Humor and IQ

Posted on October 7, 2026 by Mahesh

Mind & Everyday Life

Humor and IQ: Are Funny People Smarter?

Are funny people smarter? In student studies, the ability to write funny captions correlates about .25 with abstract reasoning and .38 with vocabulary. That is a real but moderate link: a high scorer is somewhat more likely to be funny on demand, but IQ cannot identify funny people. See the odds, the evidence and its limits.

By Alden Cross, Intelligence Research WriterPublished October 7, 2026 · 8 min readReviewed by Aarav Vardhan, Editor-in-Chief

A bar chart of the correlation between humor production ability, the judged funniness of cartoon captions, and two intelligence measures in 400 students (Greengross and Miller 2011). Everyone: vocabulary .38, matrix reasoning .25. Men: vocabulary .44, matrix reasoning .27. Women: vocabulary .31, matrix reasoning .24. All six are statistically significant, and in the full sample vocabulary correlates more strongly with humor than matrix reasoning does.

Humor and IQ are related, but only moderately, and only for one kind of humor. In student studies where judges rated how funny people’s cartoon captions were, funniness correlated about .25 with abstract reasoning and .38 with vocabulary in the largest study we reviewed, which had 400 people. That means smarter people are somewhat more likely to be funny on demand, not that funny people can be identified by their IQ, and not that a good sense of humor in the everyday sense is a sign of high intelligence. This guide explains what was measured, how big the link is in plain odds, and where the evidence stops.

What was measured: humor production, meaning writing something funny on demand, such as a caption for a cartoon that has none, scored by judges. It is not the same as enjoying jokes or being thought of as having a good sense of humor.
How strong: .25 with a matrix reasoning test and .38 with a vocabulary test in a 2011 study of 400 students; a 2018 study of 270 adults estimated a large effect of general intelligence on humor production (.51 in a latent-variable model).
In plain odds (our calculation): with a correlation of .38, someone in the top quarter on vocabulary has about a 42% chance of being in the top quarter for humor production, against 25% by chance.
Verbal over abstract: vocabulary predicted humor production better than matrix reasoning, and in the 2018 study general intelligence and idea-retrieval ability carried the effect while fluid reasoning had no separate effect beyond general intelligence.
Limits: student samples, subjective judging, and correlation rather than cause. A 2025 twin study found no additive genetic effect on caption-writing scores.

Humor draws on abilities we describe elsewhere: the word-finding fluency measured in the verbal fluency test, the verbal side of verbal and non-verbal IQ scores, and the idea generation covered in creativity and IQ.

What do studies of humor and intelligence actually measure?

Almost all of the evidence concerns humor production: coming up with something funny on the spot. Researchers hand people a cartoon with the caption removed and ask for a new one, or ask for a witty reply to an absurd question, and then have judges rate how funny the answers are. Feingold and Mazzella (1991), as summarized by Greengross and Miller, did this with 147 adults in three groups, including Harvard graduate students and volunteers recruited in Central Park, and found correlations of .31 to .52 between a vocabulary test and rater-judged humor tasks.

Howrigan and MacDonald (2008) used 185 college students and tasks such as funny responses to emails, mock descriptions of stereotyped characters and funny drawings. Greengross and Miller report correlations of .12 to .23 with the Raven matrices. In the authors’ analysis general intelligence predicted judged humor even after the Big Five personality traits were counted, and extraversion also predicted it, to a lesser degree.

How strong is the link between humor and IQ?

The largest of these studies is Greengross and Miller (2011). They gave 400 university students (200 men and 200 women) a matrix reasoning test (Raven’s Advanced Progressive Matrices), a 46-item vocabulary test and a humor task in which judges rated the funniness of captions the students wrote for three cartoons. Across all 400, humor ability correlated .38 with vocabulary and .25 with the matrices, and the difference between those two correlations was statistically significant. For men the figures were .44 and .27; for women .31 and .24.

Four studies of humor production and intelligence
Study	Sample	Intelligence measure	Result
Feingold and Mazzella (1991)	147 adults in three groups	Vocabulary	r = .31 to .52 (as summarized by Greengross and Miller)
Howrigan and MacDonald (2008)	185 college students	Raven matrices (general intelligence)	r = .12 to .23 (as summarized by Greengross and Miller); general intelligence predicted judged humor beyond the Big Five
Greengross and Miller (2011)	400 university students	Vocabulary and Raven matrices	r = .38 and .25
Christensen et al. (2018)	270 young adults	Fluid reasoning, vocabulary and retrieval factors	General intelligence predicted humor production (standardized effect .51, 95% CI .32 to .70)

Those numbers are real but moderate. A correlation of .38 means vocabulary accounts for about 14% of the differences in judged funniness, and .25 for about 6%; most of what makes one person funnier than another is something else. In practical terms (our calculation, assuming both measures follow a normal curve), a person in the top quarter on vocabulary has about a 42% chance of being in the top quarter for humor production, against 25% for a random person, and about a 62% chance of being above the median. With the matrix test (.25) those odds are about 36% and 58%. That is a tilt, not a sorting rule.

Which kind of intelligence matters most for being funny?

Vocabulary mattered more than abstract reasoning in the 2011 study, which Greengross and Miller took to be unsurprising for tasks that are verbal. A 2018 study by Christensen and colleagues refined the picture. Using the Cattell-Horn-Carroll model, they measured fluid reasoning, vocabulary knowledge and broad retrieval ability (how fluently a person can generate ideas and words) in 270 young adults, along with a battery of humor production tasks. All three specific factors correlated with humor, and a model with general intelligence on top found a large effect of g on humor ability (.51). In a bifactor model, humor was predicted mainly by g and by retrieval ability, not by fluid reasoning once those were counted.

That fits the intuition that generating many candidate words and ideas quickly is a large part of being funny on demand, though the studies do not test that directly. Our guide to CHC theory explains those factors, fluid versus crystallized intelligence covers the reasoning and knowledge sides, and the verbal fluency test measures a close relative of retrieval ability.

Is a good sense of humor a sign of high IQ?

Partly, with large caveats. The studies measure the ability to produce humor, not the experience of enjoying it, and not the way friends describe someone as having a good sense of humor. A moderate correlation across a group says nothing certain about any one person: plenty of people with ordinary scores are very funny, and plenty of high scorers are not. The idea behind the research is that humor production may be an honest indicator of intelligence, and the authors tie it to mating success; correlations like these are consistent with that idea but cannot test it. We did not find, in the sources used for this page, a comparably large study of humor appreciation and IQ, so we do not claim one way or the other.

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

Is humor inherited? What a 2025 twin study adds

A 2025 study of 448 pairs of identical and 196 pairs of fraternal twins from the UK twin registry (median age 66) tested whether humor production runs in families. The twins rated their own humor ability, rated their co-twin, and wrote funny captions for captionless cartoons, and they also completed a short cognitive ability test. Self-rated humor ability showed both genetic and non-shared environmental influence. The objectively rated caption-writing score showed no evidence of additive genetic effects; individual differences were shaped by shared and non-shared environment, although the authors say a small genetic effect cannot be ruled out. They conclude that humor production may be more complex and harder to assess than other cognitive abilities. To see how twin and sibling designs work, our guide to whether siblings have similar IQs explains the logic with real numbers.

