IQ and entrepreneurship: what studies of founders, the self-employed and CEOs measured

0 The short answer

Business owners who incorporate, and the chief executives of large companies, tend to score above average on cognitive tests taken earlier in life, but no study reviewed here shows that a high IQ makes a person an entrepreneur or makes a founder succeed. In a US youth survey, the incorporated self-employed had scored higher on a learning aptitude test as teenagers, while the unincorporated did not stand out in the same way. In Swedish conscription records of 1.3 million men, the median large-company CEO ranked in the top 17 percent for cognitive ability, which is high but not extreme, and founder CEOs scored lower than hired CEOs of non-family firms. A 2021 meta-analysis found cognitive ability related to entrepreneurial success but less strongly than emotional intelligence, and the lists of billionaire IQs that circulate online are not measurements.

1 What Counts as an Entrepreneur, and How Was Ability Measured?

The word entrepreneur covers at least four different groups in the research, and each study measured ability with a different yardstick, so no single number answers the question. The first group is the self-employed, anyone who reports working for themselves, from a freelance driver to a consultant. The second is the incorporated self-employed, owners whose businesses are legal corporations, which Levine and Rubinstein use as their stand-in for entrepreneurs. The third is founders, people who started a company, as in the Swedish records of founders and venture sales. The fourth is chief executives, who may be founders, heirs or hired professionals. A fifth, looser group appears in popular lists: billionaires and celebrity founders, who are not a research sample at all.

The studies also ask two separate questions. Entry asks who ends up owning a business or running a company. Performance asks how much the owner earns, how large the venture's sales become, or how big a firm the executive runs. A trait can predict entry without predicting performance, and the reverse. Keeping those apart is the main discipline of this page, and it is why the same finding can look supportive in one study and flat in another.

The cognitive yardsticks differ just as much. Levine and Rubinstein used the Armed Forces Qualification Test (AFQT) taken by participants in a national youth survey in 1980, reported in their 2013 discussion paper as a percentile with a median of 50. The discussion paper describes the score as built from tasks involving arithmetic reasoning, paragraph comprehension and numerical operations, plus a knowledge measure, and the page on IQ versus the ASVAB and the AFQT explains how that score relates to IQ. It is a military aptitude composite, not an IQ test.

The Swedish CEO study used the military enlistment test taken at about age 18. According to Adams, Keloharju and Knüpfer, the cognitive test consists of four subtests designed to measure inductive reasoning, verbal comprehension (a synonym test), spatial ability (a metal folding test) and technical comprehension. The subscores and a composite are reported on a stanine scale, which divides a normal distribution into nine intervals, each 0.5 standard deviations wide except the first and last. In the vocabulary of the CHC model, inductive reasoning sits closest to fluid reasoning (Gf), the synonym test to crystallized knowledge (Gc) and the folding test to visual processing (Gv). That mapping is our reading, not the authors' classification, and the technical comprehension subtest has no clean CHC label.

The third yardstick is a proxy. Wai and Rindermann did not test any chief executive. They classified Fortune 500 CEOs by whether they attended a school whose admission tests would likely place them in the top 1 percent of ability, then inferred ability from attendance. A study of 1.3 million conscripts and a study of school attendance can both be summarized as "CEOs are smart", but they are different kinds of evidence, and the table in the next section keeps them in separate rows.

2 The Studies at a Glance

Read as a set, the studies agree on a modest, positive direction for average ability and disagree on almost everything about size, group and outcome. The table lists each study with its design, its cognitive measure, its definition of entrepreneur and what it printed. Figures appear only where the source printed them, and the 2013 discussion paper is labeled as such where its figures are used.

Three cautions apply to every row. First, the executive studies start from people who reached a post, so none of them observes the people who tried and did not get there. Second, the cognitive measures are different instruments, so the percentiles do not stack. Third, correlations in a table are population summaries, and the page on reliability and validity explains why a population coefficient is not a prediction for one person.

