Comment by tomrod

It's not vanishingly tiny, simply small.

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