GUIDES FOR UNDERSTANDING
IQ percentiles: What they mean and what they do not
A percentile places a score within a reference distribution. It does not tell you the percentage of questions you answered correctly or the percentage of intelligence you possess. Without an appropriate comparison group and a documented method, adding a percentile to a total creates an appearance of precision.
Relative position, not percentage correct
A report placing a score around the 75th percentile locates it approximately at the position corresponding to three quarters of its reference distribution. The exact definition depends on the norming convention and its treatment of tied scores. NIST explains that different methods exist for estimating sample percentiles.
By contrast, 15 correct answers out of 20 tasks is 75% correct. That percentage comes from the tasks and their solutions; it needs no information about other people. Locating that total relative to a group requires additional evidence. The two numbers can coincide by chance while representing different things.
The central question: Compared with whom?
Think of a race. A position among runners training at a club and a position among entrants to a local event describe different comparisons. Similarly, a position among a website’s volunteers does not establish a position among all Spanish-speaking or English-speaking adults. This illustrates the reference-group question; it is not an exact analogy for assessment procedures.
Before using a percentile, find the population, sample selection, ages, date, version and administration conditions. Thousands of participants do not by themselves resolve biased selection. Previous attempts also matter; do not ignore the difference between someone repeating a test and someone seeing it for the first time.
Percentile steps are not equal score steps
Percentiles order positions; they do not create equal distances in the ability being measured. A difference between the 50th and 60th percentiles need not correspond to the same score change as a difference between the 85th and 95th. Extremes require particular care and depend on what an instrument can estimate.
An ideal normal model allows a mathematical relationship between distance from its centre and cumulative proportion. That is a property of the model. Applying it to a person requires justification of both their score and the reference distribution. Drawing a bell curve over this pilot’s totals does not provide that justification.
What to check when two reports disagree
Check whether the reports use the same instrument and version, reference group and comparable conditions. Look for repeated attempts or instruction changes. Then find how uncertainty is communicated. Do not declare one value true and the other false merely because one is higher.
For a formal decision, ask the organisation requesting an assessment which procedure it accepts and how results should be interpreted. A percentile alone does not explain your needs, everyday skills or the cause of a difficulty. Nor does it independently establish a diagnosis or certification.
Why your percentile is not shown here
This website’s pilot form counts complete solutions across 22 OMIB tasks. We have not established our own Spanish or English norms or a representative global population comparison. The on-screen result therefore reports correct answers, answered tasks and omissions, without a population percentile.
Use this explanation to read reports that do have appropriate norms. It is not a conversion table, an IQ calculator or a ranking of people taking this activity. The distinction between describing your session and comparing you with a population remains even when your total is very high.
Frequently asked questions
Does the 90th percentile mean an IQ of 90?
No. A percentile indicates relative position, while IQ uses a score scale. Relating them requires the instrument’s norms and interpretation.
Can this website’s percentage correct be converted to a percentile?
This pilot has no justified population conversion. Percentage correct describes the tasks solved; it does not show how many people would obtain a lower total.
Sources you can check
Educational content by Sator editorial; it is not a professional assessment or clinical review. Commercial examples are not hands-on reviews. How we prepare these guides.