Getty Images

Recent discoveries in genetic sciences have the potential to revolutionize how schools help students achieve their potential.

Since James Watson and Francis Crick announced the discovery of the double helix structure of DNA, in 1953, scientists have delved into the genetic code to improve people’s lives in innumerable ways. Genetically modified foods have increased crop yields exponentially. Law enforcement regularly uses DNA to identify criminals and exonerate the innocent. Physicians can now use DNA tests to identify patients who are at risk for breast cancer, Alzheimer’s disease, and other maladies before symptoms appear. Remarkably, the DNA revolution has even created entirely new disciplines. Forensic genealogy, for example, uses genetic data from crime scenes to trace the identity of unknown criminals or deceased victims through DNA samples voluntarily submitted to genealogical databases by distant relatives. And molecular behavioral genetics identifies segments of DNA associated with psychological traits and behaviors, including mental health diagnoses and personality traits.

So far, applied genetic science has left K-12 education untouched — but not for long. Recent findings in behavioral genetics and technological advances have the potential to change schooling in ways that were inconceivable just a decade ago. In particular, one scientific finding (the heritability of educational outcomes) and one new technology (polygenic scores) are likely to have dramatic effects on teaching and learning, perhaps making it far easier to help every child reach their full educational potential. But educators ought to prepare themselves for how these advances are likely to be applied in the schoolhouse, both to take advantage of their great promise and to guard against their potential risks.

The heritability of educational outcomes

The term heritability refers to the degree to which genetic variations lead to variations in physical traits, psychological characteristics, and life outcomes. Decades of behavioral genetics studies have led experts to conclude that almost every observable human characteristic is heritable to some extent (with a few narrow exceptions, such as religious denomination). Our DNA doesn’t determine our fate, but it does have at least some influence on most aspects of our lives. The evidence for this finding is so strong that researchers have termed it the First Law of Behavioral Genetics (Turkheimer, 2000).

Expressed as a proportion, heritability ranges from 0 (indicating that differences are caused entirely by the environment) to 1 (when differences are caused entirely by genetics). For instance, traits with typically high heritability include attention deficit/hyperactivity disorder (.68), obesity (.63), and allergies (.63). Traits in the middle range include intelligence (.51), Type I diabetes (.50) and blood pressure (.47), while depression is on the low end (.34), shaped more strongly by environmental factors than by one’s genes. When it comes to heritability’s influence on personal behaviors and life outcomes, examples include educational attainment in adulthood (.52), tobacco addiction (.44), and health self-care (.41) — all are moderately influenced by genetics. (All values from Polderman et al., 2015.)

Educators are likely to be particularly interested in the finding that academic achievement has a heritability value of .50 or higher in most studies (e.g., de Zeeuw et al., 2016; Grasby et al., 2019; Krapohl et al., 2014; Little, Haughbrook, & Hart, 2017; Rimfeld et al., 2015). This means that the most important reason that children differ from one another in academic achievement is genetic. All environmental reasons combined explain 50% or less of variance in academic outcomes.

The science allows us to estimate the average level of influence that genes have in determining differences within a group, not in determining the actual existence of a trait.

Indeed, the contribution of genetics to academic outcomes helps explain one of the most powerful findings in educational research. The 1966 Coleman Report showed that individual student characteristics have a stronger relationship with academic achievement than do school-level variables. That’s no surprise when considering that all environmental factors put together have only the same degree of influence on students’ achievement as do genetics. Since student-level characteristics include both the child’s genes and a range of other environmental influences (e.g., neighborhood characteristics, family income level, and so on), it stands to reason that they account for significantly more than 50% of students’ differences in performance and that schools will tend to have somewhat less influence.

There are also important caveats about the research findings to be aware of. For educators in particular, it is important to recognize that each heritability value is a population-level statistic. It doesn’t make sense to point to a specific student and say (for example) that 60% of their reading achievement is because of their genetics and 40% is because of their environment (Warne, 2020a). Rather, the science allows us to estimate the average level of influence that genes have in determining differences within a group, not in determining the actual existence of a trait. In a typical 2nd-grade classroom, for instance, individual students are likely to range widely in their reading levels. But if those children have been receiving the same instruction from the same teacher at the same time, then what accounts for their differences in achievement? A heritability of .60 would indicate that their genetic differences explain 60% of the variation seen in the given classroom, and the remaining 40% stems from environmental differences (including both in- and out-of-school factors, such as how often their parents read with them, how well they get along with the teacher, and so on).

