Robert Heaton

Software Engineer /
One-track lover / Down a two-way lane

Thousands of elderly twins assure me that my kids will be alright

18 Oct 2023

I know that time spent with my kids is supposed to be its own reward, and it is. But I also want to believe that what I do in this time matters, as much as possible. Elegantly handling a tantrum feels more worthwhile if I’m helping my son learn to express his feelings, not just making it through another day. I find more contentment at the end of a long afternoon if I think that I did a good job and that this good job will echo through the ages, or at least after bedtime.

I want my kids to be happy and fulfilled, skilled and accomplished, and I want to be able to help. Shouldn’t I try to pass my good habits onto them while shielding them from my dark thoughts? This shouldn’t be too hard; I’m their dad, they see me every day. My eldest child, Oscar, is only 4, but I already think he might be a little remarkable, and thick black lines seem to lead back to what Gaby (my wife) and I do with him. The spark and smarts are his, but I feel like I must surely be making a difference.

However, I recently read “Selfish Reasons To Have More Kids” by Bryan Caplan, an economist and blogger, and it’s turned me a little upside-down. Caplan observes that many parents wreck themselves trying to boost and polish their children. He argues that this isn’t just a bad tradeoff, but an almost total waste of time. He presents reams of remarkable research suggesting that, in Western middle-class families, parents’ choices have almost no influence on their children’s long-term health, intelligence, happiness, success, or character. Parents achieve nothing by sending their kids to extra maths lessons, hiding the TV remote, or even teaching them the value of hard work. Caplan shows that upbringing counts for almost nil (at least within the Western middle-class), and that genetics and randomness are everything. It appears that nothing within parental control matters.

Caplan presents his arguments as a gift, one that frees parents from eighteen years of guilt and wasted effort. In his telling there’s little that parents can do to influence their children in the long-run, so there’s no point and no duty for them to try. Kids have genes and free will; now let go and enjoy your time together.

Caplan knows that some parents will rebel against his arguments. I certainly did. I heard him telling me that I don’t matter, at least not in the ways that I’d hoped. I want parenting to be a deep, complex vocation, and I want to spend the coming decades playing a domestic game of skill and consequence. The idea of having children who I have no influence over is scary, like living with werewolves. Randomness and outside forces are everywhere and the kids are mutating while I sleep.

But even though I want to be relevant, I don’t want to waste my time. Begrudgingly, I kept reading.

What’s the evidence?

Caplan’s claim that parents have little long-term influence on their children seems absurd at first. Contra Caplan, I see my influence in my children every day. Oscar likes the same music as me. He used to be terrified of playgrounds but Gaby screwed a wooden ladder to his bedroom wall and now he’s mostly normal. I stubbed my toe and shouted “fuck!” and he whispered “fuck indeed daddy, you sound frustrated,” failing to calm me in the same way that I fail to calm him. This is surely common sense.

But common sense grows in unscientific environments. Nature and nurture are conflated, we don’t see the aggregates, and we don’t see the long-term. Kaplan agrees that parents have huge influence over their children in the short-term, but he also argues that this influence fades, sometimes fast, sometimes slow, but it does fade, and it vanishes completely when they grow up and finish becoming whoever they are. Kids are resilient to setbacks, but they’re resilient to assistance too.

In order to rigorously test theories like this, researchers study large groups of children. However, most kids are useless to them. Suppose that two happy parents have and raise a child. The child grows up with their parents, and in time they become a happy adult too. It’s impossible to know whether the child’s happiness comes from happy genes that they inherited from their happy parents, or from the happy environment that their happy parents raised them in. Their parents’ genes and choices are irreversibly mixed together. Even with a huge database of children, parents, and measurements of happiness, causalities are impossible to itemise.

Fortunately, researchers can still extract good data from special children, like identical twins who were separated at birth. These kids give researchers two copies of the same genes, raised in different environments. Since separated identical twins share genes but not environments, any systematic differences between them must be due to their different upbringings. If identical twins raised separately bear no resemblance to each other but are similar to their adopted siblings, this would suggest that the twins were shaped by their divergent upbringings. If the twins remain similar, despite growing up entirely separately, this would suggest that they were made by their identical genes.

Researchers slice and measure these children, pulling apart the effects of nature and nurture. Twins separated at birth are the gold standard, but non-twin adoptees and non-adopted twins can work too. The researchers find or build databases of useful children (who may now be adults), and compare their grades (perhaps from school records), income (perhaps from tax records) or personalities (perhaps from administering personality tests directly). The evidence from this data is strong and consistent: a near-zero effect of upbringing on character, happiness, and almost everything else.

Should I pay attention to the evidence?