3 Who Becomes an Entrepreneur? Levine and Rubinstein on Smart and Illicit

Levine and Rubinstein found that people who later incorporated a business had scored higher on a learning aptitude test as teenagers, but the sorting depended on a combination of traits rather than on aptitude alone. Their 2017 paper in the Quarterly Journal of Economics disaggregates the self-employed into incorporated and unincorporated owners. The abstract states that people who become incorporated business owners tend to be more educated and, as teenagers, score higher on learning aptitude tests, exhibit greater self-esteem and engage in more illicit activities than others. It adds that the combination of smart and illicit tendencies as youths accounts for both entry into entrepreneurship and the comparative earnings of entrepreneurs.

The data are the National Longitudinal Survey of Youth 1979. In the 2013 discussion paper, the authors describe it as a representative survey of 12,686 individuals who were 15 to 22 years old when first surveyed in 1979, followed annually through 1994 and every other year after, with the paper using survey years 1979 through 2010. Cognitive ability is the AFQT percentile from the 1980 wave. The illicit activity index was measured when respondents were between 15 and 22 and rests on youthful self-reports. The paper reports that the incorporated self-employed were twice as likely as salaried workers to report having taken something by force as youths, were 44 percent more likely to have been stopped by the police, and had an overall illicit activity index almost three times that of salaried workers. These are group averages of survey answers about adolescence. They say nothing about the character of any individual business owner, and the authors treat them as a marker of a tendency to bend the rules and take risks.

The regression result is the most instructive part, and it is easy to miss. In the discussion paper's multinomial logit comparing incorporated self-employment with salaried work, a dummy for above-average AFQT, entered alone, had a negative coefficient that was at most marginally significant. The interaction of above-average AFQT with an above-average illicit index had a positive, statistically significant coefficient of 0.58 (standard error 0.23). The authors summarize this as an almost 60 percent greater probability of becoming incorporated for people with both traits, after controlling for many characteristics, and they add that the combination of the two traits only boosts the probability of becoming incorporated, not unincorporated. In the authors' words, the mixture of high learning aptitude and break-the-rules behavior is tightly linked with entrepreneurship. Our reading is that aptitude alone did not carry the entry result in that specification. That is a finding about one survey, one aptitude composite and one operationalization of illicit behavior, and it is not a rule about intelligence.

Other factors in the same regressions are large and have nothing to do with ability. The discussion paper reports that women were 76 percent less likely than comparable men to incorporate, and that a 100,000 dollar increase in family income, enough to move a family from the 10th to the 90th percentile, was associated with an almost 60 percent higher probability of incorporation after controlling for the person's cognitive and noncognitive traits, educational attainment and family environment. The authors note that, to the extent family income proxies for credit constraints, the result suggests that difficulties in obtaining finance matter for who becomes an entrepreneur. The page on poverty and IQ covers a different question, whether low income changes measured cognition, and the two findings should not be conflated. Here the question is who gets to start and fund a company.

Two further descriptive facts from the discussion paper help locate the group. When the survey asked whether respondents had launched a business with considerable initiative and risk, 66 percent of the incorporated said yes, against 45 percent of the unincorporated. And 46 percent of the incorporated had a college degree or above between 1994 and 2010, against 32 percent of salaried workers. Education is part of the profile, and the paper's regressions include it, which is why "smart" in this literature cannot be separated cleanly from schooling.

4 Do Incorporated Entrepreneurs Earn More, and Does Ability Explain It?

Incorporated owners in this survey earned more per hour than salaried workers and unincorporated owners earned less, so the average over all self-employment hides two opposite results. This is the second half of the Levine and Rubinstein contribution, and it addresses a long-standing puzzle in the economics literature, the finding that the self-employed on average earn less than comparable employees. The 2017 abstract states that the incorporated and their businesses engage in activities that demand comparatively strong nonroutine cognitive abilities, while the unincorporated and their firms perform tasks demanding relatively strong manual skills, and that individuals tend to experience a material increase in earnings when becoming entrepreneurs, an increase that occurs at each decile of the distribution.

