Kathryn Paige Harden argues how far we go in formal education – and the huge knock-on effects that has on our income, employment and health – is in part down to our genes. Harden is a professor of psychology at the University of Texas at Austin, where she leads a lab using genetic methods to study the roots of social inequality. Her provocative new book is The Genetic Lottery: Why DNA Matters for Social Equality.
To even talk about whether there might be a genetic element to educational attainment and social inequality breaks a huge social taboo – particularly on the political left, which is where you say your own sympathies lie. The spectre of eugenics looms large, and no one wants to create a honeypot for racists and classists. To be clear, it is scientifically baseless to make any claims about differences between racial groups, including intelligence, and you are not doing that. But why go here?
I wrote this book first for my fellow scientists, who haven’t necessarily seen the relevance of genetics for their own work or have been afraid to incorporate it because of these associations. There is a large body of scientific knowledge being ignored lest the eugenics genie be let out of the bottle.
But also people are hearing every day about new genetic discoveries and seeing in their own families and lives that genetics matter. When asked to estimate how much genes influence intelligence, people’s answers are not zero. I’m trying to help them make sense of that information in a socially responsible way. If you care about social equality, what do you do with information about genetics?
You have been accused of promoting eugenics, including by prominent sociologist Ruha Benjamin, who has written that you are engaging in “savvy slippage between genetic and environmental factors that would make the founders of eugenics proud”.
Those fears are coming out of a very real place – historically, genetics has been misused. But [eugenics] is literally the opposite of what I’m advocating. The core idea of eugenics is that there is a hierarchy of people who are inferior or superior that is rooted in biology and that inequalities are justified on that basis. Mine is an anti-eugenics approach seeking to use our knowledge of genetic science to build policies and social interventions that create more social equality. Sweeping genetic differences between people under the rug does not make the genome, as a systemic force causing inequality, go away. That genetic and environmental factors are braided together at every level is simply a description of reality.
How do you predict a person’s educational attainment via their genome?
It starts with a statistical exercise in correlation called a genome wide association study (GWAS). That takes many hundreds of thousands of people with similar genetic ancestry and measures tiny genetic differences – of which there are millions – scattered throughout their entire DNA sequences. It then looks to see which of those variants correlates with their number of years of schooling.
We then take the results and – for a new person’s genetic sequence – add up that information to produce a single number, a polygenic score, that predicts how far they will go in school.
Crude though it is, the GWAS approach has found genetic variants that are correlated with going further in school. That isn’t surprising – we see evidence that there’s a genetic influence on academic achievement in twin studies. Identical twins are more similar in how far they go in school than fraternal twins.
How many variants are we talking about and what is the size of the effect?
Scientists have identified more than 1,000 genetic variants spread over the entire genome, each of which has a tiny effect. Taking the combined influence, it captures about 10-15% of the variance in educational attainment. The rate of college graduation is nearly four times higher for people who have a high compared with a low polygenic score. That competes with other variables we think of as important for educational attainment, such as family income, which has an effect size of about 11%. But it is still lower than the twin study estimate that about 40% of the variation in educational attainment is due to genes.
Can we say differences in educational attainment are caused by our DNA? Correlation does not equal causation and we know the environment makes a huge difference.
We’re reasonably certain at this point that the causal genetic influence is not nil. It is the size that is at issue. There are questions with the twin studies about whether they are attributing to genes what should really be claimed by the environment. And for polygenic score studies, people may just happen to differ genetically in ways that match environmental factors, and it is really those that are driving the effect.
More confidence in our conclusions comes when we get similar answers with different methods. Polygenic score studies within families are now also suggesting genetic cause. For example, studies of siblings who are raised in the same environment, but who are more different in their polygenic scores, show that these siblings have more different life outcomes.
Are people who have these genes more intelligent?
The word “intelligence” is a lightning rod, because it is so easily misrepresented as being a marker of all human skill. But it’s clear that formal schooling in the US and UK reinforces a very particular type of reasoning. And it is the same type of reasoning that IQ tests also pick up on.
But we have also done a genetic study that found there is a basket of non-cognitive, personality related abilities helping pull people through school – being conscientious and open to new experiences, for example. Anything that makes you more likely to get to the next stage of your education, to the extent that is reflected in your biology, a GWAS is going to pick it up. Importantly, people with these genes don’t have “good” genes. They have genetic variants that happen to be correlated with going further in school as it is currently constructed.
Will we be rushing to read our children’s genomes to discover their polygenic scores in the future?
People’s imaginations jump to this world of individualised testing and tailored interventions. I don’t think this knowledge is best used as a diagnostic tool about an individual person. There is always a danger that people will be given bad or incomplete information. I want to use genetics as a way of seeing what’s happening within our environments and social structures better.
How should this knowledge be applied, then?
One of the most useful applications is in improving the basic research we do to design our wider policies and interventions for everyone. There are many policy initiatives, and more are being proposed all the time. But their research base is limited because it assumes children only receive environments from their parents, and not anything genetic.
Consider, for example, policies to close the famed “word gap”, which is the estimated 30-million word difference in what poor children versus children from high income families hear before they turn three. The jury is still out on whether “word gap” interventions will be effective, but one glaring problem is the same vocabulary outcomes that are allegedly the outcomes of being exposed to more speech could also be the result of genetics. Parents and children share genes and the same genes that are associated with adults’ educational attainment and income are also associated with early acquisition of speech and reading in their children. Before we spend millions on interventions designed to change a parental behaviour in the hopes of improving child outcomes, it would be prudent to at least check this effect out.
Ruha Benjamin also suggested the hunt for more data to explain things just ends up being a barrier to acting on what we already know we need to do to fix the academic achievement gap…
I disagree that we already know what to do. If you look at meta-analyses of educational interventions, you see most of their effect sizes are zero. Most of the things we try in education, even when they are well intentioned and well funded, make no difference to students’ lives. It is a fiction we have this army of effective, scalable solutions just waiting in the wings. Figuring out what works for whom and when is very hard. The risk of not talking about genetics is continuing the status quo, where we are much less effective at intervening than we could be.