Hacker News Reader

Comment by dcolkitt | Hacker News Reader

One thing that we know from decades of twin studies is that childhood family environment explains a vanishingly small percent of long-run adult outcomes.

Contextualizing this single study's result against the much larger, more established evidence from twin studies strongly suggests that one or more of the following is true.

- The results are spurious and won't replicate.

- The results are true, but the association is not causal, and is simply proxying correlation with an upstream factor.

- Lack of touch does lead to methylation in the short-term, but by adulthood there's strong reversion to the mean.

- The methylation has no significant impact on any actual metric that we care about like success, health, personality or wellbeing.

View on HN

There are issues with twin studies, not the least of which the number of twins raised separately, but not adopted, is tiny. However adoptive parents are a small self-selected subset of the population.

Most twin studies simply compare identical and fraternal twins in non-adopted households.

Since identical twins share twice the genetic covariance of fraternal twins, the impact of family environment can be backed out from the respective intra-pair correlation coefficients. In the limit case if genetics played no factor, then fraternal and identical twins should have identical pairwise correlations.

For example suppose pairs of identical twins have 45% pairwise correlation for adult IQ. And say fraternal twins have 25% correlation for the same measure. That would tell us that the population level variance of adult IQ is 40% attributable to genetic heritability, 5% to environmental heritability , and 55% to non-heritable factors (i.e. not genetics and not family environment).

In case you're interested, IQ is much more heritable than that for adults. The following quote is from Wikipedia:

"Twin studies of adult individuals have found a heritability of IQ between 57% and 73%[6] with the most recent studies showing heritability for IQ as high as 80%[7] and 86%.[8]. IQ goes from being weakly correlated with genetics, for children, to being strongly correlated with genetics for late teens and adults. "

Right, those studies can tell us about heritability, but cannot distinguish between common environment and genetics.

Why would parents of twins pay much more attention to one than the other?

Maybe because, for whatever reason, one cries a lot more than the other?

But that would be an edge case. I would have assumed that identical twins behave similarly. How do they find enough neglected twins to make a reliable conclusion.

I’ve known three sets of identical twins. In each case the personalities were very different. Outgoing vs reserved, aggressive vs laid back, etc. I could easily see one wanting to be held more than the other.

Anecdotal, but... My (fraternal) twins have very distinct and different personalities. One requires an order of magnitude more attention than the other. We try to make up the difference, but it's hard.

It's not vanishingly tiny, simply small.

Can you think of a reason selection bias could be an issue, thus invalidating twin study cohorts?

Adoptive parents are people who have gone through a lot of effort and expense to get their children, so they're far more likely to be heavily invested in their kids. It's also expensive so they're likely to have money. If the effects of parenting are strong, but "cap out" -- i.e. above a certain level of care, parenting does little, but before that level it has a strong effect -- then twin studies will show little of that effect.

I'm not sure I follow the causality argument here.

A twin study should normalize everything genetically. Ergo epigenetic expression and environment are the only two causal levers. It sounds like you're arguing that parental involvement has a causal impact, which is the point of the studies to a degree.

Perhaps my original question was unclear: is there something the twins do or have done to them that cause a latent selection bias before adoption, thus invalidating the twin studies approach as a whole?

We are questioning how much the parenting has an affect on the child. The parents are self-selected and so are more likely to be similar to each other.

Agreed. That was my point. Not all twins under adoption are adopted, meaning there is nothing particularly invalidating about the twin study framework.

I don’t know. I have read this before. But this goes so much against my gut feeling and what I have experienced that it is hard to believe. It’s counter intuitive to say the least.

The documentary Three Identical Strangers also agrees with you, a fascinating account of triplets separated at birth to study parenting.

https://en.wikipedia.org/wiki/Three_Identical_Strangers

Reply on news.ycombinator.com