The 2013 discussion paper gives the magnitudes, which should be read as that version's numbers. The median incorporated self-employed person earned 41 percent more per hour and worked 27 percent more hours than the median salaried worker, while the median unincorporated person earned about 21 percent less per hour. After conditioning on standard characteristics, the incorporated had average residual hourly earnings 48 percent greater and median residual earnings 28 percent greater than salaried counterparts. At the top of the distribution, the 90th percentile of residual hourly earnings of the incorporated was 70 percent above the 90th percentile of the salaried. The paper warns that it does not account for the full array of costs and risks of starting a business, so these are not risk-adjusted investment returns.

The design matters for the causal verbs. The paper separates three things. First, sorting: those who ever incorporate earned about 30 percent more as salaried workers than comparable workers who never incorporate, so successful employees move into incorporated business ownership, while it was comparatively unsuccessful salaried workers who sorted into unincorporated self-employment. Second, a within-person boost: with controls for individual effects and individual trends, the paper reports an 18 percent boost in average residual hourly earnings when workers switch from salaried to incorporated self-employment. Third, a structural point: only 0.1 percent of the unincorporated self-employed incorporate in a given year, which the authors read as evidence that incorporation reflects planned business activity rather than an after-the-fact label for successful small businesses.

What this does and does not say about ability is narrower than headlines suggest. The earnings analysis shows that a combination of traits predicts both who incorporates and what incorporated owners earn. It does not show that raising someone's IQ would raise their earnings as an owner, because the traits are measured in youth and entangled with schooling, family income, gender and race. The within-person boost estimates what happens to the same worker's earnings after a switch, which is a stronger design than a cross-sectional comparison but still an observational one. The page on IQ and income covers the general relationship between measured ability and pay, including why income is not the same as wealth, and this page does not repeat it.

5 What Do Swedish Conscription Records Say About CEOs?

In Swedish military records, the typical large-company CEO scored well above average at age 18 but was not an outlier, and the authors conclude that the traits are necessary but not sufficient for reaching the top. The study by Adams, Keloharju and Knüpfer in the Journal of Financial Economics combines a near-exhaustive sample of male CEOs of Swedish companies with their cognitive and noncognitive ability and height at age 18. Military service was mandatory during the sample period, so the test pool includes virtually all Swedish men. The sample is 1.3 million men born between 1951 and 1978, of whom 26,000 served as CEOs of companies of varying sizes at some point in 2004 to 2010. The paper compares CEOs with the population and with more than 6,000 lawyers, 9,000 physicians, 40,000 engineers and 9,000 college-educated finance professionals.

The headline numbers are specific. The median large-company CEO, defined as one running a company with at least 10 billion Swedish kronor in total assets (about 1.3 billion US dollars), was in the top 17 percent of the population in cognitive ability. Less than one fifth of Swedish large-firm CEOs belong to the cognitive elite, the top 5 percent as defined by Herrnstein and Murray. The authors write that CEOs are smarter than average but not as smart as one might infer from prior literature. Moving from small to large companies, CEOs' average cognitive and noncognitive ability rise about two thirds of a standard deviation. Large-company CEOs were about at par or higher than physicians, lawyers, engineers and finance professionals on all three traits, though when compared with equally well-paid professionals they had a slightly lower cognitive ability than all but the finance professionals.

For readers used to IQ numbers, a rough translation is possible, with a warning. A person in the top 17 percent sits at about the 83rd percentile, which corresponds to a score near 114 on a scale with a mean of 100 and a standard deviation of 15, and the top 5 percent corresponds to about 125. That is our arithmetic, not the authors' conversion, and it assumes the conscript composite behaves like a normally distributed IQ composite. The page on the standard deviation of 15 and the percentile calculator show how these conversions work, and the rarity calculator turns a score into a one in X figure. The point of the translation is scale, not precision: the median large-company CEO in this data looks like a high-average to superior scorer, not like the 150 to 170 figures that circulate for famous founders.

