Is more nature bad news?

More on nature versus nurture. This post suggested there is compelling evidence that shared environment (family, neighbourhood and the rest) plays a smaller role for social inequality than people might expect. At least when it comes to educational attainment and income. Now, I’m sure there are some valid counterarguments and conflicting data points, and I’m open to alternative explanations compatible with the data.

I am convinced, however, that many people have, what we may call, an anti-nature bias—they tend to prefer nurture to be the more critical component. On the surface level, that is understandable. The idea that life outcomes are contingent on our genetic setup (over which we have no control) may seem rather depressing. There is also the concern that approaching genetic explanations may lead us down a dangerous path, possibly ending up with eugenics, “race science”, or something equally wrong.

These are, of course, separate questions. If we are genuinely curious to find out how things work, our preferences for how they ought to work don’t necessarily help.

I also suspect some of the aversion to hereditary explanations may be based on misunderstandings about how to interpret the results. I’ll get to that in a bit but first I think it’s helpful to look at a less contentious example. Let’s consider football.

A fictitious study

Suppose we would like to find how much of the observed difference in “football ability” is explained by the factors:

1. Shared environment: family, neighbourhood, school etc.

2. Unique environment: things that only impact one individual.

3. Heritability: genes.

So we need a method for assessing “football ability”. For instance, we could observe people perform some drills and assign them scores. Sure, there will be errors involved in the scoring, and the scores won’t perfectly align with people’s actual abilities and so on . These are the kinds of quirks social scientists need to grapple with, and there are methods for controlling for such errors.

Say we now perform a twin study. That is, we look at a set of identical and fraternal twins. As explained here, the idea is that identical twins share 100% of their DNA and fraternal twins, on average, 50%. As we also learned, we need to be careful with that assumption. In case there is assortative mating for relevant football genes, that number is going to be higher than 50% for the genes we care about:

Assortative mating occurs when couples tend to form pairs based on (genetic) similarity of a trait. If they do, then non-identical twins (and siblings) are expected to be more than 50 % genetically similar, for that trait. In twin-studies, this will reduce the difference in correlations between twin types and thus make shared environment look more important than it actually is.

Assortative mating aside, what if the correlation between identical twins is, say, 0.8? That alone means that if you’re a lousy player, your identical twin brother/sister is likely not very good either. It suggests that the impact of unique environment—unique experiences that only apply to one twin but not the other—is small. To be perfectly clear, it doesn’t rule out unique environments with substantial impacts. If one twin is hit by a car, that is a unique environmental experience that might have detrimental effects on her football playing abilities. So when we say that the impact is small, I take it to mean that it is rare to have experiences with significant influence.

Importantly, finding a strong correlation between identical twins does not shed any light on the nature versus nurture question. How much of the differences in ability are explained by shared environment, and how much is explained by genes? If shared environment has a significant effect, then it is, of course, expected for twins to have similar outcomes.

How can we gain insight into the underlying breakdown or genes vs environment? Since non-identical twins share only half their genes (assuming no assortative mating) , comparing the two correlations helps. The intuition is the following: if the gap is small—say the correlation for identical twins is 0.8, and the correlation between non-identical twins is 0.7— it’s a strong indication that the shared environment plays a prominent role. To put it differently: if you dial down the genetic similarity and don’t observe much of a change in the correlation, genetic similarity is probably less critical. Conversely, if there is a big gap between the two correlations (say 0.5 vs 0.8), we conclude that genetic similarity is a better predictor; more of the observed differences in the ability to play football can be explained by heritability. In theory, it implies that if you had DNA sequenced a toddler and also knew what genes code for football ability, you could accurately forecast their future football skills.

Interpretations

Suppose the study finds that genetics indeed play a major role. With some further analysis, it also derives an estimate for unique environmental impact. It concludes that shared environment, unique environment, and heritability explain 10%, 20% and 70% of the observed variance, respectively (this is all made up). What’s the conclusion?

