Ask the Cognitive Scientist: What Do IQ Scores Mean?

How does the mind work—and especially how does it learn? Teachers’ instructional decisions are based on a mix of theories learned in teacher education, trial and error, craft knowledge, and gut instinct. Such knowledge often serves us well, but is there anything sturdier to rely on?

Cognitive science is an interdisciplinary field of researchers from psychology, neuroscience, linguistics, philosophy, computer science, and anthropology who seek to understand the mind. In this regular American Educator column, we consider findings from this field that are strong and clear enough to merit classroom application.

QUESTION: What does an IQ score actually mean? On the one hand, I’ve heard people say that these scores are meaningless—a high IQ simply shows that the child is good at taking intelligence tests. On the other hand, I know that some districts use IQ test scores as gatekeepers for their gifted and talented programs. What is the science behind IQ scores?

ANSWER: An IQ score is a summary of how successfully a child can answer the types of questions that are frequently posed in school, especially questions that require thinking abstractly with words, numbers, or space. This summary number is not meaningless because it predicts students’ success in school and in many jobs.

But IQ is frequently misunderstood by educators, families, and the general public. Instead of being recognized as a summary of correctly answered questions, it’s believed to be a measure of an internal essence within the child that many call their learning potential. But there is little evidence that such an essence exists at all, let alone that IQ tests measure it.

A better way to think about an IQ score is as a snapshot of achievement now, rather than future potential. This conceptualization offers a better lens for how we might enable all children to answer more questions correctly—that is, to raise their IQs.

You have no doubt noticed that some children learn faster than others, can comprehend more difficult concepts, and are therefore easier to teach. Everyone knows this; we perceive it in daily life outside of schools too. We expect that children and adults are valued and respected regardless of how quickly or easily they learn, but the differences are there for anyone to observe. Understanding these differences may help us teach more effectively.

But how should we go about describing or measuring these differences? In the early 20th century, the French government faced many of the same problems we face today, including overcrowded and underfunded schools filled with struggling students. The French psychologist Alfred Binet was a leader of a group of teachers and administrators (La Société Libre pour l’Étude Psychologique de l’Enfant) who were interested in taking practical steps to improve education, including by developing a standardized measure that would predict how well students were likely to do in school.¹ Along with the psychiatrist Théodore Simon, in 1905 Binet developed what is now recognized as the first intelligence test; through multiple measures, it assessed reasoning, attention, and problem-solving abilities.² Binet intended his test as a diagnostic tool for children experiencing difficulties in school and did not use the term intelligence quotient (IQ). He was suspicious of attempts to boil the results of his measures down to a single number.

Unfortunately, when the idea of testing cognitive ability caught on around the world, especially in the United States, it was adopted by people who were not motivated by a desire to help children and not concerned with the consequences of describing ability with a single, reductive number. The American psychologist Henry Goddard referred to tests based on Binet and Simon’s assessment as intelligence tests, and in 1912, the German psychologist William Stern came up with intelligence quotient, or the now-familiar IQ. These new tests were sometimes used for diagnostic purposes in schools and the armed forces, but they were deployed more vigorously in eugenic breeding programs (designed “to produce a highly-gifted race of men”) and anti-immigration screening (with some would-be immigrants arriving at Ellis Island determined too “feeble-minded” to enter the United States).³

Concepts of Intelligence

The divergence between Binet’s concept of his assessment as a diagnostic educational tool and the use of IQ scores to promote eugenics and screen out immigrants comes down to a modern idea called essentialism. In the century that has followed since intelligence tests were first developed, we have become accustomed to thinking about IQ as a thing, an essence, that people carry around with them and that explains their performance in school, on the job, and elsewhere. This is an essentialist concept of intelligence, so named because it holds that intelligence is an essence within the individual. This view is incorrect.