Two other findings sharpen the picture. First, noncognitive ability, rated by a psychologist in an interview that evaluates social maturity, intensity, psychological energy and emotional stability, was the best predictor of appointment to a CEO position, followed by cognitive ability and height, and CEOs differed from other high-skill professions most in noncognitive ability. Second, cognitive ability was more important for larger companies, which are more likely to hire externally. The median large-firm CEO ranked in the top 5 percent in a combination of the three traits, but the authors count more than 100 times as many men in managerial roles with better trait combinations who did not become a large-company CEO in the seven-year window. Traits predict who gets into the pool, not who is chosen from it. The traits also explained little of CEO pay: only about one tenth of the pay premium of large-firm CEOs could be attributed to the labor market returns to the three traits.

6 Are Founders Smarter Than the Managers They Hire?

In the Swedish CEO data, founders scored lower than hired chief executives of non-family firms, while a separate study of founders found that higher founder ability goes with higher venture sales, and the two results answer different questions. The Adams study reports that CEOs managing smaller firms and family firms have lower traits, particularly if they come from the founding family and have not founded the company themselves. Founder CEOs, many of whom have an impressive track record in building up and growing the business, exhibit on average 0.1 to 0.2 standard deviations lower traits than non-family company CEOs. Non-family company CEOs had about one half of a standard deviation higher cognitive ability than founder-managed or heir-managed family companies, and founders had higher trait values than heirs by at most 0.1 standard deviations.

The authors draw a measured conclusion. As their findings on founders suggest, business acumen does not necessarily coincide with high trait values, which raises the possibility that the job primarily requires qualifications the traits do not capture, such as relevant work experience or knowing the right people through early life connections. They do not claim that founders are less able than managers in any broader sense. The comparison is of CEOs of existing companies by ownership type, drawn from firms that varied from small to very large, and it says little about ambitious startups that never reached a CEO register.

A different design gives a different comparison. Mohammadi and Shafi, in a peer-reviewed conference paper presented at the DRUID22 conference, study a near-exhaustive sample of Swedish male founders. They report that higher founder cognitive ability is associated with higher performance measured by venture sales, that the effect is larger for founders in high-technology industries, and that the relationship is mediated by founding-team size and team collective cognitive abilities. The study compares founders with other founders, so it asks whether, among people who started ventures, ability goes with larger sales. The Adams study compares founders who became CEOs with other CEOs. A founder group can be lower on average than hired executives and still show a positive relationship between ability and sales within the founder group, because the two comparisons draw different lines around different populations.

The mediation finding deserves attention. If the ability association with sales runs partly through assembling a team with high collective cognitive ability, then a lone-founder reading of the result, that a smarter founder sells more, misdescribes the mechanism the authors propose. It would also be wrong to treat an association with sales as an association with profit, survival or personal earnings. The paper's abstract reports the association as sales. We left out its numerical effect size because the research portal prints the figure without its unit.

For readers interested in individual founders, ACIS has separate pages on the circulating IQ figures for Elon Musk, Bill Gates, Jeff Bezos, Sam Altman and Mark Zuckerberg. Those pages trace where each number came from. This page does not rely on them, and none of the studies above identifies any named founder.

7 Does Ability Pay More for Entrepreneurs Than for Employees?

Two economics studies suggest that general ability is rewarded at least as strongly in entrepreneurial income as in wages, while Swedish enlistment data find that psychologist-rated noncognitive ability matters more than cognitive ability for avoiding poor labor market outcomes. Hartog, Van Praag and Van Der Sluis ask how valuable cognitive and social abilities are for entrepreneurs relative to employees, comparing returns to ability during spells in entrepreneurship and in wage employment. Their abstract states that this difference-of-difference design accounts for selectivity into entrepreneurial positions insofar as it is determined by fixed individual characteristics. They report that general ability has a stronger impact on entrepreneurial incomes than on wages.

