IQ Levels — The Reference Scale as a Ladder

IQ levels are not natural kinds waiting to be discovered — they are cuts through a continuous distribution, made where the math is cleanest. Because the common scale sets the mean at 100 and one standard deviation at 15, the lines at 85 and 115 each capture about 34 percent of the reference distribution on either side of the mean; adding the lines at 70 and 130 encloses about 95 percent total. The six IQ levels you can click above are simply the regions those lines create. The cuts are statistical, so the ladder behind the IQ levels is fixed.

Because the cuts are statistical, the ladder shape is fixed but the names are not. One manual may call 115–129 'high average', another 'superior', a third 'above average'. A 1950s chart and a 2020s chart can disagree about where a level starts. When you compare IQ levels across sources, compare the numbers — the boundaries and the shares — and treat the names as local dialect. The numbers behind IQ levels are stable; only the packaging varies. Compare the numbers, not the names, when judging IQ levels.

What every version agrees on is the geometry: levels are symmetric around 100, the central pair of levels holds roughly two-thirds of the distribution, and each step outward multiplies rarity sharply. The ladder's shares in the tool — computed from the normal model, not estimated by eye — make that gradient visible. The gradient is what makes IQ levels look like a ladder at all.

• 1Levels are cuts at ±1σ and ±2σ: boundaries 70 / 85 / 100 / 115 / 130
• 2Central levels: about 34% of the distribution each side of the mean
• 3Within ±1σ total: about 68%; within ±2σ: about 95%
• 4Names vary by publisher and era; the underlying math does not
• 5Every level here is described by position, not by worth

Clicking an IQ level sets a representative score and the readouts update: z-score, percentile, and the share of the reference distribution inside the level. The percentile answers the practical question — where does a score like this stand — while the z-score shows the arithmetic underneath: points above or below 100, divided by 15. That is the fastest way to make IQ levels feel concrete instead of nominal. One click turns any of the IQ levels into exact numbers.

The ladder is also a quick calibration tool for headlines. When an article claims a threshold at 120 or 132, click the nearest level and compare: on this model 120 sits near the 91st percentile and 132 just above the 98th. Small differences in cutoffs can mean large differences in rarity at the extremes — the top level holds about two percent of the distribution, so moving a cutoff a few points moves eligibility a lot. Headline cutoffs land next to these IQ levels surprisingly often.

For fine-grained work, skip the ladder and use the input field: type any score between 55 and 145 to land between levels. The boundaries page gives the table-first view of the same scale, and the chart pages plot it — but if your question is 'what are the levels and how big is each', the ladder is the direct answer. For fine work, type a score between the IQ levels.

The vocabulary around IQ levels has a fraught history: early classification systems used terms that aged terribly, and some circles still quote them. This page deliberately uses neutral, positional descriptions — 'above the reference mean', 'very high on the reference scale' — for two reasons. First, statistical honesty: a level locates a score, and location words are all the number supports. Second, respect: people are not rungs. Neutral words are the least the IQ levels deserve.

It is equally important not to swing into meaninglessness. IQ levels are not meaningless — within their scope they are reliable, useful statistics that help organize research, standardize assessments, and communicate positions economically. The scope is the point: valid as a description of test performance relative to a sample; invalid as a diagnosis, a destiny, or a ranking of human value. Within scope, IQ levels are reliable, economical statistics.

A final caution on extremes: levels near the top are sometimes treated as prizes and levels near the bottom as verdicts. Both readings misuse the statistic. High scores predict some outcomes weakly and others not at all; low scores never, by themselves, establish anything clinical — support decisions require comprehensive professional assessment. Click the ladder, read the numbers, keep the perspective. Click through the IQ levels once and the neutral language explains itself. Read the IQ levels as positions and keep the perspective.