

A bar chart of effect sizes, Cohen's d, from published meta-analyses of head injury and cognitive scores. Mild traumatic brain injury, delayed memory in the first 3 months: 1.03. Moderate-severe injury across all follow-up periods: 0.74. Mild injury, all domains pooled: 0.54. Mild head injury across all follow-up periods: 0.24. Two or more concussions versus one in athletes: 0.06. Mild injury after 3 months in unselected or prospective samples: 0.04. On an IQ scale, d of 1.0 is 15 points; these are averages across cognitive tests, not IQ scores.

Can a concussion lower your IQ? For most people a single concussion does not leave a lasting drop, but the effect is real in the first days and weeks, and it is larger and longer-lasting after moderate or severe brain injury. A meta-analysis of 39 studies of mild traumatic brain injury (1,463 cases and 1,191 controls) found a moderate overall effect on cognitive tests (d = .54) that had disappeared by 3 months in unselected or prospective samples (d = .04). For sports concussion, a 21-study analysis found no residual impairment when testing was done more than 7 days after injury. Moderate-to-severe brain injury is different: it averages d = −0.74 and remains markedly impaired beyond 2 years. This guide turns those effect sizes into IQ points, explains why clinic samples look worse than community samples, and shows why a true before-and-after IQ comparison is rare.

Most single concussions: cognitive scores recover within weeks. In unselected or prospective samples, mild injury showed no residual effect by 3 months (d = .04).
The first days: the biggest effects are on memory and global cognitive functioning; after a sports concussion, no residual impairment was found when testing was done beyond 7 days.
Who you sample matters: at 3 months or later, clinic-based samples showed d = .74 and litigation samples d = .78, against .04 in unselected samples.
Repeated concussions: across 8 athlete studies, overall d = .06, with poorer delayed memory and executive functioning.
Moderate and severe injury: d = −0.74 on average, more than three times the mild-injury figure, and still markedly impaired more than 2 years after injury.

What counts as a concussion, and how is it different from a severe brain injury?

A concussion is commonly defined as a mild traumatic brain injury (MTBI): a blow or jolt that briefly disrupts how the brain works, with symptoms such as confusion, headache, dizziness or gaps in memory. Clinicians grade traumatic brain injury (TBI) as mild, moderate or severe, mainly by the Glasgow Coma Scale, how long consciousness was lost and how long memory was affected. The research below keeps the groups apart because they behave differently, and because a concussion and a severe injury are different questions. These studies did not measure a place on the IQ classification scale; they measured performance on cognitive tests.

What happens to cognitive scores in the first days and weeks?

Belanger and Vanderploeg (2005) pooled 21 studies of sports-related concussion (790 cases, 2,014 controls) and found an overall effect of d = 0.49. Within 24 hours of injury the effects were largest for delayed memory (d = 1.00), memory acquisition (1.03) and global cognitive functioning (1.42). Beyond 7 days after injury, they found no residual neuropsychological impairment, although delayed memory in studies with a control group was still a problem at 7 days. Using athletes with a history of prior head injury appeared to inflate the effect sizes for the current concussion.

Belanger and colleagues (2005) found the same shape in non-sports mild TBI across 39 studies: the acute effects (under 3 months) were greatest for delayed memory (d = 1.03) and fluency (d = .89). Schretlen and Shapiro (2003) add that the natural logarithm of the follow-up interval correlated very strongly with the effect size after mild head injury, and that overall cognitive functioning “recovers most rapidly during the first few weeks” and essentially returns to baseline within 1 to 3 months.

Effect sizes in the published meta-analyses
Group and timing	Cohen d	IQ points at SD 15	Source
Sports concussion, first 24 hours, global cognitive functioning	1.42	21	Belanger and Vanderploeg 2005
Mild TBI, under 3 months, delayed memory	1.03	15	Belanger et al. 2005
Mild TBI, under 3 months, fluency	0.89	13	Belanger et al. 2005
Moderate-severe TBI, all follow-up periods	0.74	11	Schretlen and Shapiro 2003
Mild TBI, all studies and domains	0.54	8	Belanger et al. 2005
Sports concussion, all studies	0.49	7	Belanger and Vanderploeg 2005
Mild head injury, all follow-up periods	0.24	4	Schretlen and Shapiro 2003
Two or more concussions vs one (athletes)	0.06	1	Belanger et al. 2010
Mild TBI, 3+ months, unselected or prospective samples	0.04	less than 1	Belanger et al. 2005

Read the third column with care. It simply multiplies d by 15 to show the size on the familiar IQ scale (see SD 15 vs SD 16). The effect sizes themselves are group averages across cognitive tests, mostly memory, attention and executive measures, not full-scale IQ scores, and an average hides individuals who recover slowly or fast.

What happens after 3 months?

This is where the headline numbers split. In unselected or prospective samples, Belanger and colleagues found no residual neuropsychological impairment by 3 months after mild TBI (d = .04). In clinic-based samples and in samples that included people in litigation, the effects at 3 months or later were larger (d = .74 and .78), and litigation was associated with cognitive functioning that stayed the same or got worse over time. The abstract does not give the reasons, and we will not guess at any one person’s. What it shows is that the lingering effect you read about depends heavily on who was sampled: people drawn from the community and followed from the injury had recovered on average, and people who came through clinics or legal claims showed larger effects. A person who still has trouble months later deserves a proper assessment either way.

Do repeated concussions lower IQ?

Belanger, Spiegel and Vanderploeg (2010) pooled 8 studies, all of athletes, comparing 614 people with more than one self-reported mild TBI against 926 with just one. The overall effect was minimal and not significant (d = 0.06), but a history of multiple self-reported concussions went with poorer delayed memory and executive functioning. The samples were athletes tested after self-reported histories, so the result cannot speak for people outside sport or for those with severe injuries.

Your own number

Where would your own score land?

Find your IQ score now! →

Secure & encryptedInstant results10–20 minutes

What about moderate and severe injuries, and children?

