Does IQ Above 120 Stop Mattering? What a 50-Year Study Found

The "threshold hypothesis" is a popular claim in intelligence research: that once a person's IQ clears somewhere around 120, further increases stop predicting greater achievement, creativity or success. It is one of those ideas that sounds intuitively right — surely intelligence matters up to a point and then other factors take over. A 50-year longitudinal study that has followed thousands of profoundly gifted 13-year-olds into their careers found the opposite: even within the top 1% of ability, finer differences kept predicting who went on to earn doctorates, patents and academic tenure decades later.

What is the threshold hypothesis?

The idea traces to Getzels and Jackson's 1962 book, Creativity and Intelligence: Explorations with Gifted Students, an early study finding that high IQ did not guarantee high creativity in gifted students — evidence later researchers took as support for the idea that intelligence and achievement decouple above some cutoff. The specific number most often attached to the hypothesis, around IQ 120, is a convention later researchers adopted to test the theory rather than a figure Getzels and Jackson specified themselves; treat "120" as shorthand for "somewhere in the above-average-to-gifted range," not a number from the original study.

What is SMPY and how does it test the hypothesis?

The Study of Mathematically Precocious Youth was founded by psychologist Julian Stanley in 1971 at Johns Hopkins University, with funding from the Spencer Foundation. Stanley's method, since called "talent search," was to give children well above grade level a test built for much older test-takers — the SAT, sat before age 13 — on the reasoning that a grade-level test cannot tell apart a merely bright child from a profoundly gifted one once both score at the ceiling. A twelve-year-old who scores respectably on a test built for college applicants sits at an extreme edge of ability for their age in a way a perfect score on an easier, grade-level test cannot show. Camilla Benbow and David Lubinski have led the study since Stanley, and its headquarters moved to Vanderbilt University in 1998, where both are professors. Rather than sorting people into "gifted" and "not gifted," SMPY's design captures fine-grained differences even among people who all clear a very high bar — exactly the kind of data needed to test whether ability differences above a threshold still matter.

Stanley's talent-search model did not stay inside one research lab. Johns Hopkins built its own Center for Talented Youth around the same identification approach, and similar above-level testing programs at other universities used the same logic to find children whose ability a grade-level test could not measure. Millions of children have since taken an above-level test through one of these programs, so SMPY's findings are not just a historical curiosity about one research cohort — they describe the population a great deal of current gifted-education practice is built to identify.

What did the top 1-in-10,000 group achieve?

A 2013 study in Psychological Science followed 320 people identified before age 13 as being in the top 1 in 10,000 for mathematical or verbal reasoning ability — roughly 100 times rarer than the standard "gifted" cutoff of the top 1% — for nearly three decades, to about age 38. The comparison against less extreme groups is the core of the finding.

• Doctoral degree by age 38: 44% of the top 1-in-10,000 group, versus about 23% of people in the top 1% of ability, versus about 2% of the general US population
• STEM patents: 15.3% of the group held at least one, against a roughly 1% base rate in the general population, with 133 patents total across 49 people
• Academic tenure: 11.3% had earned tenure at an accredited institution and 7.5% at a research-intensive one by age 38
• Total notable accomplishments and awards across the 320 people: 2,749, an average of 8.6 each

The doctoral-degree comparison is the clearest test of the threshold hypothesis in this data: moving from the top 1% of ability to the top 1-in-10,000 nearly doubled the doctorate rate, when a threshold model would predict no further difference once both groups are already well past any plausible cutoff. A comparison group closer to home makes the same point differently: our guide to the average IQ of MIT, Oxford and Yale students works through what elite-university admission implies about ability, a very different method reaching a similar place — selection on ability alone keeps sorting people well above the range a simple pass/fail cutoff would treat as one group.

So does IQ ever stop mattering?

SMPY's own 50-year retrospective, published in Gifted Child Quarterly in 2021 and later named the journal's paper of the year, states the finding plainly: no threshold effect for abilities in predicting future accomplishment, and even within the top 1% of general and specific abilities, over a third of the full range of individual differences shows up and is meaningful. That is a direct answer to the popular claim, from the research group best positioned to test it with real longitudinal data rather than a snapshot. It also has a practical edge for schools: a gifted program that sorts children into a single "qualifies" or "does not qualify" bucket at one cutoff is, on this evidence, throwing away information that keeps predicting outcomes on the other side of that line. Our guide to universal screening for gifted programs covers how districts are rethinking that kind of single-cutoff sorting, and our piece on what brain scans show in gifted children looks at a structural, rather than outcome-based, angle on the same top-of-the-distribution question.

Is the study still running?

Yes. SMPY's cohorts, now numbering in the thousands across five separate talent searches since the 1970s, continue to be followed into midlife; the most recent published follow-up, in 2023, examines the same kind of cohort data at age 50, decades further along than the age-38 snapshot above. A companion analysis by some of the same researchers, published the same year as the 2013 doctorate findings, looked specifically at spatial ability's role in creativity and technical innovation within the same cohorts — part of why SMPY's leaders now argue that talent searches relying only on math and verbal reasoning risk missing a real, separate dimension of ability. Our guide to whether there is a leadership sweet spot near IQ 120 covers a related but separate question — ability among people already in senior leadership roles, rather than achievement among the extremely gifted more broadly — using a different dataset entirely, so the two should not be read as the same finding restated.

Five decades of tracking some of the most intellectually able people identified as children finds ability differences still mattering well past any point a casual "threshold" claim would predict, right up to the rarest end of the distribution studied. The IQ Metrics test reports a score on the same mean-100, standard-deviation-15 scale with its percentile shown alongside it, and our percentile calculator shows exactly how rare a given score is.

Common questions