Where does the evidence stop?

Five limits are worth keeping in mind:

Samples are mostly university students or volunteers, which narrows the range of scores.
Funniness is judged by raters with their own taste, and a caption task is one narrow slice of humor.
Correlations show a relationship, not a cause: a larger vocabulary may help produce wittier captions, or both may reflect something else, such as verbal experience.
Results vary between studies, from .12 to .52, because measures and samples differ.
The studies described here used English-language tasks, so vocabulary here means English vocabulary.

If you are curious how your own reasoning compares, you can take the IQ test. Humor is not scored in it, and these studies do not support the reverse inference that a lower reasoning score means a smaller sense of humor.

Quick answers

Are funny people smarter? On average, modestly. In the largest study we reviewed, the ability to write funny captions correlated .25 with abstract reasoning and .38 with vocabulary.
Does a sense of humor mean a high IQ? Not reliably. The studies measure producing humor on demand, and the link is moderate: a high scorer has a better than average chance of being funny, but IQ does not identify funny people.
Is humor related to verbal intelligence? More than to abstract reasoning: .38 versus .25 in the 2011 study, and idea-retrieval ability mattered in the 2018 study.
Is being funny genetic? A 2025 twin study found genetic influence on self-rated humor ability but none detectable on an objectively rated caption task.
Do smart people appreciate jokes more? The sources used here do not answer that; they concern producing humor.

Sources

Greengross, G. and Miller, G. (2011): Humor ability reveals intelligence, predicts mating success, and is higher in males, Intelligence, 39(4), 188-192.
Howrigan, D.P. and MacDonald, K.B. (2008): Humor as a mental fitness indicator, Evolutionary Psychology, 6(4).
Feingold, A. and Mazzella, R. (1991): Psychometric intelligence and verbal humor ability, Personality and Individual Differences, 12(5), 427-435.
Christensen, A.P., Silvia, P.J., Nusbaum, E.C. and Beaty, R.E. (2018): Clever people: intelligence and humor production ability, Psychology of Aesthetics, Creativity, and the Arts, 12(2), 136-143.
Greengross, G. et al. (2025): Heritability of humor production ability: a twin study, Twin Research and Human Genetics, 28(3), 265-272.
Our calculations: the odds of top-quarter and above-median humor production are derived here from the reported correlations, assuming both measures follow a normal curve. They are not reported in the sources.

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 →

A bar chart of average FAS verbal fluency scores, total words in one minute for each of the letters F, A and S, among Brazilian adults aged 60 to 69 by years of schooling (Machado et al., 2009). 1 to 3 years of education: 18.3 words (14 people). 4 to 7 years: 26.1 (31 people). 8 to 11 years: 31.9 (36 people). 12 or more years: 38.7 (54 people). Education moves the score by more than 20 words.

Taking a Test

Verbal Fluency Test

A verbal fluency test asks you to name as many words as you can in 60 seconds, by letter (F-A-S) or category (animals). See how it is scored, why education can move the score by about 20 words, what the task really measures and what a low count can and cannot tell you.

October 6, 20269 min read

Understanding IQ

Creativity and IQ

Creativity and IQ correlate only weakly. The largest meta-analysis ever run on the question, covering 45,880 people, found no significant threshold at IQ 120, just a modest overall link of about .17 to .25. See what the strongest evidence shows, and why the classic claim does not hold up.

September 28, 20268 min read

Scores & Scales

Verbal vs Non-Verbal IQ Scores

Two people can hold the same IQ number and have completely different profiles, and a split between the verbal and non-verbal indexes is far commoner than most test-takers expect. Here is what base-rate tables and compounding measurement error say about how large a gap has to be before it means anything.

August 29, 20267 min read

Do Siblings Have Similar IQs?

Posted on October 7, 2026 by Mahesh

Understanding IQ

Do Siblings Have Similar IQs? The Typical Gap, in Numbers

Do siblings have similar IQs? Somewhat. Across 69 correlations the average is .47, so two full siblings typically differ by about 12 IQ points, and about one pair in five differs by 20 or more. See the odds of every gap, what to expect from the sibling of a high scorer, and what pushes siblings apart.

By Leon Falkner, Data & Research AnalystPublished October 7, 2026 · 8 min readReviewed by Aarav Vardhan, Editor-in-Chief

A bar chart of the typical gap, in IQ points, between the scores of two people by how they are related, calculated from the average IQ correlation for each pair (Bouchard and McGue 1981) and a standard deviation of 15. Identical twins reared together (correlation .86): about 6 points. Fraternal twins (.60): about 11. Full siblings reared together (.47): about 12. Unrelated people (no correlation): about 17. Full siblings are much closer to strangers than to identical twins. The gaps are our calculation.

Do siblings have similar IQs? Somewhat, but far from identical. Across 69 correlations from 26,473 sibling pairs, the average correlation between the IQ scores of full siblings raised together is .47. That is moderate: siblings tend to land on the same side of average more often than not, yet two of them typically differ by about 12 IQ points, and roughly one pair in five differs by 20 points or more. That is not a sign that one was tested badly or raised differently; it is what a correlation of that size produces on its own. This guide turns the correlation into numbers you can use: the odds of each size of gap, what to expect from the brother or sister of a high or low scorer, and what actually pushes siblings apart.

The number: the weighted average correlation for siblings reared together is .47 (Bouchard and McGue, 1981, a review of 111 studies). Individual correlations ranged from .13 to .90.
The typical gap (our calculation): about 12 points on a scale with a standard deviation of 15. Half of all sibling pairs differ by 10 points or more, about one pair in five by 20 or more, and about one in 19 by 30 or more.
Twins for comparison: identical twins reared together correlate .86 (typical gap about 6 points) and fraternal twins .60 (about 11).
Regression to the mean: if one child scores 130, a sibling is expected to score about 114, with an 11% chance of 130 or more and a 14% chance of below 100.
Why not closer: siblings share about half of their genetic variants, not all of them; they have different experiences; and every score carries a margin of error of 2 to 3 points.

The question sits next to topics we cover elsewhere: whether IQ is genetic, whether intelligence comes from the mother or the father, what twin studies show, and whether birth order matters. Here we focus on the practical side: how far apart two siblings usually land.

How similar are siblings on IQ tests?

The standard reference is a 1981 review in Science by Bouchard and McGue, which gathered 526 familial correlations from 111 studies covering 113,942 pairings. For siblings reared together the weighted average correlation was .47, based on 69 values from 26,473 pairs. The authors noted that this is close to the .50 a simple polygenic model predicts, and that same-sex and opposite-sex siblings gave almost identical averages (.48 and .49). But the individual studies were far from uniform, ranging from .13 to .90, so .47 is a center of gravity rather than a law.

The table below puts siblings next to other pairs from the same review. The “typical gap” column is our calculation from each correlation, assuming scores that follow a normal curve with a standard deviation of 15: the standard deviation of the difference between two people is 15 times the square root of 2 times (1 minus the correlation), and the average gap is about 0.8 times that.

