Teachers matter (but not in the way we might think)

Behavioural geneticist Robert Plomin has distilled his more than 40 years of research into a controversial book that has some stark messages for education: schools, teachers and parents are not what make us different from one another. And that, he tells Jon Severs, means we have to change the way we think about learning
25th January 2019, 12:00am
Magazine Article Image


Teachers matter (but not in the way we might think)


His knee is a problem: he is awaiting an operation and his leg has to be kept straight. So his 6ft 4in frame has become a near-horizontal bridge across the small office, connecting a little black footstool that raises his left ankle and a chair that supports the majority of his weight.

To make eye contact, he has to turn his head 10cm to the left, drop his chin, lose his neck. But, even with all the scaffolding, he still shifts, winces and repositions repeatedly.

He looks awkward. Uncomfortable. And if you are one of the world's leading behavioural geneticists, one who is about to say what he is about to say - that education is the "last bastion of anti-genetics", that differences in systemic environmental influences like schools matter less than we think, that we can now predict an individual's educational future with DNA - you don't want to look uncomfortable. And you certainly don't want to look awkward.

Because that creates space for interpretation, and into that gap creep judgements, conspiracies, mutations and accusations. As with genetics itself, when you talk about genetics, the little things matter. So, Professor Robert Plomin explains the knee (sports injury, genetic predisposition to being overweight, overdoing it). He apologises for it. He controls the variable so that he can ensure the focus is on what he says, not what he looks like while saying it. That's important. What Plomin is saying at the moment is controversial, but it is a message that every teacher needs to at least consider carefully and objectively, whether they ultimately agree with it or not.

At 70 years old, celebrity has come late to Plomin. Sure, he has been featured in the broadsheets before but, since publishing his book Blueprint: How DNA makes us who we are in October, he's gone mainstream: he has appeared on Newsnight, in feature-length interviews in The Times, and even on the Steve Wright Show on BBC Radio 2.

The book summarises what he has learned in his more than 40 years as a geneticist, and it includes some of what he has said previously (his insistence that intelligence is highly heritable has been often cited in the press) and much that he has avoided saying because of, in his words, "cowardice". "The book is everything I know," he says. "I do not want to write another one."

Much of Blueprint is based on what he has discovered leading the Twins Early Development Study (TEDS), one of the longest-running and largest twin studies in the world. He set it up in 1994 when he became professor of behavioural genetics at King's College London. TEDS started with 15,000 sets of twins, of which around 13,000 still take part (he had run twin and adoption studies previously in the US, where he was born and where he worked until his appointment at King's).

Twins are crucial for behavioural geneticists because, as he writes in his book, "where you can most clearly see heredity in action is identical twins. Twins are a gift to science."

By comparing the similarity between identical twins brought up in the same environment and also studying the similarity between non-identical twins who are brought up in the same environment, you can work out the relative influence of nature and nurture (see end matter, below).

In TEDS, the comparative measures between sets of twins are extensive - for example, cognitive ability, educational outcomes, personality, health - and include interviews with parents, teachers and the twins themselves. The study currently has 55 million items of data.

So, what does all this research reveal? Put simply, everything is heritable. How clever you are, how good you are at sport, your level of aggression, even how much TV you watch.

That may not mean what you think it means, though. "Heritability is widely misunderstood," says Plomin. "It is essentially asking about the extent to which differences we see between people - in any trait, from physiological, to personality, to educational - are down to inherited DNA differences."

A useful example from Blueprint: heritability for weight is 70 per cent. But that does not mean that your weight is 70 per cent down to your DNA. Instead, it means that 70 per cent of the differences between the weight of individuals across a population - the variability - is down to differences in DNA.

Plomin is keen to stress two caveats to heritability: first, that the statistics are a snapshot of a particular population at a particular time (in his case, the twins in his study in England in the past 24 years); and second, that it accounts only for those in the population within the "normal" range.

On the latter, he explains: "Our results are limited to the samples we study. Because parents who abuse their children are unlikely to be included in our studies, we can't assume our results will generalise to families in which children are abused."

But this is robust science - twin studies (not just Plomin's but data from across the world) are among the most replicated in academia and heritability is deemed highly reliable. Plomin cites calculations for everything from reading disability (60 per cent heritable) to eye colour (95 per cent heritable) to personality (40 per cent heritable).

Shall I compare thee to some DNA?

If he'd just left it there in Blueprint, he would have still filled column inches - any talk about genetics is enough to get social media imploding, with keyboard keys being tapped into submission across academic offices, dorm rooms and curtained troll caves in equal measure.

But Plomin goes further. For example, he says that not only do we need to recognise that DNA differences may explain behaviour, but we also need to understand that we now have the power to use DNA as a probabilistic predictive measure, too. And for education in particular, he believes, this could be huge.

