Natalie Portman's IQ: the Strongest Record in Acting, and Still No Score

0 The Quick Answer on the Natalie Portman IQ Claim

No IQ score for Natalie Portman has ever been published by anyone who administered a test, and the figure of 140 in search results is unsupported: no version of it names an instrument, a date, an examiner or a reference sample. ACIS has not assessed her, holds no session or score belonging to her, and cannot assess anyone from a biography. Nothing on this page is an ACIS measurement of Natalie Portman, and no retrospective assessment of any person, living or dead, is possible with this or any other instrument.

The 140 appears on celebrity IQ aggregators and in entertainment roundups, always as a bare integer. It is worth noticing what number was chosen. On the standard scale with a mean of 100 and a standard deviation of 15, 140 is the folk threshold for genius, and that threshold traces to a single suggested classification table in Lewis Terman's 1916 Stanford-Binet manual, which Terman himself dropped from his 1937 revision. The page that traces the 140 line to that table covers the history. So the figure attached to her is not a measurement that came out somewhere and got misreported. It is the default number people reach for when they want to say that someone is very bright.

What sits beside that void is the strongest documented record of any working actor. Portman published a peer reviewed chemistry paper in 1998 under her birth name, Natalie Hershlag, while still in high school. She entered the Science Talent Search with that work. She was a co-author on a neuroimaging study published in NeuroImage in 2002. She earned an A.B. in psychology from Harvard in 2003. Every one of those items can be checked against a journal, a digital object identifier or Harvard's own publication. None of them is a test score, and the second half of this page explains precisely why that gap cannot be closed.

1 Where the 140 Comes From, and What It Names

Tracing the number leads nowhere, and the shape of the dead end is informative. Every page carrying the 140 cites another page carrying the 140. There is no journalist who obtained a record, no psychologist who reports having administered anything, and no interview in which Portman gives a figure. Several of the pages that publish it label it an estimate without saying whose estimate it is or what method produced it, which is a use of the word that carries no information.

A claim of this kind fails at a specific point. A real administration leaves an origin event: a person sat a named battery on a date, an examiner scored it against a reference sample, and someone reported the result with an error term. Even contested real scores leave that trail, which is why disputes about them can be settled by reading documents. There is nothing here to read.

The choice of 140 is the interesting part. It is not a random three digit number. It is the number the popular literature has used to mean genius since Terman's 1916 classification table, and it has had no manual behind it for almost ninety years. When a writer needs a figure for a famous person who is obviously accomplished, 140 is the number that arrives, which is why the same integer is attached to many unrelated people. That is a property of the writing convention, not of the person.

The audit on the page that sorts famous IQ claims into four evidence classes found that of thirty five cases examined in detail, four have any cognitive administration on record and exactly one has a named instrument, a date and a comparison group. Portman is not among the four. Neither is anyone else whose number is quoted in a listicle. If a reader wants to know how rare a genuine composite of 140 would be, the calculator that turns a score into a 1 in X rarity answers that, and the page on the gifted range explains what the bands above 130 are used for. Neither can supply a measurement that was never taken.

So the answer to the search query is a negative, and the rest of this page is the part worth reading. The documented record here is better than almost anything the celebrity IQ genre has to work with, which makes Portman the ideal test case for a question that matters more than her score: how much can a documented record actually tell you about measured ability?

2 The 1998 Journal of Chemical Education Paper

The first item in the record is a peer reviewed chemistry paper published while she was a high school student. The article is titled A Simple Method To Demonstrate the Enzymatic Production of Hydrogen from Sugar, and it appeared in the Journal of Chemical Education, an American Chemical Society journal, in October 1998, volume 75, issue 10, page 1270. The digital object identifier is 10.1021/ed075p1270. The author list carries Natalie Hershlag alongside Ian Hurley and Jonathan Woodward.

The content is a laboratory protocol rather than a theoretical claim. It demonstrates that glucose can be used to generate hydrogen through a two enzyme route, using glucose dehydrogenase and hydrogenase, in a procedure simple enough to run in a high school or undergraduate teaching laboratory. That is what the Journal of Chemical Education exists to publish: methods that instructors can actually use. Reported accounts describe the work as an independent study project begun during her sophomore year at Syosset High School.

