The nature-nurture debate: what are the broader implications?

In his latest book titled Selfish Reasons to Have More Kids, economist Bryan Caplan reviews the current evidence on the role of nature and nurture for what he considers the parental wish list, and shows that for most items on the list: (i) genetics explains a significant proportion of the variance (measured by heritability h^2), (ii) of the non-genetic proportion of variance, parenting, i.e., the shared environment between siblings, explains only a small part (measured by common environmentality c^2) (iii) the remaining variance 1 – h^2 – c^2 is explained by unique environment, which basically means it is explained by an amalgam of random factors and individual choices that cannot be systematically predicted. This is called the unique non-shared environment.

If you want a quick summary but don’t want to buy the book, check out Caplan’s blog post on the parental wish list.

The usual caveats apply. Since they aren’t mentioned in the specific blog post I linked to above, I will list them here briefly:

• Parenting has a huge short-run impact but the impact fades out as children row up and move away from the parents. The heritability of many traits increases with age as people shed off the influences of their parents and discover more their own inner genetic preferences and tendencies.

• The studies that these estimates are based on compare families in developed countries that are considered fit by adoptive agencies to raise kids. The estimates may not generalize over the entire range of possible families. Heritability may be lower across a larger selection of possible parents.

• The study compares parenting techniques currently in use by significant numbers of people. Truly weird or outlandish parenting techniques may produce big influences.

• The studies compare variance, not absolute quantities. The question is to what extent variation in a trait (such as reading skill) can be ascribed to variation in parenting strategies, not whether a child will learn to read if the parents refuse to expose her to books or send her to school. The studies do not compare people across different generations so changes in the overall environment can have an impact (a secular trend of increasing height has been observed by many, and a secular trend of increasing IQ has also been observed).

• The studies don’t support genetic determinism in its most direct form — rather, they indicate a huge proportion of variance not explained by genes or shared environment. This is basically the variation that arises from all the individual choices and experiences that are not determined by parenting strategies. They may include biological history (did the child get a bout of sickness), individual influences and inspirations, free choices, etc. No single influence has been identified within this that has a large share to contribute.

Unlike Caplan, I will not focus here so much on the implications of the nature-nurture debate for parenting. Rather, I will make one obvious key observation: if parenting has such little impact on individuals, most other efforts to change or modify individuals are likely to have completely negligible impacts.

Influences work best on people when they are already keen on those influences and seeking those influences. What matters is the range of options and influences available to people, and the freedom to choose the influences that best suit them. This helps explain why significant differences in people’s achievement, income, happiness, and other measures based on the country and time period that they live in — despite the fact that in the same country and time period, people with similar genetic endowments do similarly. An argument of this sort is provided in this famous paper by Dickens and Flynn which tries to resolve the paradox of high heritability of IQ within a time period combined with a secular trend of increasing IQ across decades.

This has important implications for people who want to do good in the world. There is a common saying of the form: If I can help one person, my efforts have been worth it. Helping even one person is an achievement of note. But given what the literature on fade out suggests, fade out is pretty strong for practically all kinds of help — not just parenting (see here for fade-out of teacher effects on students).

My take-away from this is simple — helping, or changing, individuals, is generally speaking too much work for too little gain. So what’s a better way to help people? The key in my view is leverage:

• Make and do stuff that lots of people — not just specific people you are trying to help — want, or can benefit from.

• Make this stuff readily accessible to lots of people.

• Don’t nag or bully people into using the stuff. Let people choose to use the stuff when it best suits them. When people choose to use stuff of their own accord, it’s likely to be most relevant to their needs. If you force it down their throats, they may gulp or swallow it, but fade out will follow soon.

Thinking in terms of leverage means foregoing the intense experience of helping specific individuals* and instead going for the more hands-off experience of doing stuff that helps people unknown to you, often each person being helped in such a small, insignificant way that they don’t devote much thought to it.

*I don’t mean to suggest that spending time with other people on mutually enjoyable activities is bad — that’s obviously a good thing in general. What I am talking about is spending time trying to “change” another person in ways that are obviously unpleasant or fatiguing for both parties involved, and where the other party doesn’t appear to be an eager participant to the change process.

Advertisement