I have come up with the following code to do intersite genetic correlation
for binary data - it seems to work, but it gives no error variance (fixed
at one in single site analysis). It seems to be syntactically correct, but
is it giving me the right result?
VG binary two site
SITE 2 !A
FAMILY 650 !I
LOC 50 !I
SUBRACE 22
REP 54 !A
INBLK 702 !A
DBH
VG
BIN1
BIN2
SITECOR.DAT !MAXIT 20
VG !bin ~ SITE SITE.REP SITE.SUBRACE !r SITE/INBLK SITE.FAMILY
1 1 1 !NODISPLAY
0
SITE.FAMILY 2
SITE 0 US 0.09 0.1 0.46
FAMILY
Result
...
LogL=-15717.5 S2= 1.000 7400 df 0.06835 0.21239 0.00000
1.00000 0.14355 0.22997 0.56457
Final parameter values 0.06835 0.21239 0.00000
1.00000 0.14355 0.22997 0.56457
Variance heterogenity factor [Deviance/DF] 0.38
Source Model terms Gamma Component
Compnt/StndErr
SI_1.INBLK 702 55 0.683505E-01 0.683505E-01 0.80
P
SI_2.INBLK 702 51 0.212392 0.212392 2.05
P
SITE.FAMILY US=UnStr 1 0.143551 0.143551 1.25
U
SITE.FAMILY US=UnStr 1 0.229972 0.229972 2.20
U
SITE.FAMILY US=UnStr 2 0.564568 0.564568 3.56
U
Covariance/Variance/Correlation Matrix US=UnStructu
0.1436 0.8078
0.2300 0.5646