IIF Certified Assessment
Start IQ Test
IIF Certified Assessment
Tables ranking average IQ by degree level circulate constantly and are read as though schooling simply sorts people. Pooled longitudinal evidence puts the correlation at about 0.56, and a meta-analysis of more than 600,000 people found that education also raises scores.
Search for average IQ by education level and you will find the same table reproduced everywhere: a tidy column of degree levels, a tidy column of scores rising beside it, and an unstated conclusion that schooling reveals what people already were. The correlation the table depicts is real and it is one of the strongest in the field. What the table cannot do is tell you which way the arrow points, and the published evidence says it points both ways at once.
The most-cited synthesis of this question is Tarmo Strenze's 2007 meta-analysis in Intelligence, which restricted itself to longitudinal studies — ones that measured intelligence first and socioeconomic outcomes later, so that the timing is at least unambiguous. Pooling 59 studies covering 84,828 people, it put the correlation between intelligence and eventual educational attainment at about 0.56. That was the strongest of the three outcomes examined, ahead of occupational level and well ahead of income.
A correlation of 0.56 is substantial by the standards of social science, and it deserves to be reported as such. It is also nothing like determinism. Squaring it gives about 0.31, conventionally read as the two measures sharing roughly thirty-one per cent of their variation — which leaves close to seventy per cent unaccounted for. In practical terms the score distributions at each education level overlap across most of their width. There are people with doctorates whose measured scores sit below the average for people who left school early, and the reverse, in numbers large enough that the table is useless for judging any individual.
A column of averages is a photograph of two processes running at once. Nothing in the photograph tells you which of them produced it.
The reading that treats the table as pure selection collapsed with a 2018 meta-analysis by Stuart Ritchie and Elliot Tucker-Drob, published in Psychological Science. They gathered 142 effect sizes from 42 datasets covering more than 600,000 participants, and concentrated on the study designs that can support a causal claim rather than a correlational one: policy changes that extended compulsory schooling for some birth cohorts and not others, comparisons exploiting school-entry age cut-offs, and controlled studies following the same people over time.
The pooled estimate was that an additional year of education is associated with a gain in the region of 1 to 5 IQ points. The effects appeared across the broad ability domains studied rather than being confined to obviously taught knowledge, they showed up at every stage of education examined, and — the finding that surprised most readers — they did not appear to fade with age, remaining detectable in older adults decades later. The authors described education as the most consistent and durable method yet identified for raising intelligence test scores.
Put the two findings side by side and the ranking loses its meaning as a statement about people. Higher-scoring adolescents do go on to more education. More education does raise measured scores. A cross-sectional table showing graduates averaging higher than non-graduates is fully consistent with either mechanism, with both, or with a third factor driving both — and it contains no information capable of distinguishing them. This is the same structural problem we described in what country IQ rankings cannot support: a table of group averages invites a causal reading it was never able to license.
The defensible summary is short. Intelligence measured in youth predicts how much education a person goes on to complete, at a strength of around 0.56 in pooled longitudinal data. Education in turn raises measured intelligence, by something in the region of 1 to 5 points a year. Neither figure supports reading a person's score from their qualifications, or their qualifications from their score, because the overlap between groups is far too wide for that.
If you want to know where your own reasoning sits, the qualification is the wrong proxy and the score is the right one — read as a percentile, which states the share of people who scored lower, and read with the confidence band our explainer on the margin of error describes. Our IQ percentile calculator converts a score into that position, and the useful thing about it is that it makes the overlap visible: the percentile ranges for adjacent education levels sit almost on top of each other. If you do not have a score to convert, you can take the assessment first. For the related question of what a score predicts about earnings rather than schooling, our piece on whether IQ predicts income works through a noticeably weaker relationship.
Your qualifications are a poor estimate of your reasoning and your reasoning is a poor estimate of your qualifications. Take a properly normed assessment and read the percentile instead of the proxy.
Find your IQ score now! →There is a broader reason these tables persist. A single ordered column feels like an explanation, and the alternative — two processes of comparable size running in opposite directions through a population with enormous internal variation — does not compress into a graphic. That is not a flaw in the evidence. It is the shape of the evidence, and any presentation that looks simpler than that has removed something in order to look that way.
Published tables do show average scores rising with education level, but they are far less informative than they look. The pooled longitudinal correlation between intelligence and educational attainment is about 0.56, which corresponds to roughly 31 per cent shared variation and leaves the score distributions at different education levels overlapping across most of their range. The averages are real; they cannot be used to judge an individual.
The evidence says yes, modestly and durably. A 2018 meta-analysis in Psychological Science covering 142 effect sizes from 42 datasets and more than 600,000 participants found that an additional year of education is associated with a gain of roughly 1 to 5 IQ points, appearing across broad ability domains and persisting into later life. What remains unsettled is whether that reflects a change in underlying general ability or improvement on the specific skills the tests sample.
Strenze's 2007 meta-analysis in Intelligence, pooling 59 longitudinal studies covering 84,828 people, reported a correlation of about 0.56 between intelligence and educational level. That was stronger than the same analysis found for occupational level or income, but it still leaves close to seventy per cent of the variation in educational attainment associated with other things.
No, and the correlation is exactly why not. At a relationship of about 0.56 the distributions at each education level overlap heavily, so a large share of people at any given level score far from the average for their level. A measured percentile from a normed assessment is the only defensible answer to the question, and it should be read with its confidence range attached.
Corrections: spotted an error? Email corrections@iqmetrics.org and we will update this story and note the change here.
The meta-analytic correlation between measured intelligence and earnings is around 0.23, which leaves roughly ninety-five per cent of the differences in pay to everything else. A 2026 analysis of two large US cohorts adds a second finding: your specific strengths carry a third to a half of the weight that general ability does.
The tables that circulate as national IQ figures come from one compilation built out of whatever studies existed, in whatever years, on whatever samples. Several of the numbers were estimated rather than measured.
On a mean-100, standard-deviation-15 scale, 130 sits at the 97.72nd percentile, which leaves about one person in 44 above it. Here is the rarity at every threshold — and why the figures stop describing real people long before they stop being calculable.
Our IIF-certified assessment reports your score with its scale, percentile and confidence range — and a breakdown of the cognitive domains behind it.
Start IQ Test →
