What each score means: 85, 100, 115, 130
Every IQ score marks a spot on a curve with an average of 100 and a standard deviation of 15. Knowing how to read that spot helps you avoid the most common traps in online results.
In this article
The bell shape
When many people take the same reasoning test, the results pile up in the middle and thin out at the ends. The shape that emerges is the normal curve, also called the bell curve: tall in the center, symmetrical, with thin tails on both sides.
IQ tests are built so that scores follow this shape. The average of the comparison group is set at 100, and the width of the curve is measured in standard deviations of 15 points. Nobody scores 100 because they got 100 of something right; 100 is simply the midpoint.
Fifteen points at a time
The standard deviation tells you how far results typically stray from the average. On the normal curve, about 68% of people fall within one standard deviation of the center, between 85 and 115. Between 70 and 130, two deviations out, you find about 95%. Above 130, roughly 2 in 100 remain.
The theoretical percentile tells you what share of the curve falls below a given score. The table shows the most commonly cited points.
| Score | Distance from the average | Theoretical percentile (approximate) |
|---|---|---|
| 70 | 2 deviations below | 2 |
| 85 | 1 deviation below | 16 |
| 100 | at the average | 50 |
| 115 | 1 deviation above | 84 |
| 130 | 2 deviations above | 98 |
| 145 | 3 deviations above | 99.9 |
Wechsler's classic manuals give names to the ranges: 90 to 109 is average; 110 to 119, high average; 120 to 129, superior; 80 to 89, low average. The names vary between tests and editions, and the latest versions of some instruments have switched to more neutral terms, such as very high and very low. The label is a convention; what it sums up is the distance from the average.
A theoretical percentile, not a ranking of users
A percentile can be calculated in two ways. The theoretical one uses the normal curve: 115 sits at the 84th percentile because, in a population with an average of 100 and a deviation of 15, 84% fall below that point. A user ranking counts how many people who took that site's test scored lower than you.
The IQNorth report uses the theoretical percentile. People who look for an IQ test online are not a sample of the population: research with ICAR showed that participants in online studies score higher than the general population (Condon and Revelle, 2014). A ranking among users would shift depending on who shows up on the site each week, and would stop saying anything about the population.
Why so many results say above 99%
Scoring above 99% of the population takes something around 135 on the IQ scale. By definition, only 1 person in 100 gets there. When a site hands that result to lots of people, the math doesn't add up, and the reason usually lies in the method.
The most common culprits are questions that are too easy, norms drawn from samples that were never corrected, and a percentage of correct answers stretched until it looks like an IQ. There is also a commercial motive: a high number feels good and gets shared. A serious result shows where the norm came from, states the margin of error, and says how far the test is able to tell scores apart.
If almost everyone lands at the top of a scale, the problem is the scale.
At the ends there are fewer people and more error
The farther from the average, the fewer people there are to serve as a comparison, and the fewer test questions have the right difficulty to separate one score from another. That is why the margin of error grows at the ends.
Every test has a ceiling and a floor. On IQNorth, for people aged 20 to 29, the test can't tell scores apart above about 138 or below about 57; at other ages these limits change. If you get every question right, you'll read that your result is at least the ceiling value, not a made-up number above it. The margin around each score is explained in Why an honest result comes with a margin of error.
See also: The report.
Read next: What IQ do you need to join Mensa?
Sources
- Wechsler, D. (1939). The Measurement of Adult Intelligence. Baltimore: Williams & Wilkins.
- Neisser, U., et al. (1996). Intelligence: Knowns and unknowns. American Psychologist, 51(2), 77–101.
- Condon, D. M., & Revelle, W. (2014). The International Cognitive Ability Resource: Development and initial validation of a public-domain measure. Intelligence, 43, 52–64.
- AERA, APA & NCME (2014). Standards for Educational and Psychological Testing. Washington, DC: American Educational Research Association.