Schretlen and Shapiro (2003) pooled 39 mostly cross-sectional studies (48 comparisons; 1,716 patients and 1,164 controls). Averaged across follow-up periods, the effect of moderate-severe TBI (d = −0.74) was more than three times the effect of mild head injury (d = −0.24). Cognitive functioning improved over the first two years after moderate-severe TBI but remained markedly impaired among patients tested more than 2 years after injury.

For children, Babikian and Asarnow (2009) summarized 28 publications from 1988 to 2007. In severe pediatric TBI they report moderate to very large impairments in intellectual functioning, mainly performance IQ, with substantial recovery in longitudinal studies, especially for performance relative to verbal IQ, although by the longest follow-up the severe group showed impairments in all the domains the review covered. For children with moderate TBI they say some deficits persist beyond two years compared with matched peers. Index scores such as processing speed matter in these evaluations; our guide to processing speed and IQ explains what that index measures.

If nobody tested your IQ before, how can anyone tell it dropped?

Usually they cannot, directly. Few people have a baseline IQ score from before an injury, so clinicians estimate premorbid ability from skills that hold up, such as reading irregular words aloud. The best-known example is the National Adult Reading Test (NART). In a study by McGurn and colleagues (2004), an IQ test taken at age 11 correlated r = .63 with NART scores at about age 80 in people who had developed dementia and r = .60 in those who had not, and after controlling for age-11 IQ, mean NART scores did not differ between the groups. That study was about dementia; clinicians apply the same logic after brain injury, with reading tests such as the NART or Pearson’s WTAR and TOPF.

A correlation of .60 explains about 36% of the variance (r squared), so a reading-based estimate is a range, not a number, and a drop of a few points will not stand out against it. Our guides to whether an IQ test can detect dementia and how long an IQ result stays valid cover the same limits from other angles. Brief tests also have a role in these evaluations; see our KBIT-2 guide for one example of a short IQ screen.

What should you do if you are worried?

If symptoms continue longer than expected after a head injury, or a child’s schoolwork changes, ask a physician about a neuropsychological evaluation. An online IQ quiz cannot separate injury from effort, mood, sleep, pain or medication, and it cannot show change without a baseline; what it can tell you is described in our guide to whether IQ tests are reliable. Attention and executive-control tasks, such as the Stroop test (see also our executive function vs IQ guide), are part of the toolkit clinicians use alongside IQ tests, not a substitute for them.

Quick answers

Can a concussion lower your IQ? Temporarily, yes: cognitive scores dip in the first days and weeks. In unselected samples there was no residual effect by 3 months after mild injury.
Does a concussion cause a lasting drop in cognitive scores? Not in group averages for most single concussions, but clinic-based samples show larger long-term effects and moderate-severe injuries do.
How long does it take cognitive function to recover after a concussion? For sports concussion, no residual impairment was found beyond 7 days; for non-sports mild TBI, by about 3 months.
Do multiple concussions lower IQ? Overall d = .06 across 8 athlete studies, with poorer delayed memory and executive functioning.
How many IQ points is a d of 0.5? About 7.5 points on a 15-point scale, but these are cognitive-test averages, not IQ scores.
How do doctors know your IQ before a head injury? They usually estimate it from reading tests such as the NART, which correlated about .60 with age-11 IQ.

Sources

Belanger, H.G., Curtiss, G., Demery, J.A., Lebowitz, B.K. and Vanderploeg, R.D. (2005): Factors moderating neuropsychological outcomes following mild traumatic brain injury: a meta-analysis, Journal of the International Neuropsychological Society, 11(3), 215-227.
Belanger, H.G. and Vanderploeg, R.D. (2005): The neuropsychological impact of sports-related concussion: a meta-analysis, Journal of the International Neuropsychological Society, 11(4), 345-357.
Belanger, H.G., Spiegel, E. and Vanderploeg, R.D. (2010): Neuropsychological performance following a history of multiple self-reported concussions: a meta-analysis, Journal of the International Neuropsychological Society, 16(2), 262-267.
Schretlen, D.J. and Shapiro, A.M. (2003): A quantitative review of the effects of traumatic brain injury on cognitive functioning, International Review of Psychiatry, 15(4), 341-349.
Babikian, T. and Asarnow, R. (2009): Neurocognitive outcomes and recovery after pediatric TBI: meta-analytic review of the literature, Neuropsychology, 23(3), 283-296.
McGurn, B. et al. (2004): Pronunciation of irregular words is preserved in dementia, validating premorbid IQ estimation, Neurology, 62(7), 1184-1186.

Share this article

Know someone who keeps seeing these numbers quoted without the scale they were measured on? Send it to them.

Keep reading

All articles →

Diagram showing where processing speed sits in the structure of cognitive ability, with simple reaction time, choice reaction time and inspection time on one axis and their differing strength of association with general intelligence shown alongside

Understanding IQ

Processing Speed and IQ

Fast thinkers are assumed to be smart thinkers, and there is a real finding underneath the stereotype: people who score higher on IQ tests do respond faster on tasks with no reasoning content at all. But the link is weaker than the folklore, and on a professional battery the speed index is the least g-loaded.

September 1, 20266 min read

Taking a Test

Can an IQ Test Detect Dementia?

A standard IQ test cannot diagnose dementia; it has no personal baseline and was not built to screen for decline. How the MMSE and the more sensitive MoCA compare, why a normal score can hide early decline, and what actually helps catch it.

September 22, 20266 min read

Taking a Test

How long is IQ test valid?

Do you want to know how long an IQ test is valid for? Have you ever wondered how accurate it is and how it affects your IQ? Stop searching now. To better understand the significance of IQ test results, this article will investigate the validity period of these assessments.

March 11, 20247 min read

Walking Speed and IQ

Posted on October 5, 2026 by Mahesh

Research & Evidence

Walking Speed and IQ: What the Gait Research Shows

Walking speed and IQ are linked: in 904 adults aged 45, higher full-scale IQ went with faster walking (0.38). See what else went with gait speed, why childhood brain health at age 3 mattered, what walking speed predicts in older adults and why it is a health marker, not an IQ test.