"We are now using polygenic scores to predict individual differences - it is identified genetic factors rather than heritability," he explains. A polygenic score is, essentially, a rating of your likelihood to present with a specific disease, trait, behaviour and so on: for example, educational performance or Crohn's disease.

People used to think there was a "gene" responsible for each of these things, but that is not the case. While some things are caused by a single gene (eg, Huntington's disease), these instances are rare. The vast majority of what makes us different from one another is down to thousands of tiny differences in our DNA that each have a small effect but that, together, can have a significant effect.

Spotting those small differences requires huge swathes of data and, in the past decade, that has become accessible via biobanks - libraries that hold the genetic and personal (health, education, environmental) information of thousands of individuals. The UK Biobank holds the data of 500,000 volunteers.

That data is made available to researchers, who run genome-wide association studies to identify DNA differences associated with a trait. They then use the information to predict the likelihood of that trait in others. This is a polygenic score.

"We need hundreds of thousands of cases to spot these small DNA effects," says Plomin. The more data you get, the more small differences you can find associated with a trait, the more accurate the polygenic score becomes.

Plomin's own polygenic score for schizophrenia (he writes that polygenic scores are now the most accurate predictor of schizophrenia that we have) is in the 85th percentile. But he does not have schizophrenia. This, he says, illustrates an absolutely fundamental point about polygenic scores. "It is a probabilistic prediction," he says. "That is really hard for people to understand. They understand the notion that you have the gene or you do not, but this is probabilistic."

All Plomin's polygenic score really means is that he is much more likely to be diagnosed with schizophrenia than someone in the 5th percentile of the distribution, as he has more of the DNA differences associated with schizophrenia. But it does not mean that he will definitely acquire it or that the person on the fifth percentile definitely will not.

Inevitably, polygenic scores have emerged for education. As Dr Kathryn Asbury explained in Tes in September, it is claimed that the EA3 polygenic score can predict individual differences in education better than family income. It explains 11 to 13 per cent of individual differences in how long people stay in education, and 7 to 10 per cent of individual differences in cognitive ability. EA3 is currently made up of more than 1,000 genetic variants: as more variants are found with bigger data sets, the potential is that those percentages will go up.

Are you what you are?

Plomin explains that polygenic scores derived from a study of 1.1 million adults for educational attainment can predict 15 per cent of the variance in GCSE scores in 16-year-olds (for comparison, he says Ofsted inspections predict less than 2 per cent of the variance in children's GCSE scores).

"If all you know about people is their DNA," Plomin writes in Blueprint, "you can, indeed, predict their school achievement."

Some have taken this point and run with it: "DNA is all that matters! 'You cannot cheat your genetic destiny,' says geneticist!"

This is not what Plomin is saying.

"It is probably one of the best predictors we have for educational achievement, but you can take kids in the lowest 10 per cent of the distribution for EA3 and some of them will get GCSE scores as good as some of those at the high end. But if you divide it into deciles, and say 'what is the average GCSE score', it is predictive."

EA3 and what comes after it has the potential to completely change education. We are increasingly relying on socioeconomic measures to intervene ever earlier in a child's life in an effort to create equal access to opportunity, and pupil premium and free school meals have long been established in schools. EA3 may offer a better predictor of not just who needs intervention but what for, too.

Of course, knowing there is a problem and creating an intervention to "fix", or at least alleviate, it is another matter.

Plomin concedes that what we have now is not fit for purpose. "[We need to] expose bullshit like growth mindset - the gimmicks. To think there is some simple cheap little thing that is going to make everybody fine, it is crazy," he says.

"Good interventions are the most expensive and intensive - if it were easy, teachers would have figured it out for themselves."

Polygenic scores could help us to create better interventions, he argues.

"I would like education researchers to use genetics as a variable and see what happens - I predict it will wipe the floor with other measures. With dementia, you cannot do a study of interventions for dementia if you do not know their genetic risk because the chances are, whatever intervention you use, it will work very differently if someone is at high risk or not.

"Research is being transformed in psychology and, eventually, it will happen in education. It will clean up attempts to assess the efficacy of an intervention."

Polygenic scores, alongside increased literacy around heritability, can give us a "peek at the ending of the book", says Plomin, and we need to use that knowledge wisely.

Does that mean using polygenic scores to change environmental factors? After all, our differences are only partly down to genetics, the rest is the impact of the environment. But what we think is important environmentally is actually not, according to Plomin - at least at the current time. Stating this, he says, has "really pissed people off".

"This is the thing that has got a lot of press attention," he explains. "The evidence is that growing up in the same family and going to the same schools does not make kids similar, it makes them as different as if they were reared in different families. If you were adopted at birth and were raised in a different family and went to different schools, you would, essentially, be the same person."