Typical gap by relationship (our calculation from the Bouchard and McGue averages)
Pair	Correlation	Typical gap (points)	Pairs 20+ points apart
Identical twins reared together	.86	about 6	1%
Fraternal twins reared together	.60	about 11	14%
Full siblings reared together	.47	about 12	20%
Siblings reared apart (two correlations only)	.24	about 15	28%
Unrelated people	0	about 17	35%

Read across the table and the point is clear: full siblings are much closer to strangers than to identical twins. A sibling pair is about 12 points apart on average; two people picked at random from the same population average about 17.

How big is the typical IQ gap between siblings?

Because a correlation of .47 leaves most of the variation unexplained, the gaps are wide. The next table shows the share of full-sibling pairs that differ by at least a given number of points, under the same assumptions as above.

How often two full siblings differ by at least this many points (our calculation)
Gap of at least	Share of sibling pairs	Roughly
5 points	75%	3 pairs in 4
10 points	52%	1 pair in 2
15 points	33%	1 pair in 3
20 points	20%	1 pair in 5
30 points	5%	1 pair in 19

Part of every gap is measurement noise, but only a small part. On a well-built test the standard error is 2 to 3 points (see the margin of error explained), which adds roughly 3.5 points of spread to the difference between two single scores, against about 15 for the gap itself. Most of the distance between siblings is real difference in what they scored.

Larger families widen the spread. Using the within-family standard deviation implied by a .47 correlation (about 11 points), the expected distance between the highest and lowest scorer is about 12 points in a family of two, 18 in a family of three, 22 in a family of four and 25 in a family of five. In a family with several children, a 20-point range is typical rather than remarkable.

What IQ should you expect a brother or sister to have?

If one child’s score is known, a sibling’s expected score sits part of the way back toward 100, in proportion to the correlation. This is regression to the mean, which our page on regression to the mean explains in general. With a correlation of .47, the expected sibling score is 100 plus .47 times the first child’s distance from 100, and the sibling’s own spread around that expectation is about 13 points.

Expected sibling scores by first child score (our calculation)
First child scores	Sibling expected to score	Sibling 100 or more	Sibling 115 or more	Sibling 130 or more
70	86	14%	1%	under 0.1%
85	93	30%	5%	0.3%
100	100	50%	13%	1%
115	107	70%	27%	4%
130	114	86%	47%	11%
145	121	94%	68%	25%

A child who scores 130 will, on average, have a sibling around 114: probably above average (an 86% chance of 100 or more), about even odds of 115 or more, and roughly one chance in nine of reaching 130 as well. That is why a gifted child often has an above-average sibling and sometimes an average one. See signs of a gifted child and gifted cutoff scores for what the label requires, and how rare a high score is for the rarity of each level. In the other direction, a child scoring 85 has a sibling expected around 93, with about a 30% chance of 100 or more.

These are illustrations from one average correlation. Real families vary, the relationship may not be perfectly linear at the extremes, and a model cannot tell you anything about a particular child. To know a child’s score, test the child; see IQ testing for children.

Your own number

Where would your own score land?

Find your IQ score now! →

Secure & encryptedInstant results10–20 minutes

Why are siblings not more alike?

Genes are one reason. Full siblings share about half of their genetic variants on average, the same as fraternal twins, and each child receives a different shuffle. The balance between genes and shared home environment also shifts with age: in a sample of 11,000 twin pairs, the heritability of general cognitive ability rose from 41% at age 9 to 55% at 12 and 66% at 17 (Haworth and colleagues, 2010), and Plomin and Deary (2015) put the range from about 20% in infancy to perhaps 80% in later adulthood. Our guide to whether IQ is genetic explains what a heritability figure does and does not mean.

The shared home matters too, so siblings are not simply a genetic lottery. In a Norwegian study of more than 334,000 pairs of brothers tested at military conscription, the correlation between brothers decreased as the age difference between them increased (Sundet and colleagues, 2008), a pattern that points to family environment effects in young adults. A Swedish study compared full brothers reared in different homes: those raised by adoptive parents scored 4.4 IQ points higher than their brothers raised by a biological parent (Kendler and colleagues, 2015), and each extra unit of the rearing parents’ education went with 1.7 IQ units. The same design with half-brothers gave 3.2 points and 1.9 units. Our guide to adoption and IQ covers what that does and does not imply.

Then there are the things that happen to one child and not the other. Different friends and teachers, a head injury or a long illness (concussion and IQ), a premature birth (premature birth and IQ), or simply a bad stretch: depression and IQ shows how low mood can pull down the speed- and attention-heavy parts of a score for a while. Finally, the yardstick can differ. If siblings sat different tests, or the same test with different norms, part of the gap belongs to the instrument rather than the child.

Is the firstborn smarter?

On average, slightly, but nowhere near enough to explain a typical gap. Within-family analyses find a firstborn advantage of roughly one to two points, and the exact size is disputed; our piece does birth order affect IQ sets out the evidence. Against a typical sibling gap of 12 points, birth order is a small effect, and it is no larger than the error on a single test.

Quick answers

Do siblings have the same IQ? No. Their scores correlate about .47 on average, which leaves a typical gap of about 12 points.
Can siblings have very different IQs? Yes, and often. About one pair in five differs by 20 points or more, and about one in 19 by 30 or more.
Is it normal for one sibling to score much higher? Yes. A gap of 15 points happens in about one pair in three, and in a family of four the highest and lowest scorers are about 22 points apart on average.
Do identical twins have the same IQ? Not exactly. Their correlation is about .86, which gives a typical gap of about 6 points.
If one child scores 130, will a sibling? The chance is about 11%; the expected sibling score is about 114.
Does birth order matter? Within families, roughly one to two points for firstborns, much smaller than the typical gap.

Sources

Bouchard, T.J. and McGue, M. (1981): Familial studies of intelligence: a review, Science, 212(4498), 1055-1059.
Sundet, J.M., Eriksen, W. and Tambs, K. (2008): Intelligence correlations between brothers decrease with increasing age difference: evidence for shared environmental effects in young adults, Psychological Science, 19(9), 843-847.
Kendler, K.S. et al. (2015): Family environment and the malleability of cognitive ability: a Swedish national home-reared and adopted-away cosibling control study, Proceedings of the National Academy of Sciences, 112(15), 4612-4617.
Haworth, C.M.A. et al. (2010): The heritability of general cognitive ability increases linearly from childhood to young adulthood, Molecular Psychiatry, 15(11), 1112-1120.
Plomin, R. and Deary, I.J. (2015): Genetics and intelligence differences: five special findings, Molecular Psychiatry, 20(1), 98-108.
Our calculations: every gap, odds and expected-score figure above is derived here from the .47 correlation, assuming normally distributed scores with a standard deviation of 15. None of them is reported in the sources.

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 showing how the heritability of IQ rises from childhood to adulthood

Research & Evidence

Is IQ Genetic?

Heritability estimates for IQ rise from roughly 0.4 in childhood to 0.7 or higher in adulthood, which sounds decisive until you know what heritability measures. Here is what twin and adoption studies actually show, and the four conclusions people draw from them that the statistic does not support.

August 27, 20267 min read

Research & Evidence

Is Intelligence Inherited From Mother or Father?

Is intelligence inherited from mother or father? From both. A review of 111 family studies found mothers and fathers resemble their children in IQ equally (.41 each). See where the X-chromosome claim came from, why it fails, and how much IQ parents pass on.