By Arjun Vardhan, News & Research WriterPublished October 5, 2026 · 7 min readReviewed by Aarav Vardhan, Editor-in-Chief

A bar chart of standardized associations with gait speed in 904 adults aged 45 in the Dunedin study, adjusted for sex. Full-scale IQ at 45 measured with the WAIS-IV: 0.38. Grip strength: 0.36. Chair-stand test: 0.34. Neurocognitive health at age 3: 0.26. Total brain volume: 0.15. Cortical thickness: 0.09. These are associations, not proof of cause.

Walking speed and IQ are related: in a New Zealand cohort of 904 adults aged 45, people with a higher full-scale IQ walked faster, with a standardized association of 0.38 (95% CI 0.32 to 0.44). That was larger than the association with total brain volume (0.15), and poor neurocognitive functioning measured as early as age 3 also predicted slower walking in midlife (0.26). It is an association measured at one point in adulthood. It does not show that walking faster raises IQ, or that a slow walker has a low IQ. Gait speed seems to work as a marker of general physical and brain health. This guide covers what the study measured, what else walking speed went with, what it predicts in older adults and how to read the numbers.

The study: Dunedin cohort, 904 adults at age 45; gait speed measured three ways (usual pace, walking while reciting letters, maximum pace).
The IQ link: full-scale WAIS-IV IQ was associated with a composite gait speed at 0.38; each of the four WAIS-IV indexes was also associated (all P < .001).
Beyond IQ: grip strength (0.36), the chair-stand test (0.34), a faster pace of aging (−0.33) and childhood neurocognitive health at age 3 (0.26) all went with gait speed.
Older adults: across 9 cohorts of 34,485 people aged 65 and over, each 0.1 m/s faster gait was associated with a hazard ratio for death of 0.88 (Studenski et al., 2011).
The catch: an association at one time point; gait speed was not measured before age 45.

What did the Dunedin study find?

Rasmussen and colleagues (2019) used the Dunedin Multidisciplinary Health and Development Study, a population-based 1972 to 1973 birth cohort in New Zealand. Of 1,037 original participants, 997 were alive at 45, and 904 (90.7%) had their walking speed measured. Walking was timed under three conditions: usual pace from a standing start; the same pace while reciting alternate letters of the alphabet aloud (dual-task gait); and as fast as safely possible. The means were 1.30 meters a second (SD 0.17) for usual gait, 1.16 (SD 0.23) for dual-task gait and 1.99 (SD 0.29) for maximum gait. The three conditions were combined into a composite gait speed.

At 45, each person also took the WAIS-IV, which gives a full-scale IQ (mean 100.1, SD 14.9 in this cohort) and four index scores: processing speed, working memory, perceptual reasoning and verbal comprehension. Higher full-scale IQ went with faster composite gait, with a standardized regression coefficient of 0.38 (95% CI 0.32 to 0.44). For full-scale IQ the coefficients ranged from 0.23 to 0.39 across the individual walking conditions, and each of the four indexes was also associated with gait speed (all P < .001). The authors adjusted for sex, and report further adjustments for leg length, body composition and childhood socioeconomic status in supplementary tables.

What else goes with a slower walk?

IQ was one of many correlates, which is the main point. Gait speed also went with physical function, biological aging and brain structure, as the table shows. The authors’ conclusion is that gait speed is associated with more than geriatric functional status; it is also associated with midlife aging and lifelong brain health.

Standardized associations with gait speed at 45
Measure	Coefficient	Direction
Full-scale IQ at 45	0.38	higher IQ, faster walking
Grip strength	0.36	stronger grip, faster walking
Chair-stand performance	0.34	better performance, faster walking
2-minute step test	0.33	better performance, faster walking
Pace of biological aging	−0.33	faster aging, slower walking
Balance	0.28	better balance, faster walking
Self-reported physical limitations	−0.27	more limitations, slower walking
Neurocognitive functioning at age 3	0.26	poorer functioning, slower walking in midlife
Visual-motor coordination	0.24	better coordination, faster walking
Total brain volume	0.15	smaller volume, slower walking
Cognitive change from childhood to 45	0.10	greater decline, slower walking
Cortical thickness	0.09	more thinning, slower walking

The childhood results matter most for interpretation. A composite of neurocognitive functioning at age 3, built from a 45-minute exam that included a pediatric neurologist, standardized tests of intelligence, language and motor skills, and ratings of emotional and behavioral regulation, predicted walking speed four decades later (0.26). Cognitive decline from childhood to midlife also went with slower gait (0.10). For the speed side of WAIS-IV scores, see our guide to processing speed and IQ.

Does walking faster mean you have a higher IQ?

No, and the study does not claim it. The data are associations, and the authors list a clear limitation: they had no walking measurement before age 45 and no brain imaging before then, so they could not follow change over time. A person who walks faster is also, on average, stronger, better balanced and biologically younger in this cohort, which is why the authors describe gait speed as associated with midlife aging and lifelong brain health. Gait speed is better read as a broad marker than as a measure of intelligence: it does not tell you someone’s IQ, and it is not part of any IQ test.

Whether exercise itself changes cognitive scores is a different question, and a different kind of evidence. Our guide to exercise and IQ covers the trials.

Why would walking relate to the brain?

Walking looks automatic but draws on balance, sensory input and attention, which is one reason researchers add a mental task and watch what happens. In the Dunedin cohort, reciting letters aloud slowed the average walk from 1.30 to 1.16 meters a second, about 11%. The study does not isolate a mechanism, and we will not offer one; what it shows is that slower walking at 45 went, on average, with weaker brain, body and childhood indicators. The executive side of that dual-task burden is the subject of our guide to executive function vs IQ.

Your own number

Where would your own score land?

Find your IQ score now! →

Secure & encryptedInstant results10–20 minutes

What does walking speed predict in older adults?

Studenski and colleagues (2011) pooled individual data from 9 cohort studies of 34,485 community-dwelling adults aged 65 and over, followed for 6 to 21 years. The mean gait speed was 0.92 meters a second (SD 0.27), and there were 17,528 deaths. Gait speed was associated with survival in every study, with a pooled hazard ratio of 0.88 per 0.1 meters a second faster (95% CI 0.87 to 0.90). Survival increased across the whole range of speeds, and predictions from age, sex and gait speed were as accurate as predictions from age, sex, chronic conditions, smoking history, blood pressure, body mass index and hospitalization.