When discussing academic achievement, it’s also important to know that heritability values tend to increase as children age. In one recent study, heritability for math and reading scores was 0 to .255 at age four and .185 to .478 at age five — that is, as kids get older, their DNA starts to have more influence on their performance, increasing by approximately .07 per grade (Little, Haughbrook, & Hart, 2017) and reaching .60 or higher by the end of high school (Rimfeld et al., 2015). Thus, genetic influences on academic achievement grow over time — perhaps because the later grades create an environment in which genetic differences are more evident — and environmental influences weaken in importance as children advance through the education system.

Genome-wide association studies and polygenic scores

Heritability values can help quantify the relative importance of genes and environment. In and of itself, however, a heritability statistic reveals nothing about which genes affect which sorts of academic achievement (or any other trait) or where on our chromosomes those genes might be located. Currently, the best method for identifying specific genes that may influence phenotypes (i.e., people’s individual characteristics) is called a genome-wide association study (GWAS). In a GWAS, researchers collect information about a trait — such as educational achievement — and DNA samples from many individuals (often hundreds of thousands). Then, the researchers examine thousands of genetic variations on people’s DNA to identify whether any variations are associated with specific scores on the trait being studied. For example, in a GWAS study on reading achievement, researchers would collect reading scores and DNA and then identify locations within the genome where one variant of DNA is more common in people with high reading achievement and a different variant is more common in people with low reading achievement.

The first GWAS for a nonbiological phenotype was published in 2013 (Rietveld et al., 2013). But even at this early stage, the results are impressive. To date, the largest GWAS for an educational outcome has identified 1,271 genetic variations that are correlated with educational attainment, defined as the total number of years of schooling an adult has had (Lee et al., 2018).

GWASs are the key to understanding how the genetic revolution can influence K-12 education. Once genetic variations are identified, this information can be used to make predictions called polygenic scores. (See Plomin & von Stumm, 2018, for an explanation of how polygenic scores are calculated.) Polygenic scores simplify a person’s genetic variation on any heritable trait into a single number that can then be used to make predictions about that person’s educational outcomes. One of these polygenic scores — for educational attainment in adulthood — is already a better predictor than parental income (Lee et al., 2018). Today, in other words, a DNA test at birth can give us a sense of how much education that baby will eventually obtain, and this prediction will be more accurate than one based on their parents’ salaries.

This isn’t to say that the child’s environment doesn’t matter, but only that the amount of information we can get from genetic analysis has increased dramatically. At this point, when we combine polygenic scores with information about childhood socioeconomic status, we can make predictions about the child’s long-term educational outcomes that are more accurate than using either predictor alone (von Stumm et al., 2020).

The promise and limitations of polygenic scores

It is the predictive power of polygenic scores that is likely to have a dramatic effect on K-12 education. In theory, polygenic scores can be calculated for almost any outcome, with predictions expected to be more accurate for outcomes that are more susceptible to genetic influence. For instance, DNA-based predictions of 6th-grade standardized achievement test scores (heritability of .74) will likely be more accurate than predictions about students’ study skills in 6th grade (heritability about .60). Likewise, 1st-grade spelling achievement (heritability around .35) is expected to be poorly predicted from genetic data because it is more susceptible to the influences of environment than genetics. (All heritability values from de Zeeuw et al., 2016.)

Overall, though, the educational potential for polygenic scores is immense. For example, one likely application is to use genetic analysis to identify young children who are at risk for disabilities (von Stumm et al., 2020). If parents and physicians learn that an infant is likely to develop symptoms of autism, they can begin to consider therapies and interventions long before the child displays social difficulties. Children at risk for a learning disability can begin to receive supports before they fall behind their peers. In middle and high school, DNA-based predictions can be used to plan interventions for students who are likely to develop mental health problems and drug addictions, thereby reducing the risk that a crisis will derail their schooling. Note that in each of these cases, the genetic prediction is not treated as destiny, but instead as an indicator of the kinds of support the child might need to help them succeed.

A DNA test at birth can give us a sense of how much education that baby will eventually obtain.

So, too, could polygenic scores be used, in combination with other sources of data, to assign children to gifted programs. Unlike standardized tests, parent nominations, or grades from teachers, polygenic scores do not disadvantage non-English speakers, are impervious to cheating, cannot be raised with expensive study programs, are not biased against students from racial or ethnic minority groups, do not fluctuate over time, and do not yield to the influence of pushy or well-connected parents. In other words, polygenic scores can overcome almost all the imperfections that characterize current methods of selecting children for gifted programs.

Polygenic scores are not yet strong enough to generate these sorts of predictions with a high degree of confidence, and for that reason, I join with experts who warn against using current polygenic scores to make educational decisions for individual students (e.g., von Stumm et al., 2020). Further, polygenic scores are not yet portable across racial and ethnic groups. Today, most GWAS samples consist mainly of DNA from people of European descent, and they may be less accurate when used to make predictions about non-white students (Domingue et al., 2015). That’s a second strike against the current use of polygenic scores.

