Attachment | Size |
---|---|
Bivariate_RegularQuantity_75.R [6] | 7.35 KB |
Hi all,
I'm working on a model of smoking quantity given initiation of regular smoking along the lines of Bares, Kendler, and Maes 2015 among others. I'm having a bit of trouble figuring out how to move from a more straightforward bivariate model to the CCC model by implementing the estimated path from initiation to quantity. For reference, this is designated b in Figure 1 at https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4631696/
I have a few questions to help me code this properly. I've attached a script for reference that I'm using as a starting point for the CCC, its a standard bivariate model of binary regular initiation and continuous cigarette quantity, with a covariate of sex on the means. I have a few questions that I think will help me figure out the code, so I would appreciate answers to these questions and/or code alterations that would properly code the CCC model.
First question, I've regressed a covariate onto phenotypic means in fore using definition variables, is this the same process I would use to code the b path? Secondly, I've only used covariates previously where they do not vary within a twin pair (i.e., effect of sex on mean phenotype in twin models of same sex twins only). Twins can vary on smoking initiation, so how would this be coded? Lastly, there should not be cross paths between the ACE components for initiation and quantity, because the effect of initiation on quantity is through only path b, is that correct?
Thanks for your help,
Stephanie