A new study by Ben Domingue, assistant professor at Stanford Graduate School of Education, finds that genetic differences between siblings is associated with how many years of education achieved.
The study, published Aug. 20 by AERA Open, a peer-reviewed journal of the American Educational Research Association, found that, within families, an adolescent with a higher “polygenic score” —which summarizes previously identified genome-wide associations for educational attainment—than her or his sibling tended to go on to complete more years of schooling.*
While the predicted difference in actual educational attainment between siblings was small—roughly one-third of a year of schooling—the study provides new evidence that recently discovered genetic factors actually do cause differences in educational outcomes, Domingue said.
“By examining siblings, this study was able to control for external social aspects, such as schools, neighborhoods, and level of parental education, to hone in specifically on the role of genes in this complex process,” said Domingue in a news release from the AERA. “The study provides strong evidence that genotype can predict educational attainment within families.”
The full AERA release is here. We talked to Domingue about his findings and how genetic research in an educational context could impact policy decisions. You'll find that edited interview below.
What does that mean that genotype may predict educational attainment?
Domingue: That's a good question and I think that the most honest answer is that we still don't know. A next step is to learn more about the intermediate personal characteristics that develop as a result of the polygenic score, which ultimately lead to more educational attainment. There is some evidence to suggest that the polygenic score also predicts cognitive functioning - so those with higher polygenic scores tend to have more cognitive ability and in turn this translates to additional years of educational attainment. But there are still many unanswered questions at this point. For example, we don't know if the score is associated with other personality traits (for example, "grit") that one might think matter in determining how far one goes in school.
What's the best way to interpret your study's findings?
Domingue: I think it's a good introduction with respect to the evidence that science is beginning to accumulate about how genetics shape our lives albeit in ways that are only detectable when you have pretty large groups of people. The medical field is already grappling with issues related to genetics and how to deal with this information. I think education is going to have to deal with this problem in the long run as well. Given the wealth of genetic research on psychiatric disorders, I think it's quite likely that schools will be approached in the not-so-distant future by parents worried about their children's genetic predisposition towards, for example, ADHD. How should the school handle, say, a request for early intervention because of this genetic risk? I don't have the answer to that, but I think that a first step in the process is understanding what genetic risk is and isn't, and how this might change when we're measuring different outcomes.
What don’t you want people to take away from the findings?
Domingue: It's crucial that people don't see this score as being deterministic. One of our biggest goals in this research was to properly characterize for the education research community what this score "means". And it most certainly doesn't mean an insurmountable "head start" towards success in life.
It's also crucially important to recognize that the environmental context in which the biology is happening is crucial. Body mass index (BMI) is a great example. Weight and height (and thus BMI) are both strongly heritable, strongly linked to genetics. However, in the last 30 years there has been an absolutely massive shift in the distribution of BMI in this country. The underlying genetics didn't change during that time period in a substantive way (this isn't a result of evolution), and so it really emphasizes that a genotype isn't a fixed destiny.
How do these findings add to previous research on the subject?
Domingue: This is the first study to demonstrate that the score continues to predict educational attainment in a nationally representative sample of younger Americans. Second, the score is associated with neighborhood characteristics. And finally, the score predicts educational attainment in African Americans as well (most earlier work focused on European Americans or Europeans).
Although the sibling model has been used in this context before, it is worth re-iterating that evidence from the sibling model is strong in the sense that sibling controls allow us to remove things that don't vary within families from the list of potential "alternative" explanations.
How could this study potentially change how we approach policies aimed at improving educational outcomes?
Domingue: A strand of research that I'm excited to undertake is to start asking additional questions about how different environments, particularly schools, may moderate the genetic effect. So, for example, do we see some schools reducing the effect of genetics? If so, what about the school environment might explain this?
With other traits (for example, smoking) there is some evidence to suggest that public policies have consequences for how genetics are associated with a given outcome. I think it will be fascinating to see if the same thing is true in the context of educational attainment.