Intelligence tests are not measures of an essence. Intelligence tests measure people’s tendency to give correct answers to questions—for example, spelling words correctly, recalling information accurately, and solving abstract puzzles quickly. But an intelligence test is not some kind of mental X-ray machine, identifying an inner quality that explains people’s performance in the real world. IQ tests are a description of the fact that some people are more accurate thinkers than others, not an explanation of why they are. This is the non-essentialist conception of intelligence: An intelligence test score is a statistically sophisticated summary of a person’s tendency to answer questions correctly.

Once you know where an IQ score comes from, it’s easy to see that it doesn’t explain anything. If someone asks, “Why did that 10-year-old excel on that algebra test?,” you might be tempted to look up her IQ score, see that it is high, and answer “She must be very intelligent.” But this explanation is circular. We know she’s intelligent because she answered a lot of questions correctly on an IQ test. Now she’s answering more questions correctly on the math test. Intelligence, like “health,” is not an essence⁴ that causes things. It is a description.

Insisting that IQ is not an essence does not dismiss its importance: Being able to spell correctly, do math in your head, and reason accurately are important skills that we want our children to have. You sometimes hear, “Scoring well on an intelligence test only means you’re good at taking intelligence tests.” That’s not true. Good performance on intelligence tests predicts positive outcomes in a wide variety of other domains. Higher scores on intelligence tests predict better grades in school and better job performance in a variety of careers.⁵ At least within North American and European cultures, intelligence tests predict success for everyone, regardless of income, wealth, or gender.⁶ Even though intelligence tests don’t explain why some people do better than others in school or on the job, they are effective at predicting who will perform better.

Why Did an Incorrect Concept Take Hold?

Two developments in the early history of intelligence research contributed to essentialist thinking. The first was the discovery that all cognitive ability tests are positively correlated. The second was advances in genetic research.

It is natural to think of human ability as a complex profile in which some of us are good at some things and some of us are good at others, but in the long run it all evens out; if you’re bad at math, for example, you’re probably good with words. That is true, to some extent—some people are better with numbers, whereas others are better with words—but on average, it’s actually the case that people who get higher scores on a test of one skill also get higher scores on tests of others. In fact, a study from 2008 of 1,800 Americans who took the 15 ability tests that constitute the Wechsler Adult Intelligence Scale IV found that scores on all 15 tests were positively associated.⁷ Even though the tests measured abilities as different as vocabulary, arithmetic, spatial thinking,* and working memory,^† scoring well on one test was associated with scoring well on the others.

This positive correlation was well established by the 1920s, and the psychologist Charles Spearman developed a statistical method called factor analysis that quantifies this tendency.⁸ He named the statistical factor that describes the ability to perform well across a wide range of cognitive tests g, for general intelligence. Because it is tempting to think of g as an essence diagnosed by IQ tests, g probably led more people to adopt the essentialist view of intelligence. But that’s not what IQ means; it is simply a number that summarizes a person’s scores across a wide range of skills. (Again, we are not saying that a high IQ score is unimportant. Doing well across a wide range of skills is a desirable trait.)

Progress in the study of human genetics also seemed to favor essentialism. You might remember learning about Gregor Mendel’s pea plants, which involved single genes (dominant or recessive) that were expressed in the plants by a trait being present or absent. Quantitative genetics is concerned with traits like height or intelligence that are not just present or absent but can have a continuum of values. These continuous traits are determined by multiple genes and environmental factors. New techniques in quantitative genetics were developed side by side with the study of intelligence in the early 20th century, by many of the same researchers.

One method they developed was to compare the intelligence of people who varied in the similarity of their genetic makeup. Identical twins share 100 percent of their genes, of course. Siblings and fraternal twins share 50 percent, on average, and half-siblings share about 25 percent. Adoptive siblings share none of their genes, on average. Research produced a somewhat surprising finding: The more genetically related people are, the more similar their IQ scores were, even after controlling for environmental factors like family wealth.⁹

This phenomenon was anointed with an essentialist-sounding name: Heritability. Heritability sounds as though it means “passed in the genes from parents to children.” Under some circumstances—like some features of farm animals raised under tightly controlled conditions—that description might not be a bad fit. But in humans, living and reproducing as they please, heritability means much less than that. Here’s why: Everything is heritable. The likelihood that you’ll get a divorce is heritable. How often you go to church is heritable. So is basketball skill.