The same abstract adds that entrepreneurs and employees benefit from different sets of specific abilities. Verbal and clerical abilities have a stronger impact on wages, whereas mathematical, social and technical ability are more valuable for entrepreneurs. The balance across the various abilities also generates a higher income, but only for entrepreneurs, a finding the authors read as supporting a theory in which entrepreneurs need to be jacks of all trades. That reading is theirs. It implies that a single global score hides something an entrepreneur-specific analysis can see, namely that the shape of a profile may matter as well as its height, and the page on the CHC model explains the broad abilities that sit behind a single overall score.

The result needs careful wording. A stronger relationship between ability and income inside entrepreneurship does not mean that entrepreneurs have higher ability, and it does not mean ability is the main driver of their incomes. It describes the slope of income on ability, conditional on being in each state. The abstract does not report a percentile difference in ability between entrepreneurs and employees, and this page does not infer one.

Lindqvist and Vestman use Swedish military enlistment data to assess cognitive and noncognitive ability for labor market outcomes, with noncognitive ability based on a personal interview conducted by a psychologist. Their abstract reports strong evidence that men who fare poorly in the labor market, in the sense of unemployment or low annual earnings, lack noncognitive rather than cognitive ability. However, cognitive ability is a stronger predictor of wages for skilled workers and of earnings above the median. The paper is not about entrepreneurs, but it draws on Swedish enlistment records like the CEO study and it frames the pattern: cognitive ability matters more higher up the earnings distribution, and temperament and drive matter more in avoiding poor outcomes. The page on IQ and personality summarizes what is known about how the two domains relate.

8 What Do the Meta-Analyses and National Comparisons Add?

The two meta-analyses that bear on the question are less direct than their titles suggest, and neither supplies a clean correlation between IQ test scores and starting or growing a company. Unger, Rauch, Frese and Rosenbusch reviewed human capital and entrepreneurial success in 70 independent samples with a total of 24,733 people. They found a significant but small relationship between human capital and success, with a correlation of .098. Human capital there means education and experience, and the outcomes of investments in human capital such as knowledge and skills. It is not an IQ score. The abstract reports that the relationship was higher for knowledge and skills than for education and experience, higher for human capital with high task-relatedness than for low, higher for young businesses than for old, and higher for business size than for growth or profitability.

The reading for this page is limited. If the relationship between a broad bundle of human capital and success is as small as .098, then the relationship between any one ingredient and success cannot be assumed to be larger. The review also shows how much the answer depends on the outcome: the same bundle relates differently to the size of a business than to its growth or its profitability. That variation is one reason an article can truthfully report "ability predicts firm size" and "ability does not predict profitability" about different studies.

Allen, Stevenson, O'Boyle and Seibert compared general mental ability with emotional intelligence in a meta-analysis of 65,826 observations. The abstract states that both matter for entrepreneurial success, but the size of the relationship is more than twice as large for emotional intelligence. The authors argue this contradicts and adds contextual nuance to meta-analyses of performance in traditional workplace settings, where general mental ability is considered more critical. The abstract available to us does not print the underlying coefficients, so this page does not report them. It also does not turn the comparison into a statement about any individual, because a meta-analytic comparison of two predictors says which predictor carries more weight on average, not what share of any one person's success each explains. The page on emotional intelligence versus IQ covers the measurement problems of emotional intelligence, which are different from those of cognitive tests, and the page on IQ and job performance covers the traditional workplace meta-analyses the authors contrast with.

A third kind of evidence compares countries. Hafer and Jones estimate cross-country regressions using a Global Entrepreneurship Development Index and a measure of national cognitive skills. For a sample of 60 countries they report that national cognitive skills robustly predict the index, with an unconditional correlation of 0.65 and a standardized beta of 0.42, after including conventional controls. The index gives weight to both entrepreneurial attitudes within a nation and the institutional and economic prerequisites for creating high-value, high-growth firms. This is a country-level finding. Countries with higher average cognitive skills scored higher on an index that mixes attitudes and institutions, and nothing in it identifies whether the entrepreneurs in those countries scored higher than their neighbors. Drawing that inference from group data is the ecological fallacy, and the page on average IQ by country treats national averages as a separate topic.

