Moderated bivariate ACE

Posted on
No user picture. Hasan Joined: 03/07/2023
Hello,
I'm trying to fit a moderated bivariate ACE model such as the one seen in the attached picture.
Is there a way to do this in umx? As far as I can tell, neither umxGxE nor umxGxEbiv do exactly this. Can they be expanded to fit this particular model?
I've tried to get my head around using OpenMx rather than umx, but with little success :-) But I found this script from the Boulder materials - https://vipbg.vcu.edu/media/course/HGEN619_2015/twin2ModBivAceCon.R, am I correct that this fits the model I'm trying to fit?
Thanks!
Replied on Mon, 04/03/2023 - 10:00
Picture of user. tbates Joined: 07/31/2009

umxGxEbiv should be doing this. If it isn't, please post back here about why.

In practice, the models are often quite hard to estimate.

Replied on Mon, 04/03/2023 - 10:01
Picture of user. tbates Joined: 07/31/2009

umxGxEbiv should be doing this. If it isn't, please post back here about why.

In practice, the models are often quite hard to estimate.

Replied on Tue, 04/04/2023 - 05:39
No user picture. Hasan Joined: 03/07/2023

In reply to by tbates

Thanks! Well, there's two issues:
- umxGxEbiv includes *one* variable plus a moderator, right? Trying to add two selDVs using c() gives the error "selDVs list must be 1 variable...". So I don't see how I can use it to investigate how a moderator impacts the relationship between two other variables.
- The moderator I have in mind is SES, so it doesn't differ between twins. Although I'm early in my research so I could go for a slightly different hypothesis.
Replied on Tue, 04/04/2023 - 05:42
No user picture. Hasan Joined: 03/07/2023

In reply to by Hasan

(Is it better to just run seperate the sample along different levels of SES and run the normal bivariate model for every subgroup? I thought about that but figured that a reviewer would probably ask why I didn't just fit one model that does everything at once :-))