But that doesn’t mean these heritable traits were passed along genetically. The effects can be indirect. For example, genes might influence (but not determine) your height (which is also impacted by factors like nutrition and health as a child), and being tall might give you a slight edge over other kids when you play basketball. You start out a little better than other kids, so you’re picked first when choosing teams during a recess basketball game. That makes you like basketball more. So you practice, which makes you better, which in turn brings more pleasure. It’s the practice that makes you good at basketball, and your height (the genetic effect) nudged you toward practicing.¹⁰

Intelligence may be the product of dozens or hundreds of such indirect effects. Heritability applies to the averages of large groups of people, where all of these effects get combined, so their net effect is easy to observe. But the relationship breaks down when you try to describe the particular causes in an individual case.

During the 20th century, researchers seeking to explain the heritability of intelligence had to rely on methods like comparing identical and fraternal twins. But in the last 25 years, the human genome has been sequenced, and we have much-improved images of the human brain as it thinks. Scientists have mounted an effort to estimate essential intelligence using only brain images or DNA sequences. The results, while “statistically significant,” are nowhere close to a level where they might be informative or useful (or dangerous, as in a new eugenics movement), and there is little indication that this state of affairs is going to change anytime soon.¹¹

For now, what scientists know is the obvious: Differences in intelligence arise from processes that involve genes, brains, and environments in complex ways that we can’t untangle in individual people. Is that 10-year-old algebra whiz a prodigy because of her genes, her teachers, her nutritional regimen, her tiger mom, or some combination? Neither scientists, teachers, nor parents have any idea.

Still, we have to add a caveat similar to those we attached to IQ: We are not saying that genetics has nothing at all to do with why some people perform better on intelligence tests. Rather, we are saying that we have no specific scientific knowledge of genetic reasons that one child performs better than another, and we don’t appear especially close to discovering any.

Can Teachers Increase a Child’s IQ?

The way we’ve conceptualized IQ, this question amounts to asking, “Can teachers succeed in making children better thinkers and learners?” Once you ask it that way, two answers are obvious:

1. Of course we can.
2. It ain’t easy.

Alfred Binet put it this way: “With practice, training, and above all, method, we manage to increase our attention, our memory, our judgment, and literally to become more intelligent than we were before.”¹²

But of course, the relationship between schooling and intelligence is not easy to study because it works both ways: Education improves thinking, and thinking well increases the tendency to stay in school longer. With the right methods, however, we can pull apart these reciprocal effects.

Here’s an example of how it can be done. Naturally, children can answer more questions correctly as they get older, but is that because they have been to school longer or because they have had more experiences outside school from which they’ve learned? We can examine the relative contributions of in-school and out-of-school learning to IQ by capitalizing on school enrollment age cutoffs. Suppose that to enroll in kindergarten, a child must be five years old by September 30. Every June, we can test the IQs of a group of children who are very close in age, but some of whom started kindergarten a year later (because their birthday was just after the cutoff). Is the average IQ of the pre–September 30 children (who have attended one more year of school) higher than their later-birthday peers?

In a landmark 2018 paper, researchers analyzed studies using this method and two other research designs that enable researchers to estimate the impact of schooling on IQ.¹³ Combining the three methods, they concluded that attending school does make you smarter. By their reckoning, each additional year of education improves IQ between one and five points.

Other interventions known to increase intelligence include transferring to a better-resourced, more challenging school, or increasing the number of years of compulsory education.¹⁴ But if we look for a more general description of environmental changes that increase cognitive performance, we learn three things: Effective changes start early, are comprehensive, and last a long time, ideally forever. Although it is obviously not something that could be broadly implemented to make kids smarter, the best example of effective environmental change is adoption into a family situation that provides rich opportunities for learning.