A 2016 meta-analysis by Beauchet and colleagues pooled 12 studies and found that poor gait performance, defined by standardized walking tests, predicted dementia: a pooled hazard ratio of 1.53 for any dementia, 1.79 for vascular dementia and 1.89 for non-Alzheimer dementias. The finding was weaker for Alzheimer disease (1.03). That is a statement about groups followed over time; a single slow walk is not a diagnosis. For what IQ tests can and cannot show about dementia, see our guide to whether an IQ test can detect dementia.

How should you read a walking-speed result?

Treat it as a health indicator, not an intelligence score. If walking has become slower or less steady, that is a reason to talk to a physician, whatever your IQ. If what you want to know is where you stand on the IQ scale, a walking test cannot tell you; a full test can, and the IQ test on this site or an official IQ test does. Our guide to the reaction time test covers another simple speed measure.

Quick answers

Are walking speed and IQ related? Yes, in one large cohort: full-scale IQ at 45 was associated with composite gait speed at 0.38 (95% CI 0.32 to 0.44).
Do people with a higher IQ walk faster? On average, in that study, yes, but it is an association, and many other things (strength, balance, aging, brain structure) went with gait speed too.
Does walking faster raise your IQ? The study cannot say. It measured both at one time point and had no earlier gait data.
What walking speeds did the studies record? In the Dunedin adults at 45, the mean usual pace was 1.30 meters a second; in pooled adults aged 65 and over, 0.92.
Does early childhood brain health predict adult walking speed? A composite of neurocognitive functioning at age 3, which included standardized tests of intelligence, predicted walking speed at 45 (0.26).
Does slow walking predict dementia? Poor gait predicted any dementia with a pooled hazard ratio of 1.53 in a 12-study meta-analysis, and Alzheimer disease only weakly (1.03).

Sources

Rasmussen, L.J.H. et al. (2019): Association of neurocognitive and physical function with gait speed in midlife, JAMA Network Open, 2(10), e1913123.
Studenski, S. et al. (2011): Gait speed and survival in older adults, JAMA, 305(1), 50-58.
Beauchet, O. et al. (2016): Poor gait performance and prediction of dementia: results from a meta-analysis, Journal of the American Medical Directors Association, 17(6), 482-490.

Share this article

Know someone who keeps seeing these numbers quoted without the scale they were measured on? Send it to them.

Keep reading

All articles →

Bar chart of aerobic exercise effects on four executive function skills, showing real improvements in inhibitory control, working memory and cognitive flexibility, and no significant change in planning

Mind & Everyday Life

Does Exercise Raise Your IQ?

Aerobic exercise has some of the best-replicated cognitive evidence of any lifestyle factor, but not for a full IQ score. What reliably improves is executive function: working memory, inhibitory control, cognitive flexibility. One related skill, planning, barely moves at all.

September 4, 20266 min read

Understanding IQ

Processing Speed and IQ

September 1, 20266 min read

Taking a Test

Can an IQ Test Detect Dementia?

September 22, 20266 min read

Executive Function vs IQ

Posted on October 5, 2026 by Mahesh

Understanding IQ

Executive Function vs IQ: How They Relate and Where They Differ

Executive function and IQ overlap but are not the same thing. Of the three executive functions, only updating working memory was highly correlated with intelligence in one study of young adults. See the three functions, a twin study, how reliable the tasks are and why training claims fall short.

By Alden Cross, Intelligence Research WriterPublished October 5, 2026 · 8 min readReviewed by Aarav Vardhan, Editor-in-Chief

Three finding cards about executive function and IQ from Friedman and colleagues' 2006 study of young adults. Updating, replacing what you hold in mind: highly correlated with fluid and crystallized intelligence and with WAIS IQ, a strong link. Inhibiting, holding back the automatic answer: a small, not significant relation to intelligence once the three functions were modeled together, a weak link. Shifting, switching between rules or tasks: a small, not significant relation, a weak link.

Executive function vs IQ: the two overlap, but they are not the same thing. Executive functions are control abilities: updating what you hold in mind, holding back an automatic response and switching between tasks or rules. In one study of young adults, only the first of those three, updating working memory, was highly correlated with intelligence; inhibiting and shifting showed small relations that were not significant. In a separate analysis of 137 college students the three functions were moderately correlated with one another but clearly separable, and a later twin study found that what they share is a highly heritable common factor that goes beyond general intelligence. This guide explains the three functions, how each relates to IQ, how heritable and reliable they are, and why training claims have not held up.

What it is: executive functions are the cognitive control processes that regulate thought and action. The standard model has three: shifting, updating and inhibiting (Miyake et al., 2000).
Unity and diversity: in 137 college students the three were moderately correlated but clearly separable.
The IQ link: updating was highly correlated with fluid and crystallized intelligence and with WAIS IQ; inhibiting and shifting were not (Friedman et al., 2006).
Heritability: a common executive-function factor was 99% heritable in a twin analysis and went beyond general intelligence and perceptual speed (Friedman et al., 2008).
Training: a 2016 review of 87 working-memory training publications found no convincing far transfer against treated controls.

What is executive function?

Executive functions are the cognitive control processes that regulate thought and action. In the model most of this research uses, set out by Miyake and colleagues in 2000, there are three: shifting between mental sets, updating and monitoring information in working memory, and inhibiting prepotent, meaning dominant and automatic, responses. Each is measured with simple laboratory tasks. The table lists the tasks in the battery of the same lab’s later twin study (Friedman et al., 2008).

The three executive functions and example tasks
Function	What it asks of you	Example tasks in that battery
Inhibiting	Hold back the dominant response	Antisaccade, stop-signal, Stroop
Updating	Replace what you hold in mind with new information	Keep track, letter memory, spatial 2-back
Shifting	Switch between task rules	Color-shape and category-switch tasks (three tasks in all)

Two of those tasks have guides of their own on this site: our news guide to the Stroop test covers inhibition, and the dual n-back task guide covers an updating task and the claim that training it raises IQ.