9 Why Billionaire IQ Lists Are Not Data

A figure for a billionaire or celebrity founder becomes data only if it names a test, a date and an examiner, and the studies that exist on elite business groups use proxies, not scores. The question of whether the very top of business is staffed by the very smartest people has been studied with a proxy. Wai and Rindermann retrospectively assessed the education and cognitive selection of Fortune 500 CEOs at four time points spanning 1996 to 2014, with a total of 1,991 CEOs. They report that between 37.5 and 41.0 percent of these CEOs attended an elite school that likely placed them in the top 1 percent of cognitive ability, which would make the top 1 percent about 37 to 41 times overrepresented. Within each of the four samples, higher CEO education and ability, by this proxy, was associated with higher gross revenue of the CEO's company; the paper's 2014 table reports r = .13 for 493 companies.

The same paper then puts those numbers in context. Using the same school classification, 20.6 percent of House members, 41.0 percent of Fortune 500 CEOs, 59.9 percent of CEOs who attended the World Economic Forum in Davos, 85.2 percent of the Forbes most powerful men and 90.1 percent of academics who attended Davos had attended an elite school. The authors conclude that Fortune 500 CEOs were highly selected but, among the broader group of elite occupations, were not at the top, and that this shows the wide cognitive range across occupations that compose the US elite. They also state the weakness of the proxy: some students attend an elite school with lower than typical test scores for reasons such as athletics, legacy status or affirmative action, others with higher than typical scores do not attend for reasons such as financial limitations, and this lowers the reliability of the measure as an ability indicator, especially at the individual level.

Set that beside the Swedish result. Wai and Rindermann infer that the top 1 percent are heavily overrepresented among Fortune 500 CEOs. The Swedish records measure ability directly, and a footnote in Adams and colleagues compares an earlier Wai estimate, that 38.6 percent of Fortune 500 CEOs in 2012 attended a school likely to place them in the top 1 percent, with their finding that 17 percent of large-firm CEOs belong to the top 4 percent in cognitive ability, which is a less demanding cutoff. The two studies differ in country, company size, time, proxy and method, and the authors of the Swedish paper say the discrepancy between prior estimates of CEO IQ and their evidence suggests leadership ability cannot be boiled down to a single trait. Both agree that CEOs are selected on ability, and neither places the typical CEO at a genius level.

None of this applies to a famous founder's circulating number, because those numbers are not measurements of this kind. A usable figure would come with the name of an instrument, the date it was given, who administered it and a reported score, and the ACIS pages on Elon Musk, Bill Gates, Jeff Bezos and Sam Altman examine what is actually on record for each name. A figure circulating without those elements is a claim about a number, not a test result. The page on accurate IQ tests explains how to tell whether a score is real, and the page on JD Vance takes up perceived leadership, which is a separate topic from measured ability.

10 Age, Risk and Experience: What Else Moves Entrepreneurial Outcomes?

The evidence on founder age and risk suggests that experience and decision tendencies shape who succeeds at least as visibly as ability does, which is a reason to keep the ability question narrow. Azoulay, Jones, Kim and Miranda integrated administrative data on firms, workers and owners to study start-ups in the United States and found that successful entrepreneurs are middle-aged, not young. The mean age at founding for the 1-in-1,000 fastest growing new ventures is 45.0. The findings are similar for high-technology sectors, entrepreneurial hubs and successful firm exits, and prior experience in the specific industry predicts much greater rates of entrepreneurial success. The authors state that the findings strongly reject common hypotheses that emphasize youth as a key trait of successful entrepreneurs.

