Hi all,

I’m trying to run a saturated model with restrictions in a multivarite analysis with 6 continuous variables.

In this saturated model with restrictions (see below) I had the following problem (that I did not have in the saturated model without restrictions):

Error: The job for model 'Con' exited abnormally with the error message: Expected covariance matrix is not positive-definite in data row 431 at major iteration 0 (minor iteration 1).

However, when I use the same script but instead of 6 with 5 variables or less, the script runs perfectly. I don’t know what it’s happing… I think it’s something related to the starting values, but I don’t know how I can solve this.

Any help would be very helpful.

Many thanks in advance.

#######################################################################

# Restrictions: means and variances equated across birth-order & zygosity groups;

# One set of Cross-trait cor, symmetric Cross-trait Cross-twin correlations (MZ and DZ)

# Create start values

# To avoid starting the optimization at the solution add some random noise

jiggle
Stmean
Stsd
StWithinperson
StBetweenMZ
StBetweenDZ

# Create Labels for Column and Diagonal Matrices

mLabs
sdLabs

# Create Lables for a Correlation Matrix

rphLabs

# Create Labels for Lower Triangular Matrices

MZbLabs
DZbLabs

# Specify Matrices to hold parameter estimates for MZ and DZ data

# elements for one overall Means

meansG

# elements for the SD

sdZ
sdG
sdT

# elements for the correlations

Rph
MZb
DZb
corMZ
corDZ

# generate expected covariance matrices

covMZ
covDZ

# Data objects for Multiple Groups

dataMZ
dataDZ

# Objective objects for Multiple Groups

objMZ
objDZ

# Combine Groups

pars
modelMZ
modelDZ
minus2ll
obj
ConCorModel
ConCorFit
ConCorSumm