Are the three executive functions the same thing?

Not quite. Miyake and colleagues (2000) tested 137 college students on simple tasks meant to tap each function, plus several frequently used executive tasks: the Wisconsin Card Sorting Test, the Tower of Hanoi, random number generation, operation span and dual tasking. Confirmatory factor analysis showed that the three functions are moderately correlated with one another but clearly separable. The functions also contributed differently to the complex tasks. Card sorting related most strongly to shifting, the Tower of Hanoi to inhibition, random number generation to inhibition and updating, and operation span to updating. The authors drew the lesson that executive functions show both unity and diversity.

Two results stand out. The Tower of Hanoi puzzle, often treated as a planning task, related most strongly to inhibition. And dual-task performance was not related to any of the three functions, even though doing two things at once is often treated as an executive demand; see our guide to multitasking and IQ.

Which executive functions track IQ?

Friedman and colleagues (2006) examined how fluid intelligence, crystallized intelligence and WAIS IQ relate to the three functions in young adults. Updating was highly correlated with the intelligence measures; inhibiting and shifting were not. In structural equation models that controlled for the correlations among the three functions, updating remained strongly related to intelligence, while the relations of inhibiting and shifting to intelligence were small and not significant. The authors concluded that current intelligence measures do not equally assess the wide range of executive control abilities that many “intelligent” behaviors probably require.

The updating result fits a wider pattern about working memory and reasoning. Schmiedek, Lövdén and Lindenberger (2014) gave nine working-memory tasks to 101 younger and 103 older adults and found that a working-memory factor predicted 71% of the variance in a reasoning factor among younger adults (83% among older adults). Kane and colleagues (2007) found that the n-back task and a verbal span task correlated only weakly, yet each accounted for independent variance in Raven’s matrices. Our news guide to working memory and reasoning and our guide to fluid vs crystallized intelligence cover the neighboring ideas. These are single studies, mostly of young adults, and we did not find a meta-analysis that pools the relations of all three functions with IQ.

How heritable is executive function?

In a multivariate twin study, Friedman and colleagues (2008) modeled the three functions as latent variables. They concluded that the functions are correlated because they are influenced by a highly heritable (99%) common factor that goes beyond general intelligence or perceptual speed, and that they are separable because of additional genetic influences unique to particular functions. That combination, they wrote, places executive functions among the most heritable psychological traits.

Heritability is a statement about variation in a particular sample and population, not about any one person’s fate, and it does not say what a changed environment could do; our guide to what twin studies of IQ show explains how to read such figures.

Your own number

Where would your own score land?

Find your IQ score now! →

Secure & encryptedInstant results10–20 minutes

How reliable are executive function tasks?

Less than their fame suggests. Hedge, Powell and Sumner (2018) measured the test-retest reliability of seven classic tasks, including the Stroop, in three studies. Reliabilities ranged from 0 to .82 and were, in their words, surprisingly low for most tasks. The Stroop reaction-time cost came out at .60 and .66 in two studies. Their “reliability paradox” says why: a task becomes a classic because nearly everyone shows the effect, which leaves little difference between people to measure on a retest. Studies like those above use several tasks per function and model what the tasks share, which is the usual way around the problem, and a single online result should not be read as a trait.

Can you train executive function to raise IQ?

The strongest test is the 2016 meta-analysis by Melby-Lervåg, Redick and Hulme: 87 publications and 145 experimental comparisons of working-memory training. Training produced reliable improvements on measures of intermediate transfer (verbal and visuospatial working memory). For far transfer (nonverbal ability, verbal ability, word decoding, reading comprehension, arithmetic) there was no convincing evidence of improvement when training was compared with a treated control condition, and the size of the working-memory gain did not predict the size of any far-transfer effect. Our dual n-back guide covers the best-known training task, and our review of brain-training apps covers the wider market.

What does this mean for IQ tests and for you?

An IQ score summarizes many abilities, and executive control is only partly in it. The twin and factor studies above suggest that updating, the working-memory side, travels with intelligence, while inhibiting and shifting are separate abilities that an IQ score does not track closely. A high score on one therefore does not guarantee a high score on the other. The working-memory side shows up in subtests such as digit span, and the speed side in the processing speed index. For ADHD, where executive control is often discussed, see our guides to ADHD and IQ and high IQ and ADHD.

A real-world window on executive load is walking while doing something else. In a New Zealand cohort of 904 adults, adding a spoken task cut average walking speed from 1.30 to 1.16 meters a second, about 11%; our guide to walking speed and IQ covers what that cohort showed. If what you want is a score on the IQ scale, take our IQ test or read how an official IQ test works.

Quick answers

What is executive function? The cognitive control processes that regulate thought and action. The standard model has three: shifting, updating and inhibiting.
Is executive function the same as IQ? No. In one study of young adults, updating was highly correlated with intelligence while inhibiting and shifting were not.
Which executive function is most related to intelligence? Updating working memory, in that study.
Are executive functions one ability or several? Both. In 137 college students they were moderately correlated but clearly separable.
Is executive function genetic? A twin study estimated a common executive-function factor at 99% heritable, going beyond general intelligence and perceptual speed. That describes a sample, not an individual.
Can brain training improve executive function and IQ? A 2016 review of 87 publications found no convincing far transfer against treated controls.
Is the Stroop test a good measure of executive function? It is a classic inhibition task, but its reaction-time cost had test-retest reliabilities of .60 and .66 in two studies.