Take the measure of what this is and is not. A teenager reaching publication in a peer reviewed journal is unusual, and the review process means at least one competent outside reader judged the protocol sound and useful. It is genuine evidence of competence in a specific technical domain: understanding an enzymatic pathway well enough to design a demonstration of it, and writing it up to a journal's standard. It is not a measurement of general ability, it does not come with a construct or a reference sample, and it says nothing about how she would perform on a timed reasoning task.

It is also, unavoidably, evidence about circumstances. Getting to a working laboratory with two credentialed co-authors at sixteen requires a school that runs independent study, adults willing to supervise, and a route into a research setting. Those are real resources and most students do not have them. Separating what the paper shows about her from what it shows about her situation is the honest analysis, and it is the subject of a later section on this page.

3 The Science Talent Search Entry, and What a Semifinalist Placement Means

The same work was entered in the Science Talent Search, the oldest pre college science competition in the United States, and this is the item on the page with the softest sourcing. Reference entries state that she was a semifinalist. The year is given as 1998 in some accounts and 1999 in others, and no listing naming Natalie Hershlag was located in the Society for Science alumni archive for this page. The correct label is reported, not documented, and the ambiguity is worth stating rather than smoothing over.

There is a structural reason to favor the later year. The competition is for high school seniors, and Portman graduated from Syosset High School in 1999, which places her eligible year in the competition judged in early 1999 rather than the one judged in early 1998. That is an inference from eligibility rules and a graduation date, not a record, so it is offered as reasoning rather than as a finding.

What a semifinalist placement means is worth being precise about, because competition tiers get inflated in retelling. The Science Talent Search receives a large annual pool of student research submissions, names several hundred semifinalists, and advances a much smaller group to the finals in Washington. A semifinalist placement is a judged assessment of a research project by a scientific review panel. It is a strong signal about the quality of a piece of work and about the student's ability to produce it. It is not a test, not a score, and not a rank on any cognitive construct.

Every row in that table is a legitimate form of evidence about something. None of them is interchangeable with any other row, and the entire celebrity IQ genre consists of pretending that they are. The distinction between a judged product and a measured construct runs through the page on what an intelligence test actually measures.

4 The 2002 NeuroImage Paper on Object Permanence

The second publication is a neuroimaging study, and it is the one most often described inaccurately. The paper is Frontal Lobe Activation during Object Permanence: Data from Near-Infrared Spectroscopy, published in NeuroImage in 2002, volume 16, issue 4, pages 1120 to 1126, with the identifier 10.1006/nimg.2002.1170 and PubMed identifier 12202098. The authors, in the order the record gives them, are Abigail A. Baird, Jerome Kagan, Thomas Gaudette, Kathryn A. Walz, Natalie Hershlag and David A. Boas.

The study used near-infrared spectroscopy, which measures changes in oxygenated, deoxygenated and total hemoglobin concentration in a defined region without radiation or restraint, to look at frontal cortex during the emergence of object permanence in infants. The abstract states the problem plainly: converging evidence points to frontal lobe maturation as central to object permanence, but methodological and ethical constraints have made neurophysiological data from awake, behaving infants hard to collect. The reported result is that the emergence of object permanence coincides with an increase in hemoglobin concentration in frontal cortex.

Three things follow for an evidence page. First, this is a real contribution to a real literature in a journal with a serious review process, and Portman is on the byline of it. Second, she is the fifth of six authors on a study run in Jerome Kagan's laboratory at Harvard, with the imaging expertise supplied by David Boas. Author position does not apportion contribution reliably in any field, and it certainly does not license a claim that an undergraduate designed or led the work. Third, and most relevant here, the paper produces no information whatever about her own cognitive test performance.

It is common to see this study described as a paper about memory that Portman wrote. The accurate description is that she was one of six authors on a developmental neuroimaging study, and that alone is a genuinely rare thing for an undergraduate to have on a curriculum vitae. The record does not need embellishment; it needs to be stated correctly, which is the same discipline applied on the page on the engineering degree behind another famous number and on the page on the screenplay behind a third.

5 An A.B. in Psychology from Harvard in 2003

The credential is documented by the institution that awarded it, which is unusual for a celebrity biographical claim. The Harvard Gazette, the university's own publication, reporting on her Class Day address on May 27, 2015, states that she attended Harvard College from 1999 to 2003 and earned an A.B. in psychology. That is the institution describing its own graduate, which removes the chain of retelling that degrades most biographical detail.

