IIF Certified Assessment
Start IQ Test
IIF Certified Assessment
Average IQ is 100 at twenty, at fifty and at eighty — not because nothing changes, but because the test re-scores you against your own age group every time. Drag through a lifetime and watch the two numbers separate.
One lifetime, three trajectories
The flat line is your reported IQ. It is flat by construction, not by luck. Drag the scrubber, or click anywhere on the chart.
| Age | Reported IQ | Fluid reasoning | Crystallised knowledge |
|---|---|---|---|
| 20 | 100 | 100 (±0) | 100 (±0) |
| 30 | 100 | 100 (±0) | 102.1 (+2.1) |
| 40 | 100 | 97 (−3) | 104.1 (+4.1) |
| 50 | 100 | 94 (−6) | 106.2 (+6.2) |
| 60 | 100 | 91 (−9) | 108.3 (+8.3) |
| 70 | 100 | 88 (−12) | 106.8 (+6.8) |
| 80 | 100 | 85 (−15) | 105.3 (+5.3) |
| 90 | 100 | 82 (−18) | 103.8 (+3.8) |
At 40, a test would still report your IQ as 100 — because it is scored against other 40-year-olds. Measured against a 25-year-old instead, the reasoning half of that score sits nearer 97, while the knowledge half sits nearer 104.1.
These are population-average trajectories, not a prediction about you. Individual variation is far larger than the lines above, and health, education and practice all move them. See the confidence range on a single score.
If you have seen a table claiming “average IQ at 45 is 106”, it was measuring something else and calling it IQ. Here is what a test actually does with your age.
You get a raw score — how many items you solved, and how quickly. This number really does change across a lifetime, and on its own it means nothing.
The raw score is looked up in a table built from people of roughly your age. A 70-year-old is never scored against 25-year-olds. This is the step almost every “IQ by age” chart ignores.
Each age band is re-centred on 100 with a standard deviation of 15. So the average is 100 at 20 and at 80 by definition. A flat line is not a discovery; it is the arithmetic working correctly.
If the averages differ between age rows, one of three things is true: they are raw scores mislabelled as IQ, they came from an instrument that was never age-normed, or they were invented. There is no fourth option. A properly normed test cannot produce a non-100 average for its own norming group — that is what norming means.
“Intelligence” is not one quantity that rises or falls together. The part that solves novel problems and the part that stores what you already know follow almost opposite paths — which is why a single number can stay flat while both halves move.
Solving problems you have never seen: matrix puzzles, number series, mental rotation, processing speed. It peaks in the mid-twenties and falls from roughly age 30 with no plateau — the decline does not pause and then resume, it is steady.
What you have accumulated: vocabulary, general knowledge, professional judgement, verbal comprehension. It keeps improving for decades and typically peaks somewhere around age 60 before drifting down slowly.
| Age band | Average reported IQ | Fluid reasoning vs a 25-year-old | Crystallised knowledge vs a 25-year-old |
|---|---|---|---|
| 20–29 | 100 | 100 | 100–102 |
| 30–39 | 100 | 100 to 97 | 102–104 |
| 40–49 | 100 | 97 to 94 | 104–106 |
| 50–59 | 100 | 94 to 91 | 106–108 |
| 60–69 | 100 | 91 to 88 | 108 peak |
| 70–79 | 100 | 88 to 85 | 107–105 |
| 80+ | 100 | 85 and below | 105–104 |
Where these numbers come from, and how far to trust them. The trajectories follow the pattern reported by Salthouse and by the WAIS-IV normative sample (N = 1,577), where the correlation between age and reasoning performance runs at roughly r = −0.40. The honest caveat is that the two ways of measuring this disagree in opposite directions: cross-sectional studies compare different people born decades apart and so overstate individual decline (a cohort confound), while longitudinal studies retest the same people and understate it, because practice effects inflate later scores and the least healthy participants drop out. The shaded band on the chart above is that disagreement drawn honestly — roughly −0.015 to −0.025 SD per year — rather than a single confident line.
Age-norming matters most for children, because a year of development at six is worth far more than a year at forty-six. It is also why a single early score should never be treated as a verdict.
Children are normed in narrow bands — often three or four months wide — because development is so fast at that age. The same raw performance from a child who is six years two months and one who is six years eleven months produces meaningfully different IQ scores.
The younger the child, the weaker the link between today’s score and their adult score. Scores measured in the early school years move substantially by adolescence; by the teenage years they settle considerably. Treat an early result as a snapshot of current performance, not a ceiling.
The original ratio IQ was mental age ÷ chronological age × 100. It worked passably for children and broke completely for adults — it implies ability keeps rising with age forever, which is why every modern test abandoned it for the age-normed deviation score used today.
Your reported IQ score does not systematically change with age, because IQ tests are age-normed — each age band is re-centred so that its average is 100. What does change is raw performance. Fluid reasoning (novel problem solving) declines from about age 30 at roughly 3 IQ points per decade measured against a young-adult reference, while crystallised knowledge (vocabulary, facts, judgement) rises into your sixties. A single IQ number can stay flat while both of those move.
100 at every age. This is true by construction rather than by coincidence: test publishers build separate norm tables for each age band and set the mean of each band to 100 with a standard deviation of 15. Any table showing different averages for different ages is reporting raw scores, using a test that was never age-normed, or is fabricated.
There is no single peak, because the underlying abilities peak at different times. Fluid reasoning peaks in the mid-twenties. Processing speed peaks earlier still. Crystallised knowledge peaks around age 60. Vocabulary can keep improving even later. The idea of one peak age comes from treating intelligence as a single quantity, which the evidence does not support.
Not the number a test reports, no. If you take a properly normed test at 65, you are compared with other 65-year-olds, so a stable relative standing produces a stable score. What declines is your performance relative to a younger reference group. Both statements are true at once, and confusing them is the single most common error in this topic.
Most free online tests are not age-normed at all — they compare everyone against a single undifferentiated table, often weighted toward younger test-takers, and many are timed in ways that penalise slower processing speed rather than weaker reasoning. That produces a lower number without measuring lower ability. It is worth checking whether a test states its norming approach before reading anything into the result.
Less accurate than the single number suggests. Every score carries a margin of error, and a typical online test produces a 95% confidence range of roughly ±9 points. That range is usually wider than the differences people argue about. You can see this directly on the percentile calculator, or check whether a gap between two scores is real on the comparison tool.
An age-normed score, its percentile and its confidence range — stated up front, not buried. That is the only version of this number worth having.