Sources

Miyake, A. et al. (2000): The unity and diversity of executive functions and their contributions to complex “frontal lobe” tasks: a latent variable analysis, Cognitive Psychology, 41(1), 49-100.
Friedman, N.P. et al. (2006): Not all executive functions are related to intelligence, Psychological Science, 17(2), 172-179.
Friedman, N.P. et al. (2008): Individual differences in executive functions are almost entirely genetic in origin, Journal of Experimental Psychology: General, 137(2), 201-225.
Schmiedek, F., Lövdén, M. and Lindenberger, U. (2014): A task is a task is a task: putting complex span, n-back, and other working memory indicators in psychometric context, Frontiers in Psychology, 5, 1475.
Kane, M.J., Conway, A.R.A., Miura, T.K. and Colflesh, G.J.H. (2007): Working memory, attention control, and the n-back task: a question of construct validity, Journal of Experimental Psychology: Learning, Memory, and Cognition, 33(3), 615-622.
Hedge, C., Powell, G. and Sumner, P. (2018): The reliability paradox: why robust cognitive tasks do not produce reliable individual differences, Behavior Research Methods, 50(3), 1166-1186.
Melby-Lervåg, M., Redick, T.S. and Hulme, C. (2016): Working memory training does not improve performance on measures of intelligence or other measures of “far transfer”: evidence from a meta-analytic review, Perspectives on Psychological Science, 11(4), 512-534.
Rasmussen, L.J.H. et al. (2019): Association of neurocognitive and physical function with gait speed in midlife, JAMA Network Open, 2(10), e1913123 (the dual-task walking speeds).

Share this article

Know someone who keeps seeing these numbers quoted without the scale they were measured on? Send it to them.

Keep reading

All articles →

Chart of fluid and crystallized intelligence across the adult lifespan, showing fluid reasoning peaking in the mid-twenties and declining steadily while crystallized knowledge keeps rising into the sixties

Understanding IQ

Fluid vs Crystallized Intelligence

Fluid intelligence is what you use on a problem you have never seen before. Crystallized intelligence is everything you have already banked. The distinction has survived eighty years of factor analysis because the two halves genuinely behave differently, and the clearest proof is what happens to each one as you get older.

September 7, 20266 min read

Understanding IQ

Memory and IQ

The world’s best memory athletes can memorize 71 shuffled words in twenty minutes. Their IQ scores are not measurably different from a matched control group who could only manage 40. If a spectacular memory is not the same thing as a high IQ, what actually is the difference?

September 9, 20266 min read

Understanding IQ

Processing Speed and IQ

September 1, 20266 min read

Reaction Time Test

Posted on September 30, 2026 by Mahesh

Understanding IQ

Reaction Time Test: How to Measure It, Why Online Scores Run Slow and What Changes It

A reaction time test depends on the equipment: 203 ms with a falling rod, 231 ms on a calibrated computer, 273 ms online. See the ruler drop conversion table, why online scores run slow, driving and sprint reaction times, and what really changes your speed.

By Arjun Vardhan, News & Research WriterPublished September 30, 2026 · 12 min readReviewed by Aarav Vardhan, Editor-in-Chief

A bar chart of average simple reaction time in milliseconds from six instruments. A falling rod gave 203 in college football players, a calibrated computer 213 after removing hardware delay and 231 as recorded, a computerized test in the same football players 268, the Human Benchmark online median 273 and a keypad device 290 at age 30.

A reaction time test measures the delay between a signal and your response, and the result depends on the equipment as much as on you. A ruler drop test needs only a ruler: catching it after 20 cm means about 202 ms of fall, though that includes closing your fingers. Online tests add lag. Robot-pressed keys on realistic browser set-ups lagged by about 80 ms on average, and college football players averaged 203 ms with a falling rod but 268 ms on a computer. Real-world reactions are slower than a lab click: drivers need about 0.7 second when they expect a signal and 1.5 seconds for a surprise. Sleep loss slows responses, and simple reaction time barely trains. For average reaction times by age and the link to IQ, see our news guide to average reaction time by age; this guide covers how to measure reaction time, why online scores run slow and what changes it.

Ruler drop: 10 cm is about 143 ms, 20 cm about 202 ms and 30 cm about 247 ms (our calculation); the simple ruler drop had poor test-retest reliability in older adults (0.57).
Online lag: web set-ups added about 80 ms on average in robot tests, and at least 30 ms on all 19 computers in an earlier check.
Same people, two tests: 203 ms with a falling rod and 268 ms on a computer in college football players.
Driving: about 0.7 second expected, 1.25 seconds unexpected, 1.5 seconds for a surprise; road design uses 2.5 seconds.
Sleep: after 17 to 19 hours awake, response speeds were “up to 50% slower for some tests.”
Practice: simple reaction time did not differ by sport or skill in baseball, tennis and non-athlete groups.

How does a reaction time test work?

Simple reaction time is one stimulus and one response, such as pressing a button when a light comes on. Choice reaction time adds decisions, such as pressing the button that matches one of four lights, and takes about twice as long. A careful lab test looks like this: in a 2015 study of 1,469 adults aged 18 to 65 in New Zealand, people sat 0.7 m from a monitor, practiced for 20 trials, then did 120 test trials with a random wait of 1,000 to 1,800 ms before each signal, using a gaming mouse (Woods et al. 2015). Responses outside a window of 110 to 1,000 ms did not count. The average was 231 ms, or 213 ms after subtracting 17.8 ms of measured screen and mouse delay.

Three features of a good test come straight from those numbers. A random wait matters, because reactions were about 15% slower after the shortest wait than the longest, a gap of about 28 ms. Fatigue matters within a sitting: mean reaction time rose from 228.1 ms in the first block of 20 trials to 237.2 ms in the sixth. And one click is noisy, since a person’s own trial-to-trial standard deviation was about 40 ms on the calibrated test and 72 ms on a keypad device at age 30, so several trials should be averaged. Human Benchmark itself tells users to perform “at least 5 clicks.”

The ruler drop test: what the distance means

The ruler drop test needs no equipment beyond a 30 cm ruler and a friend. The dropper releases it without warning, you catch it, and the distance fallen gives the time from falling-object physics, t = √(2d/g). Our conversion, ignoring air resistance:

Ruler drop distance and the time it represents (our calculation)
Distance fallen	Time	Distance fallen	Time
5 cm	101 ms	30 cm	247 ms
10 cm	143 ms	35 cm	267 ms
15 cm	175 ms	40 cm	286 ms
20 cm	202 ms	45 cm	303 ms
25 cm	226 ms	50 cm	319 ms

Hold the ruler with its zero at the top of your open thumb and finger, and have the dropper vary the moment of release so you cannot time it. The time includes closing your fingers, so it is not a pure reaction time, but it has no screen or input lag. That is why it reads shorter than a computer: in college football players a falling rod averaged 203 ms and a computerized test 268 ms, and the two correlated only .445 in the 68 whose computer tests passed an integrity check (Eckner et al. 2010). In a study of 903 adults in Sri Lanka, 95% caught the ruler at or under 40 cm (about 286 ms of fall time), and the authors proposed more than 40 cm as a screen for alcohol intoxication (Rajapaksha et al. 2023). Do not treat a single drop as precise. In 51 older adults the simple ruler drop had poor test-retest reliability (intraclass correlation 0.57) a week apart, and its correlation with a computer test was .42 (Ferreira et al. 2024).

