You are here

Scripts for Power analysis of Bivariate ACE in Twin Sample?

2 posts / 0 new
Last post
joz's picture
joz
Offline
Joined: 06/30/2015 - 16:30
Scripts for Power analysis of Bivariate ACE in Twin Sample?

Hello,
This is my first time using this forum...any help is greatly appreciated!

Are there any OpenMx scripts for determining power in a twin sample using bivariate ACE analyses?

In a twin sample, I'm conducting bivariate ACE analysis of a personality trait and depressive symptoms. I found 4.6% in shared additive genetic variance, which is smaller than expected (compared to previous reports in the literature)

My univariate ACE results: A^2 for personality trait = 62%; A^2 for depressive symptoms = 32%

My sample size is on the small side (MZ=112, DZ=100; total N=212), so I'm trying to determine if the relatively small effect size is due to power issues. I'd like to assess power of my sample.

Are there any OpenMx R scripts for assessing power in a twin sample using bivariate ACE ?

OpenMx version: 2.0.1.4157
R version: R version 3.1.3 (2015-03-09)
Platform: x86_64-apple-darwin10.8.0
Default optimiser: NPSOL

AdminHunter's picture
Offline
Joined: 03/01/2013 - 11:03
Power or Confidence Interval?

/* I had been emailing with Joyce off the forums prior to this post, and this was my email response. */

Hi Joyce,

I'm not sure what you mean by "power" in this context, but if you want a sense of the range of possible values of the variance components then you can get confidence intervals on them.

If you want a confidence interval on all the elements in a matrix/algebra called "A", then do something like this

ciA <- mxCI("A")
yourModel <- mxModel(other stuff, ..., ciA)
yourFit <- mxRun(yourModel, intervals=TRUE)
summary(yourFit)

In the multigroup case, you can put the ciA object in the model where "A" lives.