The result bears on ability in a specific, limited way. A popular picture of the young genius founder suggests that raw cognitive speed is the key input, and the data do not show a pattern consistent with that picture for successful ventures. It does not follow that cognitive ability declines or does not matter, and the study contains no cognitive test. The page on how IQ changes with age explains which abilities tend to change across adulthood and which are stable, and it helps to read the Azoulay result alongside it, as a statement about the age at which successful founding happens, not about ability.

Risk and motivation are the other recurring theme. Åstebro, Herz, Nanda and Weber review evidence that many entrepreneurs seem to enter and persist in entrepreneurship despite earning low risk-adjusted returns. They review three sets of interpretations from behavioral economics: differences in risk aversion, overconfidence, and nonpecuniary taste-based factors such as enjoying the work. This review contains no cognitive test result, and its abstract does not say who has higher ability. Its relevance is that business entry is not a pure ability story, and that the average returns reported in studies such as Levine and Rubinstein's apply to a selected, surviving group. People who try and leave are largely missing from records of incorporated owners and CEOs.

Together with the conscription study's finding that noncognitive traits predicted appointment better than cognitive ability, the picture is of a field in which ability may open doors but experience, temperament, finance and timing carry much of the weight. That is consistent with the Adams authors' own conclusion that the skills that make a CEO are not easily measurable. It is also a reason why a score from any cognitive test, including the ones ACIS reports through its six indices, cannot say whether a given person should start a business. The page on IQ and success covers the broader relationship between measured ability and career outcomes.

11 What the Evidence Supports, Stated Narrowly

The reviewed studies support five narrow statements about groups, and each one is limited by its design. The table pairs each statement with the design that backs it and the question it leaves open.

Two further statements are about what the literature does not contain. No reviewed study measured the IQ of a named founder, billionaire or startup unicorn leader with an identified test, and the figures that circulate for them are claims about numbers. And no reviewed study tested whether raising a person's cognitive ability changes whether they start, grow or keep a business. Every causal verb on this page is therefore used with the design that earns it: ability is associated with entry, goes with earnings, predicts appointment, and accompanies sales, and none of those studies manipulated ability.

The most defensible one-sentence summary is that measured ability is one of several things that help people enter business ownership and reach the top of large companies, that it is neither a threshold nor a guarantee, and that temperament, experience, family resources and chance are named by the studies themselves as large parts of the picture.

12 What This Does Not Say About You

A group result in a study of founders, the self-employed or CEOs cannot tell you whether you should start a business, how well it would do, or what your own IQ is. Every statistic above describes a sample: a survey cohort from the 1980s, Swedish men tested decades ago, Fortune 500 executives, a meta-analytic average. The Swedish authors count more than 100 times as many men in managerial roles with better trait combinations than the median large-company CEO who did not become one in the seven-year window. If traits that favorable do not guarantee the post, a single test score cannot predict one person's outcome.

The Standards for Educational and Psychological Testing (AERA, APA and NCME, 2014), published at testingstandards.net, is the framework to apply here. It treats evidence for a proposed use of scores as the core of validity, and reading a founder study as grounds for a business decision or a hiring decision is a different use from the one any cognitive test was validated for. The APA's Ethical Principles of Psychologists and Code of Conduct states in Standard 9.06 that, when interpreting assessment results, psychologists take into account the purpose of the assessment as well as test factors, test-taking abilities and other characteristics of the person assessed that might reduce the accuracy of their interpretations. Both documents point the same way: a score is evidence about performance on a set of tasks under certain conditions, and its meaning depends on what it is used for.

The practical limits of an online test follow from that. As read on October 6, 2026, ACIS offers a free trial with no card, and a Full Scale form at 50 dollars one time that covers all 20 subtests in six domains and takes about 175 minutes. Its report shows a Full Scale IQ, six index scores, percentiles and a 95 percent confidence interval, with adult norms for ages 16 to 90. It is online and unsupervised, in English only, and it is not a clinical or diagnostic instrument, not for hiring, school accommodations or admission to high IQ societies, and not a tool for choosing business partners, founders or executives. The page on the technical manual documents the instrument, and the page on how accurate online tests are explains what to expect from any unproctored score.