The same report is the best source for how she experienced the place. She arrived shortly after the release of the first Star Wars prequel and described feeling that her admission might be a mistake, that she was not, in her own phrase, smart enough to be in this company, and that she expected to have to prove she was not simply a famous person who had been let in. She said she spent her early years there taking only very difficult courses in order to demonstrate seriousness, and later concluded that seriousness pursued for its own sake was a pose rather than an achievement.

Reported items sit around that documented core. Accounts describe advanced Hebrew literature and neurobiology coursework, work as a research assistant for a Harvard law professor, and graduate courses taken at the Hebrew University of Jerusalem in the summer of 2004. Those come from compiled biographical references rather than registrar records and are labeled reported here.

Now the part that matters for the search query. A Harvard degree is one of the most powerful status signals in the English speaking world, and that is exactly why it is worth being careful about what it certifies. It certifies that an admissions committee selected her, that she completed a course of study, and that a faculty conferred a degree. It does not certify a score, because no score is part of the transaction that a diploma records. The general question of what university selectivity does and does not tell you about a number is worked through on the page on average IQ by university, and the measured relationship between ability and academic outcomes is on the page on IQ and academic achievement.

6 Choosing School Over Films, and One Quote Worth Tracing

The claim that she prioritized education over her career is well supported at the level of behavior and poorly supported at the level of the quotation everybody uses. Behaviorally the record is clear: she enrolled at Harvard in 1999 while the Star Wars prequels were in production, remained a full time student, and graduated in 2003 with the films still releasing. Reported accounts describe her reducing her acting workload during those years. That is a documented four year commitment made at the point in a career when an actor's market value is usually treated as perishable.

The quotation is a different matter. The line "I'd rather be smart than a movie star" is attached to her on quotation sites, in listicles and in social posts, usually paired with a longer sentence about not caring whether college ruined her career. No primary interview containing it could be located for this page: no publication, no date, no interviewer. It is exactly the kind of artifact this page exists to flag, and flagging it does not require doubting that she said something like it.

Notice that the true underlying claim and the unverifiable artifact travel together, which is the same pattern that carries the IQ number. She really did enroll and really did graduate, and that is checkable. The pithy sentence people quote to prove it is not checkable, and it spreads faster than the checkable version because it is short and quotable. When the two get bundled, the unverified part borrows credibility from the verified part.

The Class Day material is the more useful evidence anyway, because it is first person, dated, and published by the institution. What it documents is not a score but a disposition: someone who treated the intellectual environment as something to earn a place in rather than something to visit. Dispositions of that kind are not measured by intelligence tests, which is one of the honest limits of the whole enterprise and a point developed on the page on IQ and personality.

7 What Undergraduate Publication Is Actually Evidence Of

Two peer reviewed papers before the age of twenty one is strong evidence of something specific, and the specificity is the whole point. Peer review asks whether a piece of work is sound, whether the method supports the claim, and whether the write up meets a field's standards. Passing that filter twice, in two unrelated fields, is evidence of technical competence, of the ability to work inside a research team, and of sustained effort over long timescales with delayed reward.

What it is not evidence of is a rank on a general factor. An intelligence composite is a position in a distribution, obtained under standard conditions from a defined set of tasks, with a reference sample behind it and an error term attached. A publication is an accepted product. There is no equation connecting the two, and if one existed it would need inputs nobody has: how much of the work was hers, what resources were available, how many attempts preceded the accepted version, and how much guidance the co-authors supplied.

The literature does establish that measured ability and academic output are related at the group level, and the sizes are worth having in front of you. Deary, Strand, Smith and Fernandes reported in Intelligence in 2007, volume 35, pages 13 to 21, that in a cohort of more than 70,000 English children a latent general ability factor measured at age eleven correlated .81 with a latent measure of national examination achievement at sixteen. That is close to the ceiling of what this literature produces. Roth and colleagues, pooling 240 samples and 105,185 participants in Intelligence in 2015, reported a corrected correlation of .54 between intelligence and school grades. Both are group level relationships between two measured variables, which is not the operation this page would need.