On the one hand, it may seem perfectly intuitive. It’s probably well accepted that some are better equipped to play football. They have a superior physique, reflexes, endurance, and other essential features, which we also know (or at least can suspect) are strongly heritable. On the other hand, it’s apparent that a persons football playing skills are highly determined by how much they practised playing. So isn’t it silly to propose that it’s primarily down to nature?

One thing to point out is that the correlation is the average impact. It is fully compatible with a reality in which shared or unique environments better explain many players' abilities.

A second point: Playing football is not hardcoded in our DNA. Genetics does not only impact our physical ability to play well but likely also our propensity for enjoying it, spending time practising, being sufficiently competitive and any other factor that might ultimately impact performance. The fact that some are naturally inclined to play a lot of football, which also drastically improves their game, can itself partly be driven by psychological traits that are highly heritable.

Thirdly, if you were unlucky in the genetic football lottery, the conclusion is not that it’s futile to try to become a good player. After all, very few people dedicate a lot of time to improving their football skills, and if you have an interest and are willing to put in the hours, you may still end up at the far end of the distribution anyway.

Why nature is not that bad

Here are a few points focusing on the upside of “more nature”. That is, if we learn that some traits we care about are, indeed, driven by genetics, it’s not all that bad.

Genetics is not destiny. When we say that genetics is paramount, it’s not suggesting that only core physical or cognitive abilities matter. At least in theory, it’s compatible with the explanation that interests and preferences are the primary factors driving different outcomes. Consider educational achievement. I think it’s apparent that some are more interested in the stuff that goes on in schools and universities and others less so. Such preferential differences could be explained by genetics, but it does not contradict the intuition that “if you study hard, you’ll do better”. All the kinds of psychological traits that positively correlate with educational achievement may be heritable but still not deterministic in the sense that they are impossible to influence. It could be that most people naturally fall back to behaviours that align with their genetic predispositions and that few types of shared environments drastically force people to deviate from these. As mentioned, such environments could, of course, still exist. This whole discussion resembles the one about free will. Even if it is ultimately an illusion (I’m agnostic), it does not change our experience of being in charge of our decisions.

Meritocracy. I don’t want to suggest that genes don’t pose any constraints. For instance, I believe any attempt for me to become a professional shot-putter or a fields medalist would have been futile. It may seem unfair and disappointing but is the alternative “blank slate hypothesis” any more attractive? If everyone had the same opportunity to succeed at every endeavour, so long as they were given a decent upbringing, a failure to achieve anything but outstanding results must come down to laziness. There’s the debate about whether meritocracy is flawed, and where critics argue that a meritocratic system not only creates losers, but also permeates the notion that “losers have no one to blame but themselves”. More awareness of genetic factors might help to counterbalance that idea.

Parenting. Imagine if shared environment is crucial and your kids don’t do very well. What does it tell about your parenting ability? If genetics is more important, parents need to stress less about curating the perfect upbringing and perhaps feel a little less guilty should things not work according to plan.

Comparative advantage. David Ricardo’s theory of comparative advantage explains why trade is beneficial even when one party possess an absolute advantage, that is, they are more effective at producing all goods. The insight can be applied to international trade as well as on a personal level. For every endeavour you can imagine, there are most likely others who could perform it better. But that is no reason to dispair. Most of those people are better off pursuing something else. The game is to find a niche in which your relative advantage is stronger. In a way this is a more reality grounded and less cliche take on the saying “everyone is talented” or “everyone has some unique skills”.

How to increase social mobility? Say we learn that genetics is overrated and shared environment is, in fact, the crucial component for attaining social mobility. How do we now improve it? Easy, all we need to do to harness all the unrealized potential out there is to create healthy, well functioning and harmonious families, idyllic neighbourhoods and an excellent K-12 education! How do we do that again?

On a more sobering note, this is incredibly difficult and, in my view, nothing to be particularly cheerful about. If, instead, we established that people can succeed even in less than ideal environments, there is still the challenge of how to spot talent and provide people with enough opportunities. But it’s a less daunting task than addressing all social issues.