“Average” is not a kind of person. It is a position within a reference distribution created by comparing performance with an appropriate norm group. If an adult receives an IQ score of 100, it is easy to hear the number as though a machine has measured 100 units of intelligence. It has not. An IQ of 100 is conventionally the mean of a standardized score distribution. It tells us where a person’s test performance falls relative to the normative group used by that testing system. “Average intelligence” therefore means performance in the central region of an appropriate comparison distribution—not possession of an average-sized substance called intelligence.
That sounds technical until we ask the question that makes the whole system necessary: average compared with whom?
A ten-year-old should not have to outperform an adult
A ten-year-old and a 25-year-old can take age-appropriate cognitive assessments and both receive a standardized score of 100. The child is not being credited with adult-level knowledge or reasoning. Each person’s performance is interpreted relative to people in the relevant normative group. This is what norm-referenced scoring does. Raw performance is converted into a score whose meaning depends on a reference distribution. For many modern IQ scales, the conventional mean is 100 and the standard deviation is 15. Roughly speaking, if scores followed a normal distribution exactly, about two-thirds would fall between 85 and 115. But those numbers are properties of the scoring scale and normative distribution; they are not natural joints at which human beings divide into cognitive species.
What does the average person actually look like?
There is no single “average-intelligence profile.” Two people can obtain the same overall composite through different patterns of performance across tasks. One may perform relatively better on verbal tasks and less strongly on some spatial tasks; another may show the reverse. Even those profiles must be interpreted cautiously because subtest differences contain measurement error and may not be stable enough to support strong individual conclusions. The word average can also hide how broad the central part of a distribution is. Someone at the 40th percentile and someone at the 60th are in different positions, but both are close to the middle of the population. Treating every small score difference as psychologically meaningful would give the scale more precision than it possesses.
Average is not a verdict about competence
A standardized cognitive score is not a direct rating of whether someone can run a business, raise a child, repair an engine, write a novel or make a good decision. Real performance depends on knowledge, personality, motivation, practice, health, opportunity and the specific demands of the task. Nor does “average” mean mediocre in an everyday evaluative sense. In statistics, the mean is simply a location in a distribution. Most human activities do not require exceptional standing on a general cognitive test, and people near the center of the distribution differ enormously in expertise, interests and accomplishments.
Why norms have to be maintained
A standardized score only has meaning if the comparison group is appropriate and the test was administered and scored as intended. Norms can become outdated as populations and educational conditions change. Test publishers therefore periodically renorm instruments. The Flynn effect—the historical rise observed in many cognitive-test scores across parts of the twentieth century—made this especially visible. If an old norm table is applied to a later population, the same raw performance can receive a different standardized interpretation than it would under contemporary norms. This is another reminder that 100 is not a physical quantity. It is an assigned center point in a measurement system.
The useful meaning of “average”
Used carefully, the term is informative. It says that on the performances sampled by a particular assessment, interpreted through its norms, the estimate lies around the center of the relevant population distribution. Used carelessly, it becomes a label for a whole person. Psychometrics gives us a way to locate performance relative to others. It does not turn that location into an identity.
References
- AERA, APA, & NCME. (2014). Standards for Educational and Psychological Testing. AERA.
- Wechsler, D. (2008). WAIS–IV Technical and Interpretive Manual. Pearson.
- Flynn, J. R. (1987). Massive IQ gains in 14 nations. Psychological Bulletin, 101, 171–191. DOI
- Trahan, L. H., Stuebing, K. K., Fletcher, J. M., & Hiscock, M. (2014). The Flynn effect: A meta-analysis. Psychological Bulletin, 140, 1332–1360.