The prospective evidence about very high ability and later creative output points the same way and runs in the same direction. Kell, Lubinski and Benbow followed 320 people identified before age thirteen in the top one in ten thousand on mathematical or verbal reasoning and reported in Psychological Science in 2013, volume 24, pages 648 to 659, that by age thirty eight the group had accumulated substantial creative accomplishments and prominent professional positions. Note the design. It starts from a measured score and ends at a record of achievement. Every study of this kind runs that way, because that is the direction in which the inference is defensible. This page has the record and no score, which is the study run backwards, and running a prediction backwards is a different and much weaker operation than making one.

8 Separating Ability From Access, Opportunity and Persistence

The most useful thing to say about this record is that it is jointly produced, and the honest analysis names each contributor instead of collapsing them into one trait. A sixteen year old does not publish in an American Chemical Society journal by being clever in isolation. Four things had to be present at once, and only one of them is ability.

• Access. A route into a research setting with credentialed co-authors, in a school district that supports independent study. That is a resource, distributed very unevenly, and it is a precondition rather than a consequence of talent.
• Opportunity. Time. She was already working professionally as a teenager, and choosing to spend discretionary hours on an enzymatic hydrogen protocol rather than on anything else is an allocation decision that most people in her position would not make.
• Persistence. Peer review is slow, revision is normal, and rejection is common. Carrying a project from an idea to a printed page requires tolerance for a long feedback loop, which is a trait dimension rather than an ability dimension.
• Ability. Understanding the chemistry well enough to design a demonstration and write it to a journal's standard. Real, and impossible to isolate from the other three from the outside.

This is not a way of diminishing the achievement. It is the only reading that survives contact with what is actually known. A record of accomplishment is the joint output of a person and a situation, and pulling a single number out of it requires assuming the situation away. That assumption is what every celebrity IQ estimate quietly makes, and it is why the estimates come out so high: whatever the person had going for them gets attributed entirely to the trait being estimated.

The group level literature says the same thing with numbers. Strenze, pooling the longitudinal research in Intelligence in 2007, volume 35, pages 401 to 426, reported correlations of about .56 between measured intelligence and educational attainment, .43 with occupational status and .20 with income. Those are substantial effects for social science and they leave most of the variance to everything else. Read in the direction this page would need, from an outstanding outcome back to a score, the range of scores consistent with the outcome is very wide.

The practical upshot is a rule that generalizes past this one case. When you see a documented accomplishment used as the basis for a number, ask what else had to be true for the accomplishment to happen. If the answer includes resources, timing or a decade of effort, the number is being asked to absorb credit that belongs elsewhere. That habit also disarms most of the claims catalogued on the page of persistent IQ myths.

9 A Harvard Degree Is a Selection Signal, Not a Measurement

Admission to a selective university is a decision made by people about a candidate, and a decision is not a measurement no matter how good the decision is. The distinction is not pedantic. A measurement has a defined construct, a standard procedure, a reference sample and an error term. An admissions outcome has none of those, because it aggregates test scores, transcripts, essays, recommendations, interviews, demonstrated interests, and institutional priorities that vary by year and by applicant pool.

Selection signals do correlate with tested ability, and the size of that correlation is documented rather than assumed. Frey and Detterman reported in Psychological Science in 2004, volume 15, pages 373 to 378, a correlation of .82 between SAT scores and a general factor extracted from the Armed Services Vocational Aptitude Battery in 917 participants, and a raw correlation of .48 between the SAT and Raven's Advanced Progressive Matrices in 104 undergraduates, rising to .72 after correcting for restriction of range. That is a strong group level relationship and it is the reason admissions tests are useful at all.

It is also, as Bridgeman argued in a comment in the same journal in 2005, volume 16, pages 745 to 746, far too loose to license individual conversion. A correlation of .82 still leaves a wide band of plausible ability estimates around any single test score. And that is the best case, working from an actual admissions test result. Working from the fact of admission alone, with no test score in hand, discards even that.

Every step in that chain loses information, and the chain starts with a missing link. That is the technical reason the Harvard degree cannot be turned into a number, and it holds regardless of how selective the institution is. What selectivity does tell you, at the level of averages rather than individuals, is the subject of the page on university averages, with the related question about fields of study on the page on average IQ by college major.