September 26, 202610 min read

Research & Evidence

Adoption and IQ

Adoption can raise average IQ by about 3 to 4 points in a large Swedish sibling study and by 8 to 20 in a French study of neglected children, yet adoptees come to resemble their birth parents. See what adoption and twin studies show, and what they cannot.

September 29, 202612 min read

Bipolar Disorder and IQ

Posted on October 7, 2026 by Mahesh

Mind & Everyday Life

Bipolar Disorder and IQ: Are Bipolar People Smarter?

Are bipolar people smarter? Not on average, but the evidence is more interesting than yes or no. Swedish data on a million men show a reversed-J pattern, excellent school grades carried nearly four times the risk in males, and after diagnosis cognitive scores run about 0.6 standard deviations lower.

By Alden Cross, Intelligence Research WriterPublished October 7, 2026 · 10 min readReviewed by Aarav Vardhan, Editor-in-Chief

A bar chart of how far below comparison groups people with bipolar disorder score on cognitive measures, in standard deviations with an IQ-style point equivalent. Before onset, measured ahead of time: 0.03, about half a point and not statistically significant (Trotta 2015). Before onset, estimated afterward: 0.15, about 2 points (Trotta 2015). Estimated premorbid IQ in studies of people between episodes: 0.24, about 4 points (Swidzinski 2025). After onset, global cognition: 0.62, about 9 points (Trotta 2015). Between episodes, general cognitive functioning: 0.58, about 9 points (Swidzinski 2025). Differences are small or absent before the illness and moderate after it.

Bipolar disorder and IQ are linked in a more complicated way than the popular claim that bipolar people are smarter, and the evidence does not show higher intelligence on average. What the studies show instead is a mix. A cohort of more than a million Swedish men found that risk fell as IQ rose, with risk rising again at the very top among men with no other psychiatric diagnosis. A school-grades study found nearly four times the risk in males with excellent grades. A UK cohort linked higher childhood IQ to more manic features. And reviews of people who already have the diagnosis find moderate cognitive deficits, even between episodes. This guide lays out each result and what it can and cannot support.

A million men: across 1,049,607 Swedish men, hospital admission for bipolar disorder became less likely as IQ rose. Among men with no other psychiatric diagnosis the pattern was reversed-J: highest risk at the lowest IQ, with risk elevated again at the highest IQ, mainly for verbal or technical ability.
School grades: excellent grades at age 15 to 16 carried a hazard ratio of 3.79 for later bipolar disorder in Sweden, in males only; the poorest grades carried 1.86.
Mixed evidence: a smaller conscript cohort (50,087 men) found no link between IQ and bipolar disorder, and a UK cohort found a small positive correlation (r = .16) between childhood IQ and manic features.
After diagnosis: a 2025 meta-analysis of 75 studies found general cognitive functioning about 0.58 standard deviations lower between episodes.
The bottom line: bipolar disorder is not a marker of genius. High verbal ability shows up as a risk marker in some cohorts, but the average person with the diagnosis scores lower, not higher, on cognitive tests.

This page is a companion to depression and IQ, where the pattern runs the opposite way, and to our overview of high IQ and mental health. Creativity is the other thread people usually bring up, and we cover it in creativity and IQ.

Does a higher IQ raise the risk of bipolar disorder?

The largest study is by Gale and colleagues (2013). They linked IQ measured at military conscription (mean age 18.3) in 1,049,607 Swedish men to national hospital records over an average of 22.6 years. Across all men, the risk of admission with any form of bipolar disorder fell in a stepwise way as IQ rose. But when they restricted the analysis to men with no psychiatric comorbidity, they found a reversed-J shape: men with the lowest IQ had the greatest risk of being admitted with “pure” bipolar disorder, but risk was elevated again among men with the highest IQ, mainly those with the highest verbal or technical ability. Their conclusion was careful: at least in men, high intelligence may be a risk factor, but only for the minority of cases in whom the disorder occurs in a pure form.

Two other studies point to a high-ability link. MacCabe and colleagues (2010) used school grades for everyone who finished compulsory schooling in Sweden between 1988 and 1997 and followed them for hospital admissions up to age 31. Students with excellent grades had a nearly fourfold risk of later bipolar disorder compared with students with average grades (hazard ratio 3.79, 95% confidence interval 2.11 to 6.82), and the association appeared confined to males. Students with the poorest grades were also at moderately increased risk (hazard ratio 1.86). Smith and colleagues (2015) used a UK birth cohort, the Avon Longitudinal Study of Parents and Children. Among 1,881 people, IQ at age 8 correlated at r = .16 with lifetime manic features reported at age 22 to 23 on a screening checklist, which is not a diagnosis. Those in the highest decile of manic features had a mean full-scale IQ of 110 against 101 in the lowest decile, and the link was strongest for verbal IQ.

Not every cohort agrees. Zammit and colleagues (2004) followed 50,087 Swedish conscripts for 27 years and found no association between IQ and bipolar disorder at all, even though lower IQ predicted schizophrenia and severe depression in the same men. The Dunedin cohort in New Zealand (Koenen and colleagues, 2009) found the opposite of its depression result for mania: higher childhood IQ predicted a higher risk of adult mania.

Five studies, one question
Study	Who and what	Result
Gale et al. (2013)	1,049,607 Swedish men, IQ at about 18, hospital admissions over 22.6 years on average	Risk fell as IQ rose overall; reversed-J in men with no other psychiatric diagnosis, with higher risk again at the top, mainly for verbal or technical ability
MacCabe et al. (2010)	All Swedish students finishing compulsory school 1988 to 1997, grades at 15 to 16, admissions to age 31	Excellent grades: hazard ratio 3.79, males only; poorest grades: 1.86
Smith et al. (2015)	1,881 people in a UK birth cohort, IQ at 8, manic features at 22 to 23	r = .16; strongest for verbal IQ
Koenen et al. (2009)	1,037 people in a New Zealand birth cohort, IQ at 7 to 11, diagnoses to age 32	Higher childhood IQ predicted adult mania
Zammit et al. (2004)	50,087 Swedish men, IQ at about 18, 27 years of hospital records	No association between IQ and bipolar disorder

The studies measured different things: an IQ test, school grades, a self-report symptom checklist, hospital admissions, a diagnosis made in an interview. They looked at different ages and outcomes, and several are limited to men or are strongest in men. Two of them point specifically at verbal ability. Even where the association is real, it is modest. An r of .16 means childhood IQ accounts for about 2.5% of the variation in manic features, and bipolar disorder is uncommon, so a fourfold relative risk applies to a small absolute number. Most people with excellent grades or a high IQ never develop it.

Do people with bipolar disorder have lower IQs before the illness starts?

This is the premorbid question, and it is where bipolar disorder and schizophrenia differ most. A 2015 meta-analysis by Trotta and colleagues pooled studies that compared cognitive function in both disorders. People who later developed schizophrenia had a clear premorbid deficit (a standardized mean difference of 0.60 below healthy comparison groups). For bipolar disorder the picture was weaker and depended on how premorbid function was assessed: a small deficit of 0.15 when it was estimated retrospectively, after the illness had begun, and essentially none (0.03, not statistically significant) when it was measured prospectively, before onset. The authors called the evidence on premorbid function in bipolar disorder equivocal.

