Correlations from ADE and AE, huge differences
JuanJMV
Joined: 07/20/2016
Hi all,
I am writing this post since I am fitting an ADE bivariate model, but I get really different correlations from the ADE and the AE model.
Here you can see the correlations for both models:
> fitADE$algebras$VC
mxAlgebra 'VC'
$formula: cbind(A, D, E, A/V, D/V, E/V)
$result:
A A D D E E SA SA SD
VC 0.060119702 0.086110014 0.028546937 -0.051261209 0.171507340 0.051611766 0.23107500 0.99594547 0.10972249
VC 0.086110014 0.123336193 -0.051261209 0.111217434 0.051611766 0.409801824 0.99594547 0.19141018 -0.59288539
SD SE SE
VC -0.59288539 0.65920251 0.59693991
VC 0.17260261 0.59693991 0.63598721
> mxEval(cov2cor(MZ.expCovMZ), fitADE, T)
[,1] [,2] [,3] [,4]
[1,] 1.000000000 0.211165720 0.340797490 0.085112473
[2,] 0.211165720 1.000000000 0.085112473 0.364012792
[3,] 0.340797490 0.085112473 1.000000000 0.211165720
[4,] 0.085112473 0.364012792 0.211165720 1.000000000
> mxEval(cov2cor(DZ.expCovDZ), fitADE, T)
[,1] [,2] [,3] [,4]
[1,] 1.000000000 0.211165720 0.142968122 0.073855504
[2,] 0.211165720 1.000000000 0.073855504 0.138855743
[3,] 0.142968122 0.073855504 1.000000000 0.211165720
[4,] 0.073855504 0.138855743 0.211165720 1.000000000
> fitAE$algebras$VC
mxAlgebra 'VC'
$formula: cbind(A, D, E, A/V, D/V, E/V)
$result:
A A D D E E SA SA SD SD SE
VC 0.049583546 0.089029836 0 0 0.2104231076 -0.0016801462 0.19070107 1.01923471 0 0 0.809298933
VC 0.089029836 0.159857711 0 0 -0.0016801462 0.4835435673 1.01923471 0.24845725 0 0 -0.019234713
SE
VC -0.019234713
VC 0.751542752
> mxEval(cov2cor(MZ.expCovMZ), fitAE, T)
[,1] [,2] [,3] [,4]
[1,] 1.00000000 0.21356406 0.19070107 0.21767191
[2,] 0.21356406 1.00000000 0.21767191 0.24845725
[3,] 0.19070107 0.21767191 1.00000000 0.21356406
[4,] 0.21767191 0.24845725 0.21356406 1.00000000
> mxEval(cov2cor(DZ.expCovDZ), fitAE, T)
[,1] [,2] [,3] [,4]
[1,] 1.000000000 0.21356406 0.095350533 0.10883595
[2,] 0.213564064 1.00000000 0.108835954 0.12422862
[3,] 0.095350533 0.10883595 1.000000000 0.21356406
[4,] 0.108835954 0.12422862 0.213564064 1.00000000
>
As you can see there are huge differences between the ADE and AE correlations. In the ADE model the cross-twin cross-trait correlation is almost equal for MZ and DZ whereas in the AE model there is a notable difference (MZ=0.22; DZ=0.11) I do not know why I am getting so different correlations. What should I do?
Thank you so much in advance!
With all good wishes,
The differences aren't that
The differences aren't that great and make sense to me at least. The MZ and DZ cross-correlations are rMZ>2rDZ for the ADE model, and rMZ=2rDZ for the AE model. These are consistent with the model. The ADE fits better than the AE (lower -2lnL), but perhaps not statistically significantly so.
Log in or register to post comments
In reply to The differences aren't that by AdminNeale
Thank you so much Michael for
Thank you so much Michael for your response! Greatly appreciated.
With all good wishes.
Log in or register to post comments