Why online reaction tests read slower than the lab

Human Benchmark, the best-known online test, says its own median is 273 ms “according to the data collected so far” and its mean 284 ms across more than 81 million clicks, and warns that the test “is affected by the latency of your computer and monitor.” Its test page adds, without naming a source, that “an average human reaction time may fall between 200-250ms” and that your computer “could be adding 10-50ms on top.” Its statistics page says recorded times have gotten slightly slower over the years, “almost certainly due to changes in input / display technology.”

Independent checks agree that hardware matters. Robot-pressed keys on realistic browser set-ups showed response times that “on average lag 80 ms, and extend to 100 ms on some set-ups” (Anwyl-Irvine et al. 2021). An earlier test of 19 computers found that “all systems overestimated response times, by at least 30 ms” (Reimers and Stewart 2015). A comparison of five cognitive tasks found “a fixed additive timing offset” of 37 ms for web technology and 87 ms for recording online, against lab data (Semmelmann and Weigelt 2017). In the calibrated study the screen and mouse alone added 17.8 ms. Because much of the lag looks like a fixed offset, comparing your own score across days on the same device is more informative than comparing it with a stranger’s on another.

Measured average reaction times from different instruments
Measurement	Mean	Who and how
Falling rod, college football players	203 ms	Eckner et al. 2010: 94 US Division I players, rod caught by hand
Calibrated computer test, delay subtracted	213 ms	Woods et al. 2015: 1,469 adults aged 18 to 65 in New Zealand
Calibrated computer test, as recorded	231 ms	Same study, before subtracting 17.8 ms of delay
Computer test, same football players	268 ms	Same study, computerized simple reaction time
Human Benchmark, all visitors	273 ms median; 284 ms mean	Self-selected web visitors, over 81 million clicks, lag included
Keypad device, age 30	290 ms	Der and Deary 2017: 714 people in West Scotland

Reaction time on the road

Lab reaction time is only the first part of stopping a car. Marc Green’s review of driver studies found that when drivers fully expect a brake signal they need about 0.70 to 0.75 second to detect it and move the foot to the brake, about 1.25 seconds for an unexpected but common signal such as brake lights, and roughly 1.5 seconds for a surprise such as an object moving into the path (Green 2000). Green’s own website adds that a standard perception-brake time “cannot and does not exist.” Highway design uses a longer figure: the Iowa Department of Transportation’s manual, following the AASHTO Green Book, specifies a brake reaction time of 2.5 seconds in its stopping-distance formula. At 60 mph (26.8 m/s), our calculation gives 6.7 m covered in 0.25 second, 18.8 m in 0.7 second, 40.2 m in 1.5 seconds and 67.1 m in 2.5 seconds. A meta-analysis of 33 driving studies (94 effect sizes, about 2,000 participants) found that phone-related tasks added a mean of 0.25 second to reaction times, with handheld phones and phones used without holding them causing similar decrements (Caird et al. 2008).

Reaction time in sport

Sprinting sets the fast end. World Athletics counts a reaction under 0.100 second as a false start, a rule that a 2021 paper on hurdlers also cites, and the threshold is disputed: a study of nine sprinters and ex-sprinters found five with mean reaction times under 100 ms in at least one condition, measured as the start of force in the blocks (Pain and Hibbs 2007). That is a force-plate measurement with an athlete already braced, not the reaction of a relaxed finger. Formula 1 quotes are looser still. Valtteri Bottas said his first movement after the lights went out in 2019 took “four-hundredths,” and clarified that this was the paddle, not the car moving; in 2017 Sebastian Vettel said that “normally, the reactions are 0.2s for everyone.” Treat those as driver statements, not measurements. Athletes are not necessarily faster on a simple test: in 82 university students and 17 professional baseball players, and in 94 senior high school students followed for two years, “there were no differences in simple reaction time either for sports experience or for skill levels,” although Go/No-go reaction time improved with hitting practice (Kida et al. 2005).

Your own number

Where would your own score land?

Find your IQ score now! →

Secure & encryptedInstant results10–20 minutes

What changes your reaction time?

Sleep loss: in 39 transport and army volunteers, after 17 to 19 hours awake “performance on some tests was equivalent or worse” than at a blood alcohol level of 0.05%, and “response speeds were up to 50% slower for some tests” (Williamson and Feyer 2000). See our guides to sleep and IQ and sleep loss and test performance.
Fatigue and waiting time: both shift results within a single sitting, as the lab numbers above show.
Practice: simple reaction time barely trains, as the sport data show; choice and Go/No-go reaction times can improve.
Caffeine: small trials suggest it speeds responses through attention, with effects that vary by dose and task; see our guide to caffeine and test performance.
Video games: the evidence is contested. One review says the very act of playing action video games “significantly reduces reaction times without sacrificing accuracy” (Dye et al. 2009), while a 2018 meta-analysis found “small or null overall effect sizes” for cognitive ability (Sala et al. 2018); see video games and IQ.
Age: about 0.55 ms slower per year between 18 and 65 in the calibrated test, with trial-to-trial variability growing after 60.

Age, IQ and what a fast reaction time means

Two questions come up next, and our news guide to average reaction time by age and what it says about IQ answers them in detail. In brief: in the calibrated test, group means rose from about 218 ms at a mean age of 21 to 239 ms at 62, and on a keypad device three Scottish cohorts averaged 290, 318 and 354 ms at ages 30, 50 and 69 (Der and Deary 2017). The same study found simple reaction time correlated between −0.27 and −0.32 with a timed reasoning test, and four-choice reaction time between −0.44 and −0.53. Those are observed correlations, worth about 7% to 28% of shared variance (our calculation), so a click score cannot tell you your IQ. Speed is one strand of a full test; see our guide to processing speed and IQ and the timed IQ test page.