If you want a reference point for your own profile, a report across verbal, fluid reasoning, quantitative, visual spatial, working memory and processing speed domains gives more information than a single number, and it can be set beside the studies on this page as context, not as a forecast. The page on what IQ scores mean explains how to read each band, and the full scale IQ test page explains what the longest form covers.

13 Sources Behind This Page

These references separate peer-reviewed economics and psychology studies from a discussion paper and a conference paper, and each figure on the page comes from the document named next to it. The 2013 discussion paper supplies the Levine and Rubinstein percentages, and the published abstract supplies their qualitative findings. Meta-analytic figures come from abstracts opened on October 6, 2026. Percentile and IQ translations on the page are our arithmetic and are labeled as such.

• Levine R and Rubinstein Y. Smart and Illicit: Who Becomes an Entrepreneur and Do They Earn More? The Quarterly Journal of Economics, 2017, volume 132, issue 2, pages 963 to 1018.
• Levine R and Rubinstein Y. Smart and Illicit: Who Becomes an Entrepreneur and Does It Pay? CEP Discussion Paper No 1237, Centre for Economic Performance, London School of Economics, August 2013.
• Adams R, Keloharju M and Knüpfer S. Are CEOs born leaders? Lessons from traits of a million individuals. Journal of Financial Economics, 2018, volume 130, issue 2, pages 392 to 408.
• Wai J and Rindermann H. The path and performance of a company leader: A historical examination of the education and cognitive ability of Fortune 500 CEOs. Intelligence, 2015, volume 53, pages 102 to 107.
• Hartog J, Van Praag M and Van Der Sluis J. If You Are So Smart, Why Aren't You an Entrepreneur? Returns to Cognitive and Social Ability: Entrepreneurs Versus Employees. Journal of Economics and Management Strategy, 2010, volume 19, issue 4, pages 947 to 989.
• Lindqvist E and Vestman R. The Labor Market Returns to Cognitive and Noncognitive Ability: Evidence from the Swedish Enlistment. American Economic Journal: Applied Economics, 2011, volume 3, issue 1, pages 101 to 128.
• Mohammadi A and Shafi K. Entrepreneurs' Cognitive Ability and New Ventures' Performance. Conference paper, DRUID22, Copenhagen Business School research portal, 2022, opened October 6, 2026.
• Unger J, Rauch A, Frese M and Rosenbusch N. Human capital and entrepreneurial success: A meta-analytical review. Journal of Business Venturing, 2011, volume 26, issue 3, pages 341 to 358.
• Allen J, Stevenson R, O'Boyle E and Seibert S. What matters more for entrepreneurship success? A meta-analysis comparing general mental ability and emotional intelligence in entrepreneurial settings. Strategic Entrepreneurship Journal, 2021, volume 15, issue 3, pages 352 to 376.
• Hafer R and Jones G. Are entrepreneurship and cognitive skills related? Some international evidence. Small Business Economics, 2015, volume 44, issue 2, pages 283 to 298.
• Azoulay P, Jones B, Kim J and Miranda J. Age and High-Growth Entrepreneurship. American Economic Review: Insights, 2020, volume 2, issue 1, pages 65 to 82.
• Åstebro T, Herz H, Nanda R and Weber R. Seeking the Roots of Entrepreneurship: Insights from Behavioral Economics. Journal of Economic Perspectives, 2014, volume 28, issue 3, pages 49 to 70.
• American Educational Research Association, American Psychological Association and National Council on Measurement in Education. Standards for Educational and Psychological Testing, 2014, testingstandards.net, opened October 6, 2026.
• American Psychological Association. Ethical Principles of Psychologists and Code of Conduct, Standard 9.06, Interpreting Assessment Results, opened October 6, 2026.

14 Frequently Asked Questions