The measurement error is assumed to be independent between the twins, which puts it in the “non-shared environment” (E) bucket. So the E-correlation is biased towards zero while the genetic (A) correlation is fine.
The measurement error is assumed to be independent between the twins, which puts it in the “non-shared environment” (E) bucket. So the E-correlation is biased towards zero while the genetic (A) correlation is fine.