That is in line with the 2004 Swedish conscript cohort, which found no IQ association with bipolar disorder, although the much larger 2013 cohort did find a gradient, so the evidence is not uniform. Zammit and colleagues took the contrast with schizophrenia to mean that at least some aspects of the neurodevelopmental origins of bipolar disorder may differ from those of the other conditions. The same cohort is described in depression and IQ, where lower IQ predicted severe depression.

What happens to thinking after bipolar disorder starts?

After onset the direction is clear. In the Trotta meta-analysis people with bipolar disorder showed moderate cognitive impairment compared with healthy controls (0.62 standard deviations), and people with schizophrenia a much larger one (1.37). A 2025 meta-analysis by Swidzinski and colleagues looked only at people who were euthymic, meaning mood-stable between episodes. Pooling 95 groups from 75 studies (4,404 people with bipolar disorder and 4,037 healthy controls), they found impairment in general cognitive functioning (0.58 standard deviations lower) and in every domain tested: verbal memory (0.70), executive function (0.69), visuospatial memory (0.68), attention and processing speed (0.64) and working memory (0.61). Estimated premorbid IQ was about 0.24 standard deviations lower in those studies.

On an IQ-style scale with a standard deviation of 15, 0.58 to 0.62 standard deviations would be about 9 points. As with depression, that is an illustration of size, not an IQ loss: these are group averages on cognitive tests that mix many different tasks, and individuals vary widely. More years of education went with slightly less verbal-memory impairment; the other demographic and clinical variables the authors tested were not associated with performance. Our guide to processing speed and IQ and the piece on the digit span test explain two of the abilities involved.

Your own number

Where would your own score land?

Find your IQ score now! →

Secure & encryptedInstant results10–20 minutes

Is bipolar disorder linked to creativity?

This is the other half of the “bipolar genius” story, and the evidence is about occupations, not IQ. Kyaga and colleagues (2011) used Swedish registers to compare people treated in hospital for schizophrenia, bipolar disorder or unipolar depression, and their relatives, with controls. People with bipolar disorder, and the healthy siblings of people with bipolar disorder or schizophrenia, were overrepresented in creative professions. People with unipolar depression, and their siblings, were not. The authors suggested the traits may run together in families. Full siblings share about half of their genetic variants on average, which is part of why they resemble each other on many traits; our guide to whether siblings have similar IQs puts numbers on that for test scores.

Creative occupations are a coarse stand-in for creative ability, and holding one is not the same as being exceptionally intelligent. See creativity and IQ for how the two relate, and humor and IQ for another creative skill that has been measured in the lab: being funny on demand.

What does this mean for you or someone you know?

Three practical points. First, an IQ test cannot detect or rule out bipolar disorder; diagnosis is clinical and rests on a history of manic or hypomanic and depressive episodes. Second, thinking difficulties such as slowed processing or trouble with memory and concentration are well documented even between episodes, and are worth raising with a clinician rather than dismissing. Third, a high score does not signal risk, and a diagnosis does not signal either genius or decline: the group differences described here are averages, and individuals vary widely.

If you are curious about your own reasoning scores, you can take the IQ test; for how a profile of strengths and weaknesses is read in practice, see how to read an IQ test report. If you are worried about mood symptoms, speak to a doctor or mental health professional. In the United States, anyone in crisis can call or text 988 to reach the Suicide & Crisis Lifeline.

Quick answers

Are bipolar people smarter? Not on average. After onset, people with bipolar disorder score lower on cognitive tests (about 0.6 standard deviations in two meta-analyses), and the evidence on IQ before onset is equivocal.
Is bipolar disorder linked to high IQ? Weakly and inconsistently. Risk was elevated at the highest verbal or technical IQ among Swedish men with no other diagnosis, excellent school grades carried nearly four times the risk in males, and childhood IQ correlated .16 with manic features in a UK cohort; one smaller cohort found no link.
Do people with bipolar disorder have lower IQs? After onset, on average, moderately lower cognitive scores, including between episodes. Before onset, small or no differences depending on how it was measured.
Is bipolar disorder linked to creativity? Registry data show people with bipolar disorder, and healthy siblings of people with it, are overrepresented in creative professions. That is about occupations, not IQ.
Can an IQ test diagnose bipolar disorder? No.
How does it differ from schizophrenia? Premorbid deficits are clear in schizophrenia (0.60 standard deviations) and lower IQ predicted it in the Swedish cohort; for bipolar disorder the premorbid evidence is equivocal.

Sources

Gale, C.R. et al. (2013): Is bipolar disorder more common in highly intelligent people? A cohort study of a million men, Molecular Psychiatry, 18(2), 190-194.
MacCabe, J.H. et al. (2010): Excellent school performance at age 16 and risk of adult bipolar disorder: national cohort study, British Journal of Psychiatry, 196(2), 109-115.
Smith, D.J. et al. (2015): Childhood IQ and risk of bipolar disorder in adulthood: prospective birth cohort study, BJPsych Open, 1(1), 74-80.
Zammit, S. et al. (2004): A longitudinal study of premorbid IQ score and risk of developing schizophrenia, bipolar disorder, severe depression, and other nonaffective psychoses, Archives of General Psychiatry, 61(4), 354-360.
Koenen, K.C. et al. (2009): Childhood IQ and adult mental disorders: a test of the cognitive reserve hypothesis, American Journal of Psychiatry, 166(1), 50-57.
Trotta, A., Murray, R.M. and MacCabe, J.H. (2015): Do premorbid and post-onset cognitive functioning differ between schizophrenia and bipolar disorder? A systematic review and meta-analysis, Psychological Medicine, 45(2), 381-394.
Swidzinski, S. et al. (2025): Domain-specific cognitive function in euthymic bipolar disorder: a systematic review and meta-analysis, Psychological Medicine, 55, e336.
Kyaga, S. et al. (2011): Creativity and mental disorder: family study of 300,000 people with severe mental disorder, British Journal of Psychiatry, 199(5), 373-379.

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 →

A range chart of how far people with depression score below healthy comparison groups on cognitive tasks, in standard deviations with an IQ-style point equivalent. Executive function, memory and attention during an episode: 0.34 to 0.65 standard deviations, roughly 5 to 10 points on a scale with a standard deviation of 15 (Rock 2014). Executive function and attention in remission: 0.52 to 0.61, roughly 8 to 9 points (Rock 2014). Executive-function tests across 113 studies: 0.32 to 0.97, roughly 5 to 15 points (Snyder 2013). These are scores on specific tasks, not full-scale IQ.

Mind & Everyday Life

Depression and IQ

Does depression lower IQ? It lowers performance on attention, memory, speed and planning tasks while an episode lasts, and some executive and attention differences remain after recovery. A permanent loss of intelligence is not shown, and two large cohorts suggest the arrow also runs the other way.

October 7, 20269 min read

Understanding IQ

High IQ and Mental Health

A widely cited 2018 study found sharply higher rates of anxiety and depression among members of a high-IQ society. Why that finding is far from the whole story, what a century of gifted research actually shows, and what seems to matter more than IQ itself.

