I was wrong about twin studies

In the past few weeks I’ve enjoyed digging into Kathryn Paige Harden’s The Genetic Lottery and, thanks to Kathryn’s eloquent and clear writing, realized that I misunderstood how ‘twin studies’ have been used to understand the role of genetics in child development. Here’s what I wrote last month:

One common way to study the impact of genetics on humans is through “twin studies” which compare identical twins (with the exact same DNA sequence) that were separated at birth (and thus have been raised in different “environments”) to fraternal twins (who don’t share a DNA sequence) to estimate the role of genetics in development. Some studies have shown that certain life events are surprisingly correlated between identical twins.

In hindsight, I should have known of course that twins separated at birth are exceedingly rare and so the vast majority of twin studies are actually conducted on twins raised together in their birth home. These studies ask the simple question: how much more similar are identical twins than fraternal twins? Remember that fraternal twins are only 50% genetically similar compared to identical twins which are 100% genetically identical.1

Understanding how these different kinds of twin studies are conducted is important because each method of study comes with its own unique problems. Studying twins separated at birth introduces a host of complications when trying to compare them. Imagine the number of differences in your upbringing between you and your childhood friend. There are thousands and thousands of differences that led to both of you being different, how could you uncover the role of DNA? It would be difficult and those that have tried have brought on much criticism.

Similarly, studying twins raised together asks the question: given equivalent environments, how do the similarities between fraternal and identical twins compare? The “equivalent environments” assumption may not be valid given documented differences in twins’ prenatal environments and parenting. Are identical twins parented in the same way as fraternal twins? I’m suspicious of that assumption.

This isn’t to say that we should dismiss 50 years of twin research. The benefit of doing a twin study which simply compares twins between households is that there are a lot more twins to study, and the more you study, the more confident you can be in finding a real effect among the noise. A recent Nature paper conducted what’s called a meta-analysis which is when