A pause. "To clarify - parents matter and schools matter. Obviously they matter," he says. "But do individual differences in parenting [or schools] make a difference? There is very little evidence that they do."

So whether you go to a school judged "requires improvement" or "outstanding", or whether you are raised in a strict home or home with a more relaxed approach to discipline, his argument is that you are still likely to turn out the same way.

Why might this be? It could be that universal education, centralised curriculum, checks and balances on teaching and schools, societal norms of parenting - all this has led to a high bar of environmental influence that is broadly similar for all. However much we think we are worlds apart in how we teach or parent, and however much we may disparage those who do it differently, it may be that, in Plomin's words, we have got to a point where it is all "good enough".

"If you make equal opportunity happen, what you are left with are the genetic differences," he explains.

If you took a population 50 years ago, the results might look different, as the environment may have been more varied and, thus, genetic differences less influential. As Plomin says, these findings are a snapshot: "They say what is, not what could be." Nor, indeed, what was.

But he says that much of what we consider to be environmental is actually genetically influenced: what he calls "the nature of nurture". So, the child who finds learning easy because their DNA gives them that advantage will curate an environment in which they can thrive, whichever school they are in (obviously if you lock them in a closet, he says, they won't, but in the UK school system, they will seek out the "good teachers", seek out books, choose friends wisely, and so on). Likewise, the child whose DNA pushes them towards risk taking or bullying will curate environments in which those traits win out.

"I know it is a difficult pill to swallow, but there is just so much evidence now to support this argument: that what looks like systematic effects of the environment are actually mediated genetically," he says.

"The most highly heritable trait is vocabulary, and people say, 'how can that be? Are you saying vocabulary is in the genes?' Well, obviously not, but it can mean you tune into the verbal channel. Some kids just love the nuances of words, some are just not interested. The genetics comes out - they select and modify and create environments that fit their genetic propensities."

Genome what I mean?

What are the "pure" environmental factors, then? Chance, accidents, random opportunities, tragedies - a whole range of small interactions we cannot control, says Plomin. "I think they are probably idiosyncratic. The bottom line is, with genetics, we are not talking about major gene effects, we are talking about thousands of tiny effects, and my guess is the same is true of the environment - thousands of tiny things, different ones for individuals."

This could be a bleak message: we are in the chains of our DNA and the only salvation is by chance. But Plomin stresses that this is not the case: he reiterates that DNA only tells you what is, not what could be. You can make a difference, he says, but your success at doing so will depend on how much time you are willing to spend, how much resource you are willing to put into it and how well you recognise the challenge of battling against an individual's genetic preference.

What he urges teachers to do, though, is ask why. Why would you want to force a child on to a path that is 'unnatural'?

"What I try to talk about in the book is tolerance: people are different genetically and we have to recognise and respect that," he says. "Sure, you can help kids to get the same grade but should they have to struggle to do as well as their peers who can just do it with their hands tied behind their back? Maybe they like doing something else better?

"The bad part about the national curriculum is that it is taking all these square pegs and pushing them into the same round hole, or trying to. It might be that we don't judge every person by that academic, conceptual, theoretic learning. But for university educated people, the thought that their darlings might not go on to university is so hard to accept - there is something really wrong with this."

For Plomin, genetics should not influence education resource to ensure pupils fit a cultural concept of success. Rather, it should influence attempts to ensure basic literacy and numeracy for all and then efforts to nurture children in whatever way fits their nature.

Unsurprisingly, Plomin has come under heavy fire for these views. While the basic point that "everything is heritable" is accepted, some reviewers have been highly critical of Plomin's other conclusions.

Writing in The Spectator, Professor Kathryn Paige Harden, who leads the Developmental Behavior Genetics Lab and co-directs the Texas Twin Project at the University of Texas at Austin, argues Plomin's claim that DNA can predict the future "radically oversells the… science" and that twin studies elsewhere had shown family environments do make a difference as "poor children whose 'DNA fortune-tellers' predict that they will succeed in education still end up worse off, economically, than rich children who are genetically predicted to fail".

She also writes that "treating psychological differences between British schoolchildren as if they constitute the entire warp and weft of 'who we are' is a thin and frayed conception of the fabric of human identity". (For more criticism, see bit.ly/PlominCrit)

To the latter point, he replies almost with a shrug. He has never claimed to be speaking about anything other than the population he studies (which is true - he makes it clear in the book and repeatedly in the interview).

"But it does not follow that it is only specific to this population - it remains for people to show that it differs [elsewhere]," he says. "As much as you might think, 'of course it will', so far, the evidence of the heritability of these things suggests it is pretty common around the world. People have looked - it does not seem different."