September 18, 20266 min read

Understanding IQ

Creativity and IQ

September 28, 20268 min read

Depression and IQ

Posted on October 7, 2026 by Mahesh

Mind & Everyday Life

Does Depression Lower IQ? What the Studies Show

By Arjun Vardhan, News & Research WriterPublished October 7, 2026 · 9 min readReviewed by Aarav Vardhan, Editor-in-Chief

Does depression lower IQ? It lowers performance on tests of attention, memory, processing speed and planning while an episode lasts, and some attention and executive-function problems can remain after mood recovers. The research does not show that depression permanently lowers a person’s underlying intelligence, and the long-term cohort data point the other way as well: a lower childhood or teenage IQ predicts a higher risk of later depression. This guide separates the questions that usually get blurred together and says what each study actually measured.

During an episode: moderate deficits on tests of attention, memory and executive function, with effect sizes of about 0.3 to 0.65 standard deviations in one meta-analysis of a single test battery. These are the kinds of tasks that overlap with parts of an IQ test.
After recovery: executive-function and attention deficits (about 0.5 to 0.6 standard deviations) were still present in people whose symptoms had remitted; memory deficits were smaller and not statistically significant.
Severity matters: across studies, more severe depression went with worse episodic memory, executive function and processing speed, but not with stored knowledge.
The other direction: in a Swedish cohort of 50,087 men and a New Zealand cohort of 1,037 people, lower earlier IQ predicted more depression later.
Not shown: a permanent loss of IQ points caused by depression. The reviews measured specific tasks, not full-scale IQ.

The topic touches several pages we already publish. The reverse claim, that high intelligence brings mood problems, is covered in high IQ and mental health and IQ and happiness. How nerves and mood distort a single sitting is covered in test anxiety and IQ scores and ADHD, autism, dyslexia and IQ test scores. This page asks the direct question about depression.

What does depression do to test performance?

The most useful evidence comes from meta-analyses, which pool many small studies. Rock and colleagues (2014) combined studies that used one computerized battery, the Cambridge Neuropsychological Test Automated Battery (CANTAB), to compare people with depression with healthy controls. They found moderate deficits in executive function, memory and attention, with effect sizes (Cohen’s d) from 0.34 to 0.65. Executive function is the set of skills for planning, switching between tasks and holding rules in mind; our guide to executive function versus IQ explains how it relates to intelligence.

Another review focused on executive-function tests. Snyder (2013) pooled 113 studies and found that major depressive disorder was reliably associated with impaired performance, with effect sizes from 0.32 to 0.97 depending on the test. People with depression also had slower processing speed, but slowed movement alone could not account for the results, and there was some evidence that the deficits were larger when current symptoms were more severe.

McDermott and Ebmeier (2009) asked a different question: does severity matter? Across the studies they pooled, more severe depression went with worse episodic memory, executive function and processing speed, but not with semantic memory (stored knowledge) or visuospatial memory. That pattern echoes what a bad night of sleep does to a score, as described in our piece on sleep loss and test performance: the speeded, attention-heavy parts take the hit and stored knowledge holds up.

What three reviews found
Review	What it combined	Main finding
Rock et al. (2014)	Studies using the CANTAB battery, people with depression versus controls, during episodes and in remission	Moderate deficits in executive function, memory and attention (d from 0.34 to 0.65); executive function and attention deficits persisted in remission (d from 0.52 to 0.61)
Snyder (2013)	113 studies of executive-function tests, major depressive disorder versus healthy controls	Reliable impairment, effect sizes 0.32 to 0.97; slower processing speed too; larger deficits with more severe current symptoms
McDermott and Ebmeier (2009)	Correlations between depression severity and test performance	Worse severity went with worse episodic memory, executive function and processing speed; no link for semantic or visuospatial memory

A Cohen’s d of 0.5 means half a standard deviation, which on an IQ-style scale with a standard deviation of 15 is 7.5 points. The ranges above therefore correspond to very roughly 5 to 10 points in the CANTAB review and 5 to 15 in the executive-function review if they were put on that scale. They are not IQ points. They are scores on specific tests, the reviews did not measure full-scale IQ, and the conversion is only there to give a sense of size.

Do the problems go away when the depression lifts?

Not entirely, on average. In the same 2014 meta-analysis, people whose depressive symptoms had remitted still showed moderate deficits in executive function and attention (d from 0.52 to 0.61). Memory deficits were smaller and not statistically significant (0.22 to 0.54). The authors concluded that cognitive impairment is a core feature of depression rather than a side effect of low mood alone.

Two cautions apply. These are comparisons of groups at one point in time, so they cannot show whether the differences were caused by depression or were already there before it began. And an average gap of half a standard deviation leaves a large overlap between the groups, so plenty of people who have recovered score within the usual range.

For IQ testing the practical point is timing. The tasks in these reviews resemble the working-memory, processing-speed and reasoning parts of a full IQ battery, so a score taken in the middle of an episode may understate what the same person scores when well. That is our inference, not something the reviews tested directly. It fits what our guides say about how a single sitting can be pulled down; see processing speed and IQ and the factors that affect an IQ test result.

Does a lower IQ raise the risk of depression?

Two long-running cohorts suggest the arrow also runs this way. Zammit and colleagues (2004) followed 50,087 Swedish men who took an IQ test at military conscription in 1969 and 1970, using hospital records over 27 years. A lower IQ was associated with a higher risk of later hospital admission for severe depression, and also for schizophrenia and other non-affective psychoses. There was no association with bipolar disorder, a contrast we unpack in bipolar disorder and IQ.

Koenen and colleagues (2009) used the Dunedin birth cohort in New Zealand: 1,037 people who took the WISC-R at ages 7, 9 and 11 and were assessed for mental disorders at ages 18, 21, 26 and 32. Lower childhood IQ was associated with a higher risk of adult depression and anxiety, with more than one disorder at a time, and with depression that persisted. The authors framed this as lower “cognitive reserve”.

Neither study shows that a low IQ causes depression. Childhood IQ can also stand in for things that raise risk on their own, such as early adversity, health problems or schooling, and the Swedish outcome was depression severe enough to need hospital admission, which tells us little about milder cases. Our overview of childhood IQ and later-life health describes the same causal puzzle for physical health. The safest reading is two-way: depression can lower performance while it lasts, and lower earlier ability is a marker of higher risk.

Does a high IQ cause depression?

The popular version of this claim leans on a 2018 survey of 3,715 American Mensa members, in which about 27% reported a mood disorder, against roughly 10% in the general population. We cover that study and its limits (self-reported diagnoses, a sample drawn from the far tail of the distribution) in high IQ and mental health. In the two population cohorts above, higher IQ went with a lower, not a higher, risk of depression, and our summary of IQ and happiness finds no strong, consistent evidence at the population level that smarter people are more likely to be depressed. One finding runs the other way, and it concerns mania rather than depression: in the Dunedin cohort, higher childhood IQ predicted a higher risk of adult mania.

Your own number

Where would your own score land?

Find your IQ score now! →

Secure & encryptedInstant results10–20 minutes

What about depression in parents and the IQ of their children?

One small study looked at this directly. Miller and colleagues (2024) followed 209 mother-child pairs from a multicenter trial and compared depression screens during pregnancy and one year after birth with the child’s WPPSI-III score at age 5. Children of mothers with a positive screen during pregnancy were more likely to have an IQ below 85 (35% versus 18%), and the same held for a positive screen after birth (47% versus 21%). But once the analysis accounted for social determinants of health and clinical characteristics, the associations were no longer statistically significant (adjusted odds ratios 1.4 and 2.1, with confidence intervals that include 1).

