Interpreting Adoption Studies

This is supplementary.

Understanding some key facts about twin and adoption studies helps make their results seem less counter intuitive.

The data discussed here is primarily on children and parents in first world countries who are non-poor. This data does not help answer questions about parenting effects that are very different from typical first world non-poor parenting styles. The data does not help address the effect of growing up in malnourished or without access to education. Indeed, twin and adoption studies with adopted kids in extremely poor households show that nutrition is an important predictor of life outcomes (link)

It also doesn't address extreme parenting styles. Not many people raise their kids in the woods cut off from the rest of society and this kind of variable is not included in the regressions, so the data has little to say about this kind of parenting.

If adopting parents treat their adopted children with “less intense” parenting than their biological children, then adoption studies will understimate the effect of parenting. In the extreme case, if all adopting parents treat their adopted children the same as other adopting parents but vary in how they treat their biological children, we will measure a zero effect size even if parenting has important effects.

Parenting could have a big effect on combinations of outcomes while only having small measured effects on each individual outcome. For example, parenting could have a large effect on “having a major drug or alcohol or gambling problem” but the measured effect of parenting on each of these individually could still be small because the adoptees and their adopted sibblings can have different symptoms (one has a drug problem and another has a gambling problem).

It's also the case that adopting parents are probably systematically different from non-adopting parents. It could be that non-adopting parents tend to have parenting styles that do have important effects on long term child outcomes, while adopting parents have parenting styles that have very small effects.

When an adoption study finds that parenting does not affect outcome X, it does not mean that parenting cannot affect it, just that the parenting styles in the data set did not affect it.