There is a long-standing scientific literature demonstrating that adoption from impoverished circumstances into middle-class homes is associated with IQ advantages of at least 15 points (compared to the adoptee’s parents or non-adopted siblings).¹⁵ Even in modern Sweden, where legislative choices have established a strong social safety net and therefore poverty is relatively rare, adoption into more economically advantaged homes is associated with significant increases in IQ.¹⁶ Of course, no one proposes that children from low-income families should be adopted to improve their intelligence, but these studies are important for showing the power of an environment that offers abundant learning opportunities and supports.

Another reason to believe that cognitive ability can improve is the Flynn effect, perhaps the most dramatic and well-established finding in the history of intelligence research. The researcher James Flynn analyzed the results of intelligence tests administered from the 1930s to the 1980s and found that average intelligence had increased by between 5 and 25 points per generation.¹⁷ These increases cannot possibly be genetic in origin. Large-scale genetic changes in human populations take thousands of years, yet IQ scores have risen dramatically within just a few decades.

The reasons for the Flynn effect are complicated, but a consensus has emerged that they involve the modernization of society. In 1910, most Americans worked on farms or in manufacturing; thus, the cognitive challenges they met tended to be the same, day after day. In modern economies, more and more jobs present a stream of novel problems. What’s more, those problems increasingly require the mental manipulation of abstract symbols like numbers, charts, and specialized concepts used in fields like finance, engineering, and law. The ability to manipulate abstract concepts is largely what IQ tests measure. Our abilities change and improve in response to the world in which we are immersed.

What Does All This Mean for Teachers?

Some children show up on day one ready and eager to learn. They devour the educational sustenance that is provided to them. Others seem entirely unready, and our greatest efforts produce only minimal gains. The first kind of child gets a high score on an IQ test, and the second gets a low score. Isn’t the problem that children with a low IQ can only be expected to learn so much?

The answer is threefold. First, as we have described, an IQ score is not an explanation of why some children do better in school than others; it is a description of it, a quantification. Second, even if it is the case that some kids are born with more effective brains than others, there is no way to diagnose this in individual children, certainly not with intelligence tests. IQ tests measure where a child is, not where they could potentially be. Third, the fact that cognitive ability and IQ scores are difficult to change is not an indication that they are baked into the genes at conception. Positive or negative behavioral patterns are always difficult to change once they are established, no matter where they came from. That’s true of how well children answer questions, how they interact with peers, how they cope with difficult emotions, and more.

So here is our bottom line: Intelligence tests are a useful way to describe and quantify differences in cognitive functioning among people in general and schoolchildren in particular. In the hands of a qualified psychologist, they can be used to identify individual strengths and weaknesses, and they do a good job of predicting future performance—but not potential. If circumstances change, outcomes may change too.

Intelligence tests do a good job of predicting future performance not because they provide insight into mental capacity, but because the circumstances that led to a high IQ score in the first place are likely to persist. Children who have had access to environments that provide rich opportunities for learning will probably have access to those sorts of environments in the future. The child who has had a lot of success with academic tasks will likely have more confidence when taking on tasks in the future and persist longer when tasks are difficult. These processes no doubt combine with preexisting genetic differences to some extent, but the two cannot be meaningfully disentangled, especially in individual children.

Making significant improvements in anyone’s thinking is never easy, and it is more difficult with children who have had less success in the past. But it’s helpful to know that a low IQ score doesn’t equate to a hard-and-fast limit on what a child can achieve. And whether the child has a history of success or struggle in the past, the broad guidelines for future success are the same: Start as soon as possible, be comprehensive, and persist.     

Eric Turkheimer is the Hugh Scott Hamilton Professor of Psychology at the University of Virginia. He is a past president of the Behavior Genetics Association and a fellow of the American Academy of Arts and Sciences. In 2024, he won the Dobzhansky Founders Lifetime Achievement Award for outstanding lifetime contributions to the field of behavior genetics. His most recent book is Understanding the Nature‒Nurture Debate. Daniel T. Willingham is a professor of cognitive psychology at the University of Virginia. He is the author of several books, including the bestseller Why Don’t Students Like School? and Outsmart Your Brain: Why Learning Is Hard and How You Can Make It Easy. Readers can pose questions to “Ask the Cognitive Scientist” by sending an email to ae@aft.org. Future columns will try to address readers’ questions.