The sample was small and came from a trial of pregnant women with subclinical hypothyroidism, so it is a caution against over-reading the raw numbers rather than a final answer. It does fit a pattern familiar from other topics on this site, where an association shrinks sharply once family circumstances are counted: the link between waiting for a treat at age 4 and teenage achievement fell by about two thirds in our grit versus IQ guide.

Should you take an IQ test while depressed?

An IQ score is a snapshot, and its margin of error is already 2 to 3 points on a well-built test (see the margin of error explained). In the middle of a low period, poor sleep and trouble concentrating can pull the speed- and attention-heavy parts of a score down further. A sensible approach is to wait, to compare scores only when they were taken under similar conditions, and to read an unusually low result from a bad stretch as a statement about that day. The same logic applies to anyone retesting after a long gap; see can your IQ change over time.

Our IQ test gives a total score for curiosity. It cannot diagnose or rule out depression, and how much any online score can be trusted depends on the conditions it was taken in (IQ test accuracy). If low mood is lasting, a doctor or mental health professional is the right place to start. In the United States, anyone in crisis can call or text 988 to reach the Suicide & Crisis Lifeline.

For a neighboring condition with a very different pattern in the IQ data, where higher ability shows up as a risk marker in some cohorts, see bipolar disorder and IQ.

Quick answers

Does depression lower IQ permanently? The studies reviewed here do not show that. They show lower performance on specific tests of attention, memory and executive function during episodes, and some executive and attention differences that persisted in remission.
Can depression lower your score on an IQ test? It can, mostly on the speeded and attention-heavy parts, and more so when symptoms are more severe.
Do depressed people have lower IQs? On average they score lower on some cognitive tests, and in two cohorts lower earlier IQ predicted later depression. The groups overlap heavily, so a person with depression can have any IQ.
Does a high IQ cause depression? Neither population cohort found that. The claim rests mainly on a 2018 survey of Mensa members that relied on self-reported diagnoses.
How big is the effect? About 0.3 to 0.65 standard deviations for attention, memory and executive tasks in one review, roughly 5 to 10 points if put on an IQ-style scale, and larger with more severe symptoms.
Should I retest after recovering? If an earlier score came during a hard period, a retest under better conditions is reasonable; expect the usual test-to-test variation of a few points as well.

Sources

Rock, P.L. et al. (2014): Cognitive impairment in depression: a systematic review and meta-analysis, Psychological Medicine, 44(10), 2029-2040.
Snyder, H.R. (2013): Major depressive disorder is associated with broad impairments on neuropsychological measures of executive function: a meta-analysis and review, Psychological Bulletin, 139(1), 81-132.
McDermott, L.M. and Ebmeier, K.P. (2009): A meta-analysis of depression severity and cognitive function, Journal of Affective Disorders, 119(1-3), 1-8.
Zammit, S. et al. (2004): A longitudinal study of premorbid IQ score and risk of developing schizophrenia, bipolar disorder, severe depression, and other nonaffective psychoses, Archives of General Psychiatry, 61(4), 354-360.
Koenen, K.C. et al. (2009): Childhood IQ and adult mental disorders: a test of the cognitive reserve hypothesis, American Journal of Psychiatry, 166(1), 50-57.
Miller, E.S. et al. (2024): The association between perinatal depressive symptoms and child neurodevelopment, American Journal of Obstetrics & Gynecology MFM, 6(11), 101488.
Karpinski, R.I. et al. (2018): High intelligence: a risk factor for psychological and physiological overexcitabilities, Intelligence, 66, 8-23.

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 comparing self-reported anxiety and mood disorder diagnosis rates between the general population, about 10 percent each, and a 2018 survey of 3,715 members of a high IQ society, about 20 percent anxiety and 27 percent mood disorder.

Understanding IQ

High IQ and Mental Health

September 18, 20266 min read

Mind & Everyday Life

Bipolar Disorder and IQ

October 7, 202610 min read

Research & Evidence

IQ and Happiness

IQ and happiness are almost unrelated at the individual level, though smarter countries do report higher average happiness. What the research actually finds, why national comparisons mislead, and what predicts happiness better than IQ.

September 22, 20266 min read

MoCA Score

Posted on October 6, 2026 by Mahesh

Scores & Scales

MoCA Score: What the Ranges Mean and How It Differs From IQ

A MoCA score is a number out of 30 from a 10-minute screen for mild cognitive impairment. See the four publisher score bands, why the 26 cut-off flags many healthy adults, what changes the score, what to do with a result and why a MoCA is not an IQ.

By Leon Falkner, Data & Research AnalystPublished October 6, 2026 · 7 min readReviewed by Aarav Vardhan, Editor-in-Chief

A score strip from 0 to 30 for the Montreal Cognitive Assessment. The publisher's ranges are severe 0 to 9, moderate 10 to 17, mild 18 to 25 and normal 26 to 30. Markers show the original cut-off of 26, the cut-off of 23 favored by a 2018 meta-analysis, and the average score of 24.7 among 2,637 cognitively unimpaired adults in the Framingham Heart Study (2026). Many healthy older adults fall below 26.

A MoCA score is a number out of 30 from the Montreal Cognitive Assessment, a 10-minute screening test for mild cognitive impairment. The publisher’s ranges are 26 to 30 normal, 18 to 25 mild, 10 to 17 moderate and 0 to 9 severe. But newer research shows that the usual cut-off of 26 flags many healthy people: in a 2026 study of 2,637 cognitively unimpaired adults, the average MoCA score was 24.7. And a MoCA score is not an IQ score. This guide explains the ranges, how the cut-off debate works, what moves the score and how it differs from an IQ test.

The ranges: 26 to 30 normal; 18 to 25 mild; 10 to 17 moderate; 0 to 9 severe, according to the publisher. Only the normal cut-off and the mild range trace to the original validation work; the publisher says the moderate and severe bands are arbitrary.
The cut-off: the 2005 validation used 26; a 2018 meta-analysis found 23 gave the best diagnostic accuracy; a 2026 study found average scores of healthy people over 60 below 23.
Education: one point is added for 12 or fewer years of formal education, up to a maximum of 30.
Not IQ: the MoCA is a screen for impairment, scored in raw points, with no validated conversion to an IQ.
What to do: a score is a prompt for a clinical evaluation, not a diagnosis.

What is the MoCA and what does it test?

Nasreddine and colleagues developed the Montreal Cognitive Assessment (MoCA) in 2005 as a 10-minute screening tool to help first-line physicians detect mild cognitive impairment (MCI), a state that often progresses to dementia. It is scored out of 30 and covers seven areas:

How the 30 MoCA points are spread across areas
Area	Points	What the points come from
Visuospatial and executive	5	a trail-making task, copying a cube and drawing a clock
Naming	3	naming three animals
Attention	6	repeating digits forward and backward, tapping to a letter, serial subtraction
Language	3	repeating two sentences and a one-minute fluency task
Abstraction	2	saying what two pairs of words have in common
Delayed recall	5	five words learned earlier, recalled without cues
Orientation	6	date, month, year, day, place and city

The fluency item is a one-minute letter F task, scored 1 point for 11 or more words; our guide to the verbal fluency test explains how those words are counted, and the digit task is related to the digit span test. We describe the sections only in general terms, not the actual items, because publishing them would weaken the test for the people who need to take it. Version 8.1 is the most recent form, and alternate versions (8.2 and 8.3) exist to reduce learning effects when the test is repeated. A MoCA-Blind version exists for people with visual impairments.