The studies are clever, but are they valid? They control naturally for almost everything, but they still aren’t perfect. For example, maybe parents who choose to adopt are meaningfully different to the average parent, meaning that conclusions based solely on them don’t generalise to the rest of the population. Maybe parents who choose to adopt and then also agree to be part of a long-term study are even more different. Maybe women who have twins are different. Maybe twins themselves are different too.

But even if these sampling biases are material, I doubt that they’re large enough to tear down the studies’ broad conclusions. I’d guess that adoptees and twins separated at birth are a good enough sample to represent humanity, and that even if they aren’t fully representative, they probably aren’t masking a giant effect that skips twins and applies only to the rest of us. If researchers were able to fully control for sampling biases then this might shift their estimate of the effect of parental influence from “incredibly low” to merely “very, very low”.

Caplan admits that the studies are primarily focussed on the Western middle class, because that’s where the data is. This hurts the studies’ generalisability but binds me - an orthodox member of their class - even tighter. All said, I think I have to assume that the studies pointing towards the primacy of genes are valid for people like me.

But do the studies definitely apply to me, or you, specifically? They find no effect of parenting style on children’s adult outcomes, within the range of normal parenting styles in middle-class Western families. That word “normal” might provide an opening for a determined parent to squeeze through in order to regain their lost gravity. The studies suggest that there’s no difference between the free-range and regimented ends of the normal spectrum, but they can’t say anything definite about what happens beyond the edges of normality.

Caplan recommends that parents dissolve their fears and ambitions in the acid balm of the evidence. But there is another response that’s consistent with the data, although it might not necessarily be a good idea: redouble your efforts and head for the ambiguity beyond the well-studied centre, where the evidence might not stretch. More enrichment, more practice, more effort, fewer half-measures.

This makes some common sense; think about outlandish famous families. The Williams sisters must be naturally gifted tennis players, but they surely wouldn’t be the same dominant champions without their obsessive dad. The Polgar sisters would have been unremarkable chess players without theirs. These types of childhoods are so rare that they can’t possibly be adequately represented in any of the twin datasets, so the research doesn’t have anything direct to say about them. Twin studies don’t disprove the Jackson Five.

Caplan claims that kids are elastic, and that whether helped or harmed they tend to snap back to their natural state. However, I learned in physics classes that not even elastic is perfectly elastic. It pings back to its original shape after mild deformation, but it can still be altered permanently if stretched beyond a point called its elastic limit. Whilst the sum of small interventions on a child might be zero, it might still be possible to permanently deform them (in a good way) through the application of massive force. This metaphor is so perfect that it must surely be true.

In fact, even Caplan is stretching his own kids like this. In a 2015 blog post (the book was written in 2011), he describes the rigorous homeschool that he runs for his two eldest (twins, coincidentally). The main reason he homeschools them, he says, is because they are particularly academic kids, and they all think that they will enjoy an uncompromising homeschool more than a conventional one. However, he also suspects that his homeschool might be so off-the-scale remarkable that it vaults over the evidence and produces better adult outcomes, despite his claims that this is usually impossible. He writes:

I suspect – though I’m far from sure – that the Caplan Family School is such an exceptional experience that ordinary twin and adoption evidence isn’t relevant.  For example, my sons are plausibly the only 12-year-olds in the nation taking a college class in labor economics.

Should you or I try to do this too? It’s almost always delusional to put yourself and your children in a category called “exceptional”, and this might not even be a category that you want to be in. I do wonder, though, where does “normal” end and “exceptional” begin? Where’s the elastic limit, and how weird is it really? Is anything less than the Williams sisters a waste of time? Or does the curve bend much sooner than that? Even if you don’t want to do anything too odd by modern standards, a lot of the data in these studies comes from dead twins brought up decades ago. Today’s parenting zeitgeist might not necessarily be better than the old days, but it’s certainly different. How well does data from a different era in parenting generalise to today? Is it possible that even normal parenting today is different enough from several decades ago to have a material impact?

Is reading a respected parenting manual and teaching your toddler to add and multiply too normal and futile, or just crazy enough that it might work? I don’t want to be Richard Williams and I couldn’t even be Bryan Caplan, but I could be a bit weirder than average if that was worthwhile and harmless. I might be inventing straws to clutch at, but as far as I know there’s no cast iron science out here so we’re allowed to make things up again and I can assert a world in which I have agency.

What should I do now?

I’ve drilled a tenuous airhole in Caplan’s claims, but his evidence is still strong, spiky, and hard to digest without a rupture in my plans. Normally when confronted with new evidence you can wisely say “it’s probably a combination of everything” and then maybe do a bit more or less of something, or not. However, Caplan argues specifically that parenting is not a combination of everything. Everything is nature, at least in the long-term. His arguments are backed by simple and compelling studies that are hard to wishy-wash away and that block the easy path back to the status quo.