On the other points, he stresses that what he is discussing is individual difference, not group difference, and that the two are very different. "The causes of individual differences are not necessarily the same as the average differences between groups. So, if you consider gender, or class, or the most explosive example, ethnicity, the causes of the average difference are not necessarily related to the causes of individual difference. So, genetics can be important for individual differences but that does not mean that genetics is causing the average differences between groups - it could be a completely different environmental factor, like discrimination or gender stereotyping."

The criticisms do not seem to worry him. He is just happy that his book has got people talking about genetics, whether they agree with him or not. He calls Blueprint a "provocation" and stresses he does "not have all the answers". His aim is simply to persuade people to recognise the role of genetics in their lives, to understand that we have less control than we thought, and to make us question how we react to that reality.

In schools in particular, he sees the discussion as vital. "I hope we can move education more towards realising it should not be 'genetics is bad, environment is good'," he says. "Teachers cannot ignore that DNA differences are the biggest systematic difference between kids, not just in learning abilities but behaviour problems."

But is there a danger that, armed with genetic information, his message becomes a justification for sorting children into streams, that it limits opportunity, that it becomes a self-fulfilling prophecy: not probabilistic but deterministic? "That would be a disservice to teachers to think they would react that way," claims Plomin.

"They would only do that if they do not understand genetics, if they think genetics is hardwired, that you cannot do anything about it. That is so wrong."

He thinks for a moment.

"Genetics is not a sledgehammer," he says, "it is nudges and whispers."

Jon Severs is commissioning editor at Tes.

Robert Plomin is a keynote speaker at the Bryanston Education Summit 2019 on 5 June, for which Tes is media partner. For more details and to book tickets, go to bit.ly/TesBES2019

Robert Plomin on genetics and education

The phonics screening check: Professor Robert Plomin argues that the results of the phonics screening test taken by all English pupils in Year 1 are highly heritable. He gave the test to those in his long-term twin study (known as TEDS) and found that it was "among the most highly heritable traits ever reported at this age…at about 70 per cent". "This means that the test is not measuring how well children are taught reading. Instead, it is a sensitive measure of genetically driven aptitudes for learning to read," he writes in Blueprint.

Selective schools: "Selective versus non-selective schools: there's a full grade difference between them. Of course that has to be the environment - those are better schools, right?" he says. "But then you control for what they selected for and it is nothing more than a self-fulfilling prophecy. You select the kids that were doing the best and then they do the best. There is no added value in terms of educational achievement. If those kids had gone to a comprehensive school, they would have done just as well. They would have taken advantage of the environment they had, they would have found the good teachers."

Special educational needs and disability: "We have been held back by this medical model that starts with diagnosis," he argues. "It assumes there is a disease or disorder, that you either have or do not have something. We now know it is a dimension, it is not one gene but many genes of small effect. Kids have reading problems: what do we have to gain by calling it dyslexia? It is wrong. It is a dimension, it is more or less of something. There is no dividing line where you have it or do not. There is nothing to diagnose. Maybe it helps to just recognise that kids are different and then decide what you're going to do: roll up your sleeves and realise it is going to be harder. It does not mean you cannot change the child."

Behaviour: "You can change behaviour - parents can and should monitor and control their child's behaviour. They should say 'don't hit your sister', but you have to recognise and respect the differences that are there. Bullying is a good example. Zero tolerance for bullying [in schools] - in that controlled environment you can stamp a lot of it out, but have you changed the bullies? Have you made a difference in who is a bully? When they leave the school environment, I believe they will still be bullies...we have not changed people, we have changed behaviour."

Pedagogy: "The idea of a teacher standing and delivering in front of a classroom? Do they still do that?" he asks. "Jeez, it seems so crazy to me. You are boring half the kids, and some are not understanding what you are saying. And at a daily level, some kids are just more interested in it [education]. They go further, they think about it, they are not just staring out of the window. I know personalised learning is not popular now but, with computers, it is not expensive. People are so opposed to personalised learning. When I mentioned it at a teaching conference, I was shouted down, I think, in part, because they think it will take away from teachers. But it seems to me that it would allow teachers to do what they want to do, and that is to give more personalised help. The kids who are sailing ahead - just get out of their way and let them get on with the programme. But those who need help, mostly for behaviour reasons rather than pure learning reasons, it frees up the teacher to do it."

You’ve reached your limit of free articles this month

Register for free to read more

You can read two more articles on Tes for free this month if you register using the button below.

Alternatively, check if your school has a Tes subscription. If not, for just £5 per month you can subscribe personally for:

  • Unlimited access to all Tes magazine content
  • Exclusive subscriber-only articles 
  • Email newsletters

Already registered? Log in

You’ve reached your limit of free articles this month

Subscribe to read more

Check if your school has a Tes subscription. If not, for just £5 per month you can subscribe personally for:

  • Unlimited access to all Tes magazine content
  • Exclusive subscriber-only articles 
  • Email newsletters

Read more