What an IQ of 70 means — and what it doesn't

Where an IQ of 70 sits on the scale

The IQ scale is built so that 100 is the average and 15 points is one standard deviation. That makes 70 exactly 2 standard deviations below the mean — around the 2nd percentile, or above about 2% of people. The band from 70 to 79 is conventionally labelled Borderline, a term inherited from mid-century clinical practice rather than a description of anything that happens in the distribution at those two numbers.

In tail terms, a score at least this far below the average is about 1 in 45 — uncommon, but far from unheard of. In a group of 1,000 people you would expect about 23 to score 70 or lower.

Your score is a range: about 66 to 74

This is the single most important thing to understand about a score of 70. No test measures reasoning exactly. Every result carries a standard error of measurement — on a well-constructed test, usually somewhere between 3 and 5 points — which means the honest way to report 70 is "about 66 to 74", not "70".

The practical consequence: a difference of a few points between two results, or between two people, is not a real difference. Sit the same test again next week and the number will move around inside that window. Anyone treating 70 as a fixed property of a person is reading more into it than the measurement can support.

Why 70 is the one IQ number that ended up written into rules

Of all the scores on the scale, 70 is the one that appears in formal documents — diagnostic manuals, education law, disability criteria. The reason is arithmetic rather than evidence: 70 is exactly two standard deviations below 100, and a round two-SD figure is an easy number to write into a rule. Nothing about human ability changes between 71 and 70.

What is worth knowing is that the criteria themselves are more careful than their reputation. They require an assessment of how a person actually manages everyday life alongside any score, and they explicitly build in an allowance for measurement error rather than treating the number as a boundary. So even in the documents where 70 appears, it was never intended to function as a line a person falls on one side of — which is the opposite of how it is usually cited.

Why online tests are least accurate at this end of the scale

Tests are calibrated where the test-takers are, and that is near the middle. Below about 80 there are fewer items that meaningfully separate one score from the next, so a single skipped section or an expired timer shifts the result more than it would at 100. Unsupervised conditions add their own noise: nobody checks whether you understood the instructions, whether the connection dropped, or whether you were interrupted.

None of that makes an online result worthless — it makes it a wide estimate. How accurate online IQ tests are explains what they can and cannot resolve.

What an IQ score leaves out

An IQ test is a sample of timed abstract reasoning. It does not sample vocabulary you use fluently, skills you have built through practice, how you handle people, how well you manage a household or a job, creativity, or persistence. Those are not footnotes to intelligence — for most of what people actually do, they matter more than the score does, and none of them appear in it.

A composite this low is exactly where a single figure is least informative and a strand-by-strand breakdown matters most: results at this end are frequently pulled down by one specific area — processing speed under time pressure, or unfamiliarity with a question format — while other abilities sit considerably higher.

Why an IQ of 70 is not ‘half’ of 140

IQ is an interval scale, not a ratio scale: it has no true zero, so ratios between scores carry no meaning. An IQ of 140 is not twice the reasoning ability of 70, any more than 20°C is twice as warm as 10°C. What the numbers do support is ordering, and distance measured in standard deviations — 70 is two below the mean, 140 about two and two-thirds above. Nothing multiplicative follows. Treating IQ as a quantity you can have twice as much of is the most common error made with the scale, and it is a category mistake rather than an exaggeration.

When a professional assessment is the right next step

If a score in this range matters for something concrete — a school placement, a workplace adjustment, a benefits or diagnostic question — an online estimate is the wrong instrument. A qualified psychologist administers a full battery in person, verifies the conditions, and assesses everyday functioning alongside any score. Formal criteria for intellectual disability reference roughly two standard deviations below the mean, but always together with that functioning assessment and an explicit allowance for measurement error. A number on its own is never sufficient.