IQ 140: gifted, Mensa, and what the label actually means

130 is a convention, not a scientific boundary

The gifted label starts at 130 because 130 is exactly two standard deviations above the mean on a 15-point scale, and because organisations need a number they can write into a rule. It is the figure Mensa's 98th-percentile admission criterion corresponds to, and the figure many school gifted programmes cite. What it is not is a point where anything changes in the distribution of ability — the curve is completely smooth through 130, and 129 and 131 differ by less than the measurement error of the test that produced them.

The clearest evidence that the line is administrative is that it moves with the scale. Tests normed with a 16-point standard deviation put the same 98th percentile at 132, which is why both figures circulate as "the gifted threshold" and why the same underlying ability can land above the line on one test and below it on another. A score of 140 clears both.

Two consequences follow. First, treat "gifted" as a label a selection process applied, not a property a person has. Second, be sceptical of any claim that hinges on the exact number: with a standard error of about ±4 points, 140 means roughly 136 to 144, and that window straddles thresholds.

How rare an IQ of 140 really is

A score of 140 or above is about 1 in 260 — about 4 people in every 1,000, or roughly 0.4% of the population. The Gifted band as a whole covers a little over 2% of people, but it is spread very unevenly: scores just above 130 are many times more common than scores in the 140s, because the curve falls away steeply here.

That steepness is why percentiles compress at this end. One IQ point is worth only about less than 0.1 percentile places at 140, against about 2.7 near the mean — so a ten-point gap up here separates far rarer groups than a ten-point gap in the middle, even though the arithmetic looks the same.

The 140 'genius' line is a book title, not a finding

No classification system in current use contains a genius category, yet 140 is quoted as the genius threshold more often than any other number. The trail leads to Lewis Terman, who from the 1920s followed a large cohort of Californian children selected for very high scores on his revision of the Binet scale, and published the results as Genetic Studies of Genius. The selection level was around this figure, and the word in the title travelled with it.

What the study actually found is the part that did not travel. The cohort did well on average and their outcomes varied enormously, patterned by family circumstances, health and persistence rather than by score; several went on to ordinary careers, and the eminence Terman expected to follow from the number largely did not. So the strongest available evidence attached to 140 is evidence against reading it as destiny — which is close to the opposite of how the figure is used.

What "gifted" does and does not predict

On average, and across large groups, higher measured reasoning is associated with better academic and occupational outcomes. That association is real, moderate, and much weaker at the individual level than the label suggests — whether IQ predicts success sets out how much of the variance it actually accounts for, and how much is left to everything else.

The historical evidence is worth knowing here. Lewis Terman's long-running study followed a large group of children selected for very high scores through their whole lives, expecting eminence; what it found was a group who did well on average, with outcomes spread widely and strongly patterned by circumstance, health and persistence. High measured reasoning turned out to be one input among several rather than a determinant.

The popular 140 "genius" line

No classification system in current use has a "genius" band. The figure of 140 circulates because Terman's 1920s studies selected children scoring around that level and were published under the title Genetic Studies of Genius — the word travelled with the number and outlasted the methodology. A score of 140 is above that informal line, which is why you will see it described as genius level; the description is popular usage, not a classification.

Treat the word as journalism rather than measurement. What counts as a genius IQ traces where the various figures came from.

Where test ceilings start to matter

Precision is not constant across the scale. A test's items are concentrated where its test-takers are, so the further out a score sits, the fewer items and the fewer norming subjects support it. At 140 that is beginning to bite: the confidence interval around the score is wider than it would be at 110, and short unsupervised tests are the worst affected because they have fewest items to spare at the top.

Above 145 the problem changes in kind rather than degree, because standardisation samples contain almost nobody there. At 140 the norms are still resting on real observations, if fewer of them.

How many 140s in a country of sixty million

About 0.4% of people reach 140, which in a country of sixty million implies roughly two hundred and thirty thousand. That is a large number attached to a very small proportion, and the tension between the two is exactly why the ‘genius’ framing misleads in both directions: the score is genuinely uncommon, and the group it describes is far too large for the word to carry the singularity it implies. Any claim that a score at this level puts someone among a rarefied few is arithmetically wrong.

Which reasoning strengths produce a composite this high

A composite of 140 generally requires several abilities well above average at once, but the profile behind it is rarely flat. Strong fluid reasoning and pattern recognition are the usual load-bearing components; working memory and processing speed vary much more widely and can sit near average in someone scoring 140 overall.

Whether a gifted-range score is broad or sharply peaked has practical consequences that the composite cannot express — it is the difference between someone who is quick at everything and someone who is exceptional at one thing and ordinary at others.

At 140 the composite sits far enough out that each strand carries a visibly wider confidence interval than it would nearer the mean; the shape of the profile remains readable, but the precision of any single strand figure does not survive being quoted to the point.