As GWAS technology improves, however, polygenic scores will only grow in sophistication and become more accurate (Plomin & von Stumm, 2018), not to mention less costly to obtain. Thus, I have no doubt that their use will become routine in the coming years, in both K-12 education and many other sectors. In short, we should already be thinking about how best to take advantage of these tools, while also considering how they could be abused. In law enforcement, for example, the practice of identifying criminals with the aid of relatives’ genetic data posted online has already raised serious concerns about privacy (given that anybody who shares their own genome is also sharing the partial genomes of their biological relatives, likely without their permission). Likewise, medical ethicists have warned that polygenic scores could be obtained during the process of artificial insemination, allowing prospective parents to select the embryo(s) with the greatest likelihood of having desirable traits — that is, they may be tempted to practice a form of eugenics.

Ideally, Americans will look to the democratic process to determine how to move forward ethically with the use of genetic data. To an extent, this has already begun. For example, due to concerns that people might be discriminated against because of their genetic heritage, the U.S. Congress passed the Genetic Information Nondiscrimination Act in 2008, banning the use of genetic information in employment decisions. Inevitably, some other uses of genetic data will be banned in the coming years, while other uses will be carefully regulated and still others will be permitted freely.

To date, however, few educators have even begun to discuss these issues (if they’re aware of them at all). It is easy to imagine harmful and unethical uses of genetic information by school systems, though. For instance, genetic predictions could be used to bar qualified children from special education services if their polygenic scores do not predict they will need them; likewise, they could be used to bar otherwise qualified children from gifted education programs. So, too, could genetic data be used to assign students to educational tracks, restricting their opportunities. For that matter, merely informing a child of their polygenic scores could be detrimental to their motivation and/or sense of self-worth, particularly if those scores suggest that they’ll be likely to struggle in certain areas.

While such concerns should be taken very seriously, none of these detrimental uses of polygenic scores is inevitable. It would be a mistake to allow our worst fears to determine how we proceed, especially if that means banning all use of genetic data in schools. That would only exclude the educational system, and the children it is supposed to help, from the many benefits that genetic research has started to bring to other fields.

Moreover, it’s not as though we should be content with our existing approaches to making decisions about student support services, enrichment opportunities, and so on. The use of polygenic scores might have negative consequences, but we know for sure that our current practices (relying excessively on standardized tests, responding to pressures from well-connected parents, allowing teachers to make gut-level decisions about individual children) are often unfair and ineffective.

Suggested policies

The current limitations of GWASs are not unsolvable. For example, while much of the existing DNA research has focused on people of European descent, genome repositories are quickly becoming more diverse, given researchers’ desire to extend the medical and scientific benefits of GWASs to populations worldwide. Meanwhile, larger and more diverse samples and more exhaustive DNA sequencing are enabling genetic researchers to identify more segments of DNA that are associated with specific educational outcomes. These improvements will, in turn, make polygenic scores more and more accurate in predicting how students will perform. Although experts disagree on the precise time line (e.g., compare Murray, 2020, with Warne, 2020b), polygenic scores for some educational outcomes may be usable within the coming decade.

Education policy makers should start thinking now about how to make the best use of these fast-emerging tools. The implications of using polygenic scores in schools are wide-ranging, and decisions about their ethical use should involve experts from a variety of fields, beyond education itself. Still, based on what we now know about heritability, and given the current state of genetic technology, I have some suggestions for school and district leaders in particular:

  1. Put a moratorium on the use of all polygenic scores in education until predictions are more accurate. This should be reviewed periodically (e.g., every three or five years) to determine whether polygenic scores have improved sufficiently for use in an educational context.
  2. Keep in mind that children’s developed abilities should trump any predictions based on genetics. Regardless of what a person’s polygenic score predicts, their educational options should reflect their actual behavior and classroom performance. For example, if a child is motivated, excels in typical schoolwork, and has high test scores, then they should be admitted to a gifted program, regardless of what the polygenic score predicts.
  3. When making diagnostic decisions, don’t rely on genetics alone. A polygenic score may be used to help identify children who should be considered for a specific school support, therapy, or program, but it  should not be the only marker used to determine eligibility, and children should be evaluated regularly (probably at least once per year) to ensure that services are still needed.
  4. Refrain from using polygenic scores to help make decisions about academic pathways and programs for younger, typically developing children. In the early elementary grades, heritability values are too low to make valid predications about cognitive abilities. As children age, and the balance between heritability and environmental influence shifts, they should be regularly reevaluated to determine whether they might benefit from particular services, with polygenic scores playing a larger (but never the sole) role.
  5. Finally, limit the use of polygenic scores until those scores have been validated across the racial and ethnic groups that your district serves. For most districts, this will mean a total ban on polygenic scores for now. For others, it may be appropriate to use scores only as outside data points to consider when they happen to be available. (The district shouldn’t seek out that data, and it should take care to ensure that students from less affluent families aren’t put at a disadvantage just because their parents are unable to pay for genetic testing — much as they shouldn’t be put at a disadvantage because their parents can’t pay for independent assessments by private psychologists.)