As of December 2025, the publisher says certification is optional for healthcare professionals who use the MoCA only for screening, triage or referral based on the total score, and remains required for anyone who interprets individual sections or domain-specific impairments.

What do the MoCA score ranges mean?

The publisher’s FAQ gives four bands, and it is open about where they come from: the MCI ranges were defined using Petersen criteria in the 2005 validation study, while the moderate and severe groups were arbitrarily defined by default.

The MoCA publisher’s score bands
Score	Label	Where the band comes from
26 to 30	Normal	The cut-off used in the 2005 validation study
18 to 25	Mild impairment	MCI range from the 2005 validation, using Petersen criteria
10 to 17	Moderate impairment	Arbitrary, per the publisher
0 to 9	Severe impairment	Arbitrary, per the publisher

The publisher also says a score of 18 is usually considered the line between MCI and Alzheimer’s disease, with overlap, because the diagnosis depends on the presence of cognitive and functional impairment and not on a number. That is the key limit of any band: a test score is evidence, not a verdict.

Is 26 the right cut-off?

In the 2005 validation, 94 patients with MCI, 93 with mild Alzheimer’s disease and 90 healthy older controls took both the MoCA and the Mini-Mental State Examination (MMSE). At a cut-off of 26, the MoCA detected 90% of the MCI patients and 100% of the mild Alzheimer’s patients, against 18% and 78% for the MMSE. Specificity was 87% for the MoCA and 100% for the MMSE. In other words, the MoCA caught far more impairment, at the price of flagging more healthy people. Our guide to whether an IQ test can detect dementia compares the MoCA and the MMSE in more detail.

Later work found the price was higher than first thought. Carson, Leach and Murphy (2018) pooled nine studies, chosen from 304, and found that a cut-off of 23 out of 30 gave the best diagnostic accuracy, with fewer false positives than 26. Then Muller and colleagues (2026) looked at 2,637 cognitively unimpaired adults in the Framingham Heart Study (average age 53.6, 64% college-educated). The average MoCA score was 24.7 (SD 3.0). Participants over 60 averaged below 23, and people over 70 scored below 23 regardless of education. Only college-educated people under 40 averaged near the old cut-off, at 26.3 (SD 2.4).

MoCA averages in cognitively unimpaired Framingham adults (2026)
Group	Average MoCA score	Reading
All 2,637 participants	24.7 (SD 3.0)	Close to the revised cut-off of 23
Over age 60	Below 23	Healthy group average under the revised cut-off
Over age 70	Below 23, at any education level	The cut-off flags many healthy people
Under age 40, college-educated	26.3 (SD 2.4)	Near the original cut-off of 26

The authors conclude that a cut-off of 26 produces a high rate of false positives, that 23 may also be set too high, and that clinicians and researchers need demographically appropriate, population-based norms. The practical lesson is that the same score can mean different things at 35 and at 75.

What changes a MoCA score?

Education. The standard scoring adds one point for 12 or fewer years of formal education, counted from after kindergarten, and never above 30. The Framingham norms are split by age and education, which is a reason to ask which norms a clinician is using.
Age. In the Framingham sample, averages fell below the revised cut-off after 60 even among cognitively unimpaired adults.
Repeat testing. The publisher provides alternate versions to reduce learning effects, which is why a repeat test should not reuse the same one.
Vision and physical limits. A MoCA-Blind version exists for visual impairments, and the publisher says scoring out of 25 and converting back to 30 (for example, for a person with hemiplegia) has not been validated.
Who scores it. Interpreting individual areas requires trained and certified users, per the publisher.

Your own number

Where would your own score land?

Find your IQ score now! →

Secure & encryptedInstant results10–20 minutes

How is a MoCA score different from an IQ score?

They answer different questions. An IQ score compares a person with a norm group of the same age and has a mean of 100 and a standard deviation of 15. The MoCA is a raw score out of 30 built to separate people with cognitive impairment from those without it. Healthy younger adults cluster in the upper part of the scale, so the test has little room to tell a person with a high IQ from one with an average IQ. A 30 does not mean a high IQ, and a 24 does not mean a low one. The lower averages in the Framingham study track age and education, not intelligence.

For these reasons there is, to our knowledge, no validated way to convert a MoCA score to an IQ. The IQ score converter translates between IQ scales and percentiles, and our guide to how IQ tests work explains what norming means, but neither can do anything with a MoCA total. If you want a reasoning score, take the IQ test; if you want to know whether memory or thinking has changed, the right next step is a clinician, as described next.

What should you do with a MoCA result?

Treat it as a screening result. A score below the cut-off is a reason to ask for a fuller evaluation, which usually looks at history, daily functioning and other tests, and a score above it does not rule out a problem. Because averages differ so much by age and education, ask the clinician which norms or cut-off they use. Scoring in the 18 to 25 range does not mean a person has dementia; the publisher labels it “mild impairment,” and the Framingham data show healthy adults land there often.

Quick answers

What is a normal MoCA score? The publisher says 26 to 30. In a 2026 study of 2,637 cognitively unimpaired adults the average was 24.7, and adults over 60 averaged below 23.
Is a MoCA score of 22 bad? It falls in the publisher’s mild band, but the average of healthy adults over 60 in one large study was itself below 23; it is a reason to discuss norms and follow-up, not a diagnosis.
What MoCA score indicates dementia? None by itself. The publisher says 18 is usually considered the line between MCI and Alzheimer’s disease, with overlap.
Does education change the score? Yes: one point is added for 12 or fewer years of education.
Is the MoCA an IQ test? No. It is a 30-point screen for impairment with no validated IQ conversion.
How long does it take? About 10 minutes.

Sources

Nasreddine, Z.S. et al. (2005): The Montreal Cognitive Assessment, MoCA: a brief screening tool for mild cognitive impairment, Journal of the American Geriatrics Society, 53(4), 695-699.
Carson, N., Leach, L. and Murphy, K.J. (2018): A re-examination of Montreal Cognitive Assessment (MoCA) cutoff scores, International Journal of Geriatric Psychiatry, 33(2), 379-388.
Muller, E. et al. (2026): The Montreal Cognitive Assessment at the Framingham Heart Study: a re-examination of the norms, Brain and Behavior, 16(5), e71487.
MoCA Test Inc.: frequently asked questions (score bands, education adjustment, versions and certification), accessed October 2026.

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 →

Four comparison bars for cognitive screening tools: MoCA sensitivity 90 percent and specificity 87 percent versus MMSE sensitivity 78 percent and specificity 77 percent for detecting mild cognitive impairment.

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

Verbal Fluency Test

October 6, 20269 min read

Taking a Test

IQ Testing for Children

Children are not tested on the same instrument as adults, and the test itself changes twice before age eighteen. Here is which one applies when, how reliable an early score really is, and what a parent should and should not conclude from a single number.

September 8, 20266 min read