I am wondering how should we understand “environmental components” in twin modeling ? It is basically referring to anything that is not genetic? And is it referring to a specific period? Even if twins grow up in the same household, after a while it is likely that they will not share the same environment anymore.
That lead to my other question:
Are the proportions of genetic (A and D) and environmental (C and E) components that affect the variance of a trait fixed? Or can they change over time? I’m guessing it depends on the trait…
The environmental components are indeed non-genetic components. Similarly to heritability, the estimates are population estimates, which are specific to that population in space and time.
The distinction between common (C) and unique (E) latent factors refers to whether these factors contribute to similarity between twins/siblings (or family members), or not.
To answer the last part of your question, yes, variance components do change over time. We commonly see shared environment variance for traits during youth, but it dissipates once the twins move away from home, go to college etc. This is consistent with more self-determination of behavior once one leaves home, but other developmental factors may be involved. That C doesn’t seem to “stick” over time is plausible, albeit not consistent with the concepts of critical or sensitive periods that have lifelong consequences. There may be advantages to the organism in being able to overcome early adverse experiences.
Also worth noting is that unique environmental variance includes the effects of measurement error. Since error is not transmissible, unequal autoregression of A C and E components in, e.g., a developmental model, is to be expected for traits with appreciable measurement error.