Having these or similar policies in writing before they are needed will help education leaders ensure that they make decisions in an informed and fair manner. Cobbling together a policy quickly or making ad hoc decisions will only leave the process open to abuse, poor decisions, and widening inequalities.

What district and school personnel should not do is fear these scientific advances and institute permanent bans on polygenic scores or other genetic information. The fact is that the genetic revolution is coming to education, and educators must be informed and ready for it. And while they should be aware of its limitations and potential abuses, they should also know that these tools have brought benefits to every field they have touched.

References

Coleman, J. et al. (1966). Equality of educational opportunity. U.S. Department of Education, National Center for Educational Statistics.

de Zeeuw, E.L., van Beijsterveldt, C.E.M., Glasner, T.J., de Geus, E.J.C., & Boomsma, D.I. (2016). Arithmetic, reading and writing performance has a strong genetic component: A study in primary school children. Learning and Individual Differences, 47, 156-166.

Domingue, B.W., Belsky, D.W., Conley, D., Harris, K.M., & Boardman, J.D. (2015). Polygenic influence on educational attainment: New evidence from the National Longitudinal Study of Adolescent to Adult Health. AERA Open, 1 (3), 1-13.

Grasby, K.L., Coventry, W.L., Byrne, B., & Olson, R.K. (2019). Little evidence that socioeconomic status modifies heritability of literacy and numeracy in Australia. Child Development, 90 (2), 623-637.

Krapohl, E., Rimfeld, K., Shakeshaft, N.G., Trzaskowski, M., McMillan, A., Pingault, J.-B., . . . & Plomin, R. (2014). The high heritability of educational achievement reflects many genetically influenced traits, not just intelligence. Proceedings of the National Academy of Sciences, 111 (42), 15273-15278.

Lee, J.J., Wedow, R., Okbay, A., Kong, E., Maghzian, O., Zacher, M., . . . & Cesarini, D. (2018). Gene discovery and polygenic prediction from a genome-wide association study of educational attainment in 1.1 million individuals. Nature Genetics, 50 (8), 1112-1121.

Little, C.W., Haughbrook, R., & Hart, S.A. (2017). Cross-study differences in the etiology of reading comprehension: A meta-analytical review of twin studies. Behavior Genetics, 47, 52-76.

Murray, C. (2020). Human diversity: The biology of gender, race, and class. Twelve.

Plomin, R. & von Stumm, S. (2018). The new genetics of intelligence. Nature Reviews Genetics, 19 (3), 148-159.

Polderman, T.J.C., Benyamin, B., de Leeuw, C.A., Sullivan, P.F., van Bochoven, A., Visscher, P.M., & Posthuma, D. (2015). Meta-analysis of the heritability of human traits based on fifty years of twin studies. Nature Genetics, 47 (7), 702-709.

Rietveld, C.A., Medland, S.E., Derringer, J., Yang, J., Esko, T., Martin, N.W., . . . & Koellinger, P.D. (2013). GWAS of 126,559 individuals identifies genetic variants associated with educational attainment. Science, 340 (6139), 1467-1471.

Rimfeld, K., Kovas, Y., Dale, P.S., & Plomin, R. (2015). Pleiotropy across academic subjects at the end of compulsory education. Scientific Reports, 5, Article 11713.

Turkheimer, E. (2000). Three laws of behavior genetics and what they mean. Current Directions in Psychological Science, 9 (5), 160-164.

von Stumm, S., Smith-Woolley, E., Ayorech, Z., McMillan, A., Rimfeld, K., Dale, P.S., & Plomin, R. (2020). Predicting educational achievement from genomic measures and socioeconomic status. Developmental Science, 23 (3), Article e12925.

Warne, R.T. (2020a). In the know: Debunking 35 myths about human intelligence. Cambridge University Press.

Warne, R.T. (2020b). Crossing the Rubicon from the social to the biological sciences. [Review of the book Human diversity: The biology of gender, race, and class by C. Murray]. American Journal of Psychology, 133 (4), 536-543.

 

This article appears in the October 2021 issue of Kappan, Vol, 103, No. 2.

ABOUT THE AUTHOR

default profile picture

Russell T. Warne

RUSSELL T. WARNE is an associate professor of psychology at Utah Valley University, Orem. He is the author of In the Know: Debunking 35 Myths about Human Intelligence and Statistics for the Social Sciences .