But it’s drastic to change how you raise your kids based on a short book and some studies that you aren’t going to read. The book’s claims are extreme, at least compared to what I used to think, and it’s hard to build enough confidence to change your mind about things that matter to you. I rarely need to develop solid beliefs about messy, unsettled topics that I’m not an expert in. I’ve skimmed a few paper abstracts and some reviews of the book, but that doesn’t feel like enough. Caplan seems smart and honest but this isn’t settled science and how do I know he’s not missing or ignoring grave methodological gaffes?

I can’t unread the book, and as someone who likes to consider themselves a somewhat scientific, data-driven parent, I can’t ignore it. So what should I do now?

I think I value my children for who they are already, but it’s good to be reminded to start there. I don’t care whether I have any long-term influence on my friends, I just like spending time with them and being there if they need me. So why do I care about being able to shape my kids? The desire to help your children is surely natural and normal, to a degree, but that doesn’t mean it’s always helpful.

Caplan says that I can stop worrying about whether I’m wrecking Oscar’s future habits and character. I try not to fret like this, but often it’s unavoidable. Does he play by himself enough? Does he watch too much TV? Are we letting him be too picky with his food? Should we use more discipline when he won’t share? Less discipline? I extrapolate today’s small behavioural decisions ten years forward into a bleak future. I fear that parenting is a system of positive feedback loops, where deviations become liberties that congeal into nightmares. But Caplan says that everything is fluid and reverts to the mean and I shouldn’t sweat the deviations. Bribe kids to behave, give them unlimited social media time, none of it matters, they’re much less of a blank slate than you think. Nothing will come back to bite you, and if you do get bitten then there was nothing you could have done to stop it.

Still, I’m not ready to stop trying to help my kids flourish. I’m not confident enough that Caplan’s evidence applies to my family and my era, and in any case at the moment I don’t have to make any tradeoffs. Oscar and I do a lot that I’d previously assumed would benefit him in the longterm: maths, reading, piano. For now he enjoys nearly all of it and so do I, so nothing is being sacrificed. I’m sure that this will change as he gets older, but at the moment it’s more fun for me to talk to Oscar about multiplication and prime numbers than pretend to order another pasta with cheese from his play-dough restaurant. On some days he doesn’t want to do any sums and tells me to get lost. But even if I was certain that the long-term impact on his future earnings would be zero, I’d still take him to the science museum and try to remember how aeroplanes work.

This sounds relaxed and balanced, but it’s easy to be sanguine when there aren’t any dilemmas. If he stops being interested in things that I think are valuable then I’m sure I’ll feel anxious, and I’ll struggle when he starts making decisions that I think are mistakes. I’ll reevaluate when I’m forced to, but for now I hope that it’s possible to both try to help your kids excel and to live with them in the moment.

I wonder if these studies should change how I see the rest of the world too. I’m friends with my old physics teacher from high school. I went to his house for lunch and told him about Caplan’s book. He was horrified. “If that’s true, is there any point in me trying to be a good teacher?” he said. This had occurred to me too. If parents truly have no lasting influence on their children, how can schools, or local theatres, or any kind of small public policy intervention hope to have any? Maybe it’s even harder than I thought to make any long-term difference to anything.

And how should I think about traits that have value but don’t show up in survey data? For example, I can take Oscar to piano lessons and encourage him to practice. Most adults who know how to play the piano probably had lessons when they were younger, and their parents probably pushed them at least a little. Does being able to play the piano matter, morally and cosmically, even if it has no impact on income, happiness, or anything else that can easily be measured? The harder you think and the more precise the questions, the more you need a detailed moral philosophy.

It’s helpful to have thousands of elderly twins reminding me that my kids will probably be fine, whatever I do. Everything reverts to the mean, the twins murmur kindly. Don’t be too smug when things are going the way you hoped, and don’t despair when they aren’t.

I’m not ready to fully accept my obsolescence yet. We’ll watch more TV but we’ll keep doing maths together. One day we’ll start to disagree, and then we’ll reassess. Caplan does throw me one bone: “parents [have] moderate influence over how much their children like them.” Even if nothing I do adds up to anything, the days will hopefully make a happy childhood.

Read more of my essays about parenthood here. Plus, I’m writing a book about having kids! Subscribe to my newsletter for updates.

Subscribe to my new work on programming, security, and a few other topics. Published a few times a month.
Follow me on Twitter ➜ RSS ➜