10 What a Proxy Can and Cannot Do

A proxy is a measurement of one thing used to say something about another, and every proxy has a stated exchange rate that has to be respected. Admissions tests, military qualification batteries and national school examinations are all usable proxies for general ability, because each is itself a standardized measurement with a reference population. That is what makes them convertible at all.

Three conditions have to hold before a proxy can be converted into an IQ estimate for an individual, and all three are routinely ignored in celebrity coverage. The first is that an actual score must exist and be authenticated, not merely the fact that the person took the test or attended the institution. The second is that a published conversion must be applied and named, so the reader can see the assumptions. The third is that the resulting figure must be reported as an interval, because the correlation between a proxy and a composite is never one and the residual uncertainty is large even when the correlation is high.

Portman fails the first condition, which ends the exercise before the others matter. No SAT score, no ACT score, and no other standardized result of hers is public. The same is true of almost every famous person: the evidence audit on the celebrity roundup found only a handful of cases with any authenticated non-IQ score behind them. Where such a score does exist, the honest treatment is set out on the page comparing the SAT with an IQ test and in the general treatment of record score claims and the evidence behind them.

There is a version of this page that would be easy to write and dishonest. It would take the two publications, the Harvard degree and the Science Talent Search entry, apply a plausible sounding adjustment, and produce a range. It would look rigorous because it cites real studies. It would still be an invention, because the inputs required for the calculation do not exist, and dressing an assumption in a confidence interval does not convert it into evidence. That is the failure mode this whole genre runs on, and it is worth naming even when the person under discussion has the best record in it.

11 Every Claim on This Page, With Its Label

A page that grades other people's evidence owes a grade on its own, so here is every substantive claim above with its source and its label. The labels are documented, meaning a primary source states it; reported, meaning a named outlet or compiled reference states it and the primary record has not been inspected; attributed, meaning someone is quoted asserting it; estimated, meaning derived by a stated method with stated assumptions; and unsupported, meaning no traceable source exists.

Two rows deserve a second look. The journal rows carry the strongest label available on a page like this, because a digital object identifier resolves to a publisher record that nobody can quietly edit. The Harvard Gazette row is unusual in the same way the college directory entry is unusual on other pages in this series: the institution is describing its own graduate, which removes the retelling chain entirely.

Notice that no row is labeled estimated. That label belongs to a figure produced by a stated method with stated assumptions, such as an authenticated proxy score converted with a published equation. Nothing here qualifies, because no proxy score exists to convert. Grading claims this way is the same discipline that separates reliability from validity, two questions people routinely merge, which is the subject of the page on reliability and validity.

12 Getting a Number That Comes With Its Own Error Bar

The question behind most celebrity IQ searches is answerable for the person asking it, and the answer arrives with the properties the celebrity number lacks. ACIS is a self-administered online assessment for adults aged 16 to 90, built from 20 subtests across the six Cattell Horn Carroll domains, with subtest scaled scores on a mean of 10 and a standard deviation of 3 and composites on a mean of 100 and a standard deviation of 15. A result comes with a stated construct, a stated reliability and a stated error term, which is precisely what a bare integer copied between pages cannot offer.

The published index figures show what that means concretely. Verbal Comprehension has an omega reliability of .9745, a standard error of measurement of 2.40 points and a g loading of .864. Fluid Reasoning is at .9727, 2.48 and .922. Visual Spatial is at .9488, 3.39 and .906. Working Memory is at .9247, 4.12 and .788. Processing Speed, the least g saturated of the six, is at .8790, 5.22 and .648. The General Ability Index drawn from the fifteen non speeded subtests reaches an omega of .9885 with a standard error of 1.61 points. The derivation is in the technical manual, and the underlying structure is described on the cognitive domains page and the g factor page.

For a reader who arrived from a question about academic ability, the relevant form is the six subtest Quick administration at 15 dollars, which runs Similarities, Vocabulary, Matrix Reasoning, Figure Weights, Digit Span and Alphanumeric Sequencing in about 45 minutes and reports Verbal Comprehension, Fluid Reasoning and a partial working memory picture. Longer forms are described on the adult testing page. When a result has to be accepted by an institution, the supervised route is the only one that works, and it is described on the page on where to take a test.

