Key takeaways: IQ scores are designed to follow a normal distribution — a symmetrical bell curve centered on 100, with a standard deviation of 15. Understanding this one concept explains almost everything about how IQ scores are reported, compared, and sometimes misunderstood.
Why IQ Scores Are Built Around a Bell Curve in the First Place
When psychologists first started developing standardized cognitive assessments in the early 20th century, they needed a way to express results that would remain meaningful no matter how the test itself changed over time — new questions, revised norms, different age groups. The solution was to stop reporting raw scores directly and instead convert every result onto a standardized scale, calibrated so the results naturally form a bell-shaped (normal) distribution.
This wasn’t an arbitrary choice. Many human traits that result from the combined influence of many small, independent factors — height, reaction time, and yes, performance on reasoning tasks — tend to naturally distribute in this bell-shaped pattern when measured across a large population. Test-makers lean into this tendency deliberately: they calibrate scoring so that the average performance lands at a fixed center point, and results spread outward from there in a predictable, symmetrical way.
The Two Numbers That Define the Whole Scale
Every normal distribution is fully described by just two numbers: its mean (the center point) and its standard deviation (how spread out the results are around that center). For IQ scores, the convention almost universally used today is:
- Mean = 100 — the exact center of the distribution, representing average performance
- Standard deviation = 15 — the “unit” of spread used to describe how far any individual score sits from average
Once you know these two numbers, you can describe the entire distribution. About 68% of all scores fall within one standard deviation of the mean (between 85 and 115). About 95% fall within two standard deviations (70 to 130). And about 99.7% fall within three standard deviations (55 to 145). This is sometimes called the “68-95-99.7 rule,” and it applies to any normally distributed dataset, not just IQ.
What This Actually Looks Like
Picture the classic bell shape: a tall peak directly above 100, tapering symmetrically down on both sides. The peak being so tall directly above the center is the visual representation of the fact that “average” is, by a wide margin, the single most common outcome — roughly half of the entire area under the curve sits between 90 and 110 alone.
As you move outward from the peak in either direction, the curve drops off quickly. This is the part that trips people up most often: because the tails are so thin, moving from “high average” to “gifted” territory requires covering a lot of statistical ground even though it looks like a small jump on paper. Going from 110 to 130 sounds like “20 points,” but it’s actually the difference between roughly the 75th percentile and the 98th percentile — you’re crossing from “somewhat above typical” all the way to “rarer than 1 in 50.”
Why Standard Deviation Matters More Than People Realize
Standard deviation is the reason a “10-point difference” doesn’t mean the same thing depending on where it happens on the scale. A 10-point gap between 95 and 105 reflects a fairly ordinary difference well within the thick, common part of the curve. A 10-point gap between 125 and 135 reflects a much rarer, more separated pair of results, because that stretch of the curve is thin — far fewer people occupy that territory to begin with.
This is also why clinicians and researchers are cautious about treating small score differences as meaningful, especially once you’re outside the 90–110 range. A 3- or 4-point gap almost anywhere on the scale is well within the range you’d expect from normal test-retest variation — it’s simply not a large enough movement to represent a real, reliable difference in underlying ability.
Not Every Test Uses the Exact Same Scale
While mean 100 / SD 15 is by far the most common convention (used by the Wechsler scales, most modern online tests, and most contemporary research), it wasn’t always universal. Some older tests, and a handful of specific instruments, have used a standard deviation of 16 or even 24 historically. This is worth knowing because a raw score from an older test or a different scale can’t be compared directly to a modern 100/15-scaled result without converting it first — the same underlying raw performance could produce a noticeably different-looking number depending on which standard deviation convention was applied.
This is one of the more common sources of confusion when people compare scores from different tests taken years apart, or compare a modern result to an older family record — the numbers aren’t necessarily on the same scale, even if they’re both labeled “IQ.”
How This Connects to Percentile
If you’ve read how IQ percentile is calculated, this is the mechanism underneath it. Percentile is derived directly from where a score falls on this normal distribution — it’s a way of translating “distance from the mean, measured in standard deviations” into a more intuitive “percentage of people you outperformed.” The bell curve and the percentile table are two views of the exact same underlying model.
Real-World Implication: Reading Your Own Score Correctly
Understanding the bell curve changes how you should read your own result. If your score comes back at 106, you’re solidly inside the thick, common part of the curve — a genuinely typical result, and the 6-point gap from 100 isn’t meaningfully different from average in any practical sense. If your score comes back at 128, you’re deep into the thin part of the right tail, and that gap from 100 represents something statistically much rarer, even though “28 points” and “6 points” are just numbers that look linearly comparable at first glance.
This is exactly why a good results report — including the one you get from the free IQ test — shows your percentile alongside your raw score rather than the score alone. The percentile does the work of translating your position on the curve into something you can actually interpret without needing to think in standard deviations yourself.
Frequently Asked Questions
Why is the IQ mean set at exactly 100? It’s a deliberate calibration choice, not a natural law — test-makers scale scoring so that the average performance in the reference population lands on 100, making the number easy to interpret (above 100 = above average, below = below average).
What does a standard deviation of 15 actually mean in practice? It means roughly 68% of people score between 85 and 115, and moving further from 100 in either direction represents an increasingly rare result, following the standard bell-curve pattern.
Can a 10-point IQ difference be meaningless? Depending on where it falls, yes. Near the center of the distribution (90–110), a 10-point gap is well within normal test-retest variation. Further out on the tails, the same numeric gap represents a rarer, more separated result.
Do all IQ tests use the same standard deviation? Most modern tests use 15, but not all historical or specialized instruments do. Scores from tests using a different standard deviation convention aren’t directly comparable without adjustment.
Is it possible to score outside the 55–145 range? Statistically yes, but extremely rarely — that range covers roughly 99.7% of the population under a standard normal distribution. Scores outside it are exceptionally rare and, at the very low or very high ends, often require more specialized testing instruments to measure accurately at all.