Hierarchical Multivariate heritability analysis

2 replies [Last post]
koak's picture
Offline
Joined: 05/10/2012

Hello

I am wondering if anyone knows of a script that would allow me to test the next step of a common pathways model. The Common Pathways model came out as the best model for my multivariate analysis, but I am now wanting to test whether another variable that I also have (but that is not currently in the model) is actually what is currently the common latent variable. Does anyone have any ideas on how I could do this?

Also, I may then want to see if there is an additional common latent variable as well as the variable I am able to define - so ideas on that second step would also be greatly appreciated.

Many thanks

Karen

neale's picture
Offline
Joined: 07/31/2009
Unlikely to be the latent variable

Hi Karen

I would simply suggest that you include the new variable in a larger common pathway model than the one you currently have. At most I would expect it to be closely related to the latent variable (if so, it should have a big factor loading on the path from the latent factor to it). Some residual variation (maybe only E) is practically inevitable unless you have a variable which has been measured with no error whatsoever.

Cheers
Mike

koak's picture
Offline
Joined: 05/10/2012
Thanks heaps for this

Thanks heaps for this suggestion, Mike.

Karen