Quick answers

What is a good reaction time? On a computer, roughly 200 to 300 ms is ordinary; the device changes the number by tens of milliseconds, so compare like with like.
How accurate is an online reaction time test? It includes screen, input and browser lag: robot tests found about 80 ms on average, and Human Benchmark says its own scores are affected by latency.
How do you do the ruler drop test? Have a friend drop a ruler without warning, catch it, read the distance and convert it: 20 cm is about 202 ms. Average several drops.
What is the average reaction time when driving? About 0.7 second when expected, 1.25 seconds when unexpected and 1.5 seconds for a surprise (Green 2000); road design uses 2.5 seconds.
Can you improve your reaction time? Simple reaction time barely changes with practice; sleep, attention and the task matter more.
Does reaction time measure IQ? Only weakly: correlations of −0.27 to −0.32 for simple and −0.44 to −0.53 for four-choice reaction time in three Scottish cohorts.

The bottom line

A reaction time score is a reading from an instrument. The same person can score 203 ms with a falling rod, 231 ms on a calibrated computer and close to 300 ms on a phone or keypad, and the difference is mostly the device. Use the ruler drop for a quick comparison with no screen lag, average several trials, test when rested, and compare yourself with yourself on the same device. If you want a measure of reasoning rather than speed, a timed IQ test asks for both under standard conditions, and our digit span guide covers a different short-term-memory measure.

Sources

Woods, D.L., Wyma, J.M., Yund, E.W., Herron, T.J. and Reed, B. (2015): Factors influencing the latency of simple reaction time, Frontiers in Human Neuroscience, 9, 131.
Der, G. and Deary, I.J. (2017): The relationship between intelligence and reaction time varies with age: results from three representative narrow-age age cohorts at 30, 50 and 69 years, Intelligence, 64, 89-97.
Human Benchmark: Reaction time statistics and test page (median 273 ms, mean 284 ms), checked 30 September 2026.
Anwyl-Irvine, A., Dalmaijer, E.S., Hodges, N. and Evershed, J.K. (2021): Realistic precision and accuracy of online experiment platforms, web browsers, and devices, Behavior Research Methods, 53(4), 1407-1425.
Reimers, S. and Stewart, N. (2015): Presentation and response timing accuracy in Adobe Flash and HTML5/JavaScript web experiments, Behavior Research Methods, 47(2), 309-327.
Semmelmann, K. and Weigelt, S. (2017): Online psychophysics: reaction time effects in cognitive experiments, Behavior Research Methods, 49(4), 1241-1260.
Eckner, J.T., Kutcher, J.S. and Richardson, J.K. (2010): Pilot evaluation of a novel clinical test of reaction time in National Collegiate Athletic Association Division I football players, Journal of Athletic Training, 45(4), 327-332.
Rajapaksha, S. et al. (2023): The establishment of a population-specific reference value for the ruler drop test for the clinical assessment of reaction time, Journal of Forensic and Legal Medicine, 96, 102525.
Ferreira, S. et al. (2024): Validity and reliability of a ruler drop test to measure dual-task reaction time, choice reaction time and discrimination reaction time, Aging Clinical and Experimental Research, 36, 61.
Green, M. (2000): “How long does it take to stop?” Methodological analysis of driver perception-brake times, Transportation Human Factors, 2(3), 195-216; and Green’s driver reaction time page.
Iowa Department of Transportation: Design Manual, Chapter 6, Section 6D-1, Sight Distance (brake reaction time of 2.5 s).
Caird, J.K., Willness, C.R., Steel, P. and Scialfa, C. (2008): A meta-analysis of the effects of cell phones on driver performance, Accident Analysis and Prevention, 40(4), 1282-1293.
Pain, M.T.G. and Hibbs, A. (2007): Sprint starts and the minimum auditory reaction time, Journal of Sports Sciences, 25(1), 79-86.
Ntolaptsis, K. and Panoutsakopoulos, V. (2021): Relationship between reaction time, medal winning and performance in the 60 m hurdle indoor event before and after the change of false start rule, Acta Facultatis Educationis Physicae Universitatis Comenianae, 61(1), 72-85.
Motorsport.com: Bottas on his start reaction time (25 October 2019), and the 2017 Austrian Grand Prix report.
Williamson, A.M. and Feyer, A.M. (2000): Moderate sleep deprivation produces impairments in cognitive and motor performance equivalent to legally prescribed levels of alcohol intoxication, Occupational and Environmental Medicine, 57(10), 649-655.
Kida, N., Oda, S. and Matsumura, M. (2005): Intensive baseball practice improves the Go/Nogo reaction time, but not the simple reaction time, Cognitive Brain Research, 22(2), 257-264.
Dye, M.W.G., Green, C.S. and Bavelier, D. (2009): Increasing speed of processing with action video games, Current Directions in Psychological Science, 18(6), 321-326.
Sala, G., Tatlidil, K.S. and Gobet, F. (2018): Video game training does not enhance cognitive ability: a comprehensive meta-analytic investigation, Psychological Bulletin, 144(2), 111-139.

Share this article

Know someone who keeps seeing these numbers quoted without the scale they were measured on? Send it to them.

Keep reading

All articles →

Understanding IQ

Processing Speed and IQ

September 1, 20266 min read

Research & Evidence

Sleep and IQ

This site’s news desk has already covered what one bad night does to a test score on the day it is taken. This is a different and, for children especially, more consequential question: what years of too little sleep do to the cognitive systems an IQ test measures in the first place.

September 11, 20266 min read

Scores & Scales

Digit Span Test

A digit span test asks you to repeat lists of digits; on the WAIS it has three tasks of eight items. Verified adult averages run from 5.3 to 6.9 digits, not seven, and language matters: 8.84 in Mandarin versus 5.77 in English for the same people.

September 30, 202614 min read