*To learn about spatial thinking, see “Picture This: Increasing Math and Science Learning by Improving Spatial Thinking” in the Summer 2010 issue of American Educator: go.aft.org/4th. (return to article)


^†For an explanation of working memory, see “How Can Educators Teach Critical Thinking?” in the Fall 2020 issue of American Educator: go.aft.org/3su. (return to article)

Endnotes

1. M. Brysbaert and S. Nicolas, “Two Persistent Myths About Binet and the Beginnings of Intelligence Tests in Psychology Textbooks,” Collabra: Psychology 10, no. 1 (2024): 117600.

2. A. Binet and T. Simon, “Méthode Nouvelle pour le Diagnostic du Niveau Intellectuel des Anormaux,” L’Année Psychologique 11 (1905): 191–244.

3. L. Benjamin, A Brief History of Modern Psychology, 4th ed. (Hoboken, NJ: Wiley, 2024); and L. Benjamin, “The Birth of American Intelligence Testing,” Monitor on Psychology 40, no. 1 (2009): 20.

4. S. Heine, B. Cheung, and A. Schmalor, “Making Sense of Genetics: The Problem of Essentialism,” Hastings Center Report 49 (May 2019): S19–S26.

5. A. Demetriou et al., “Cognitive and Personality Predictors of School Performance from Preschool to Secondary School: An Overarching Model,” Psychological Review 130, no. 2 (2023): 480–512; D. Fergusson, J. Horwood, and E. Ridder, “Show Me the Child at Seven II: Childhood Intelligence and Later Outcomes in Adolescence and Young Adulthood,” Journal of Child Psychology and Psychiatry 46, no. 8 (2005): 850–58; and F. Schmidt and J. Hunter, “The Validity and Utility of Selection Methods in Personnel Psychology: Practical and Theoretical Implications of 85 Years of Research Findings,” Psychological Bulletin 124, no. 2 (1998): 262–74.

6. R. Nisbett et al., “Intelligence: New Findings and Theoretical Developments,” American Psychologist 67, no. 2 (2012): 130–59.

7. D. Wechsler, Manual for the Wechsler Adult Intelligence Scale-Fourth Edition (WAIS-IV) (San Antonio: Pearson, 2008).

8. C. Spearman, The Abilities of Man (London, UK: Macmillan, 1927).

9. T. Bouchard and M. McGue, “Familial Studies of Intelligence: A Review,” Science 212, no. 4498 (1981): 1055–59.

10. W. Dickens and J. Flynn, “Heritability Estimates Versus Large Environmental Effects: The IQ Paradox Resolved,” Psychological Review 108, no. 2 (2001): 346–69.

11. S. Gusev, “No, Intelligence Is Not Like Height,” The Infinitesimal (blog), August 26, 2024, theinfinitesimal.substack.com/p/no-intelligence-is-not-like-height.

12. A. Binet, Modern Ideas About Children, transl. S. Heisler (Menlo Park, CA: Suzanne Heisler, 1975).

13. S. Ritchie and E. Tucker-Drob, “How Much Does Education Improve Intelligence? A Meta-Analysis,” Psychological Science 29, no. 8 (2018): 1358–69.

14. Ritchie and Tucker-Drob, “How Much Does Education Improve Intelligence?”

15. R. Nisbett, Intelligence and How to Get It: Why Schools and Cultures Count (New York: W. W. Norton & Company, 2009).

16. K. Kendler et al., “Family Environment and the Malleability of Cognitive Ability: A Swedish National Home-Reared and Adopted-Away Cosibling Control Study,” Proceedings of the National Academy of Sciences 112, no. 15 (2015): 4612–17.

17. J. Flynn, “Massive IQ Gains in 14 Nations: What IQ Tests Really Measure,” Psychological Bulletin 101, no. 2 (1987): 171–91.

[Illustrations by Fran Pulido]