Guessing is not a substitute, and the size of the error is documented rather than asserted. Self estimates correlate roughly .30 with measured ability, loose enough that a self placed figure is routinely wrong by tens of points in either direction, which is the arithmetic set out on the page on estimating your own intelligence. The same objection that disposes of a stranger's guess about a celebrity disposes of your own guess about yourself.

13 Sources Behind This Page

Every claim and every number above traces to one of the following, and the list is short enough that a reader can check all of it. Primary records come first, then the peer reviewed literature, then the compiled references whose claims are labeled reported rather than documented.

• Hershlag, N., Hurley, I., and Woodward, J. (1998). A Simple Method To Demonstrate the Enzymatic Production of Hydrogen from Sugar. Journal of Chemical Education, 75(10), 1270. Source for the existence, journal, date and subject of the first publication. Byline order differs between Crossref and other indexes, as noted in the body.
• Baird, A. A., Kagan, J., Gaudette, T., Walz, K. A., Hershlag, N., and Boas, D. A. (2002). Frontal Lobe Activation during Object Permanence: Data from Near-Infrared Spectroscopy. NeuroImage, 16(4), 1120 to 1126, indexed at PubMed 12202098. Source for the author list, the method and the reported finding.
• Harvard Gazette, coverage of the Harvard College Class Day address, May 2015. Source for attendance from 1999 to 2003, the A.B. in psychology, and her own account of self doubt and course selection.
• Deary, I. J., Strand, S., Smith, P., and Fernandes, C. (2007). Intelligence and educational achievement. Intelligence, 35(1), 13 to 21. Source for the latent correlation of .81 between general ability at age eleven and national examination achievement at sixteen in a cohort of more than 70,000 children.
• Roth, B., Becker, N., Romeyke, S., Schaefer, S., Domnick, F., and Spinath, F. M. (2015). Intelligence and school grades: a meta-analysis. Intelligence, 53, 118 to 137. Source for the corrected correlation of .54 across 240 samples and 105,185 participants.
• Strenze, T. (2007). Intelligence and socioeconomic success: a meta-analytic review of longitudinal research. Intelligence, 35(5), 401 to 426. Source for correlations of about .56 with educational attainment, .43 with occupational status and .20 with income.
• Kell, H. J., Lubinski, D., and Benbow, C. P. (2013). Who rises to the top? Early indicators. Psychological Science, 24(5), 648 to 659. Source for the 320 person cohort identified before age thirteen in the top one in ten thousand and followed to age thirty eight.
• Frey, M. C., and Detterman, D. K. (2004). Scholastic assessment or g? Psychological Science, 15, 373 to 378, with the comment by Bridgeman, B. (2005), Psychological Science, 16, 745 to 746. Source for the SAT and general ability correlations and for the objection to individual level conversion.
• Compiled reference entries for the Science Talent Search semifinalist placement, the Syosset High School graduation year, the sophomore year independent study project, the Hebrew literature and neurobiology coursework, the research assistant work and the Hebrew University summer courses. These are labeled reported throughout because no primary record was inspected for this page.

Three things could not be resolved and are stated as such rather than smoothed over. The byline order of the 1998 chemistry paper differs between indexes and the publisher's full text is paywalled. The year of the Science Talent Search placement is given as both 1998 and 1999, and no Society for Science listing naming her was located. The frequently quoted line about preferring to be smart could not be traced to any primary interview, publication or date.

The professional framework behind all of this is written down and public. The Standards for Educational and Psychological Testing, published jointly in 2014 by the American Educational Research Association, the American Psychological Association and the National Council on Measurement in Education, require that validity evidence be presented for the specific interpretation being made, that reliability and the standard error of measurement be reported so precision is visible, that the norm sample be described so a percentile means something, and that fairness be addressed across the populations tested. A celebrity figure fails all four at once, because no construct, no instrument, no sample and no error term is ever named. APA standards on test use add that scores be interpreted by someone qualified to do so, in light of the conditions of administration, which is impossible for a number whose only history is having been copied. Those same requirements apply to this site, which is why the self-administered assessment and the non clinical status of ACIS are stated in the body of this page rather than buried in a footer. The International Test Commission guidelines on test use state the same principles in an international frame, and their practical consequences for internet delivered instruments are worked through on the page on the accuracy of online tests.

14 Frequently Asked Questions