|
|
Hello good people of the SPSS variety,
I am after a little assistance with matrix algebra, that is, implementing
the following Bollen-Stine transformation in SPSS,
I am after Z data matrix,
where
Z= YS-½∑½
where
Y denotes a n x p data matrix, lets say 10 variables with 400 cases.
S denotes a sample covariance matrix of Y
∑ denotes the implied covariance matrix of my model.
Let’s also assume that I already have these matrices, I have pasted these
examples below to be a little bit helpful, and these would be used as input
files into any solution.
S
2.710
0.628 0.771
0.695 0.340 1.534
0.822 0.371 0.556 0.948
1.315 0.621 0.703 0.804 1.679
1.113 0.566 0.685 0.798 1.212 2.308
1.225 0.482 0.594 0.655 1.072 1.151 1.459
0.650 0.352 0.900 0.570 0.799 0.682 0.636 1.193
0.798 0.445 0.991 0.709 1.025 0.937 0.877 0.968 1.690
0.912 0.485 0.379 0.562 0.700 0.900 0.766 0.382 0.598 1.239
∑
2.731
0.627 0.771
0.812 0.375 1.534
0.845 0.39 0.505 0.949
1.284 0.593 0.768 0.799 1.679
1.297 0.599 0.776 0.807 1.227 2.322
1.144 0.528 0.684 0.712 1.082 1.093 1.459
0.799 0.369 0.841 0.497 0.756 0.764 0.674 1.191
0.958 0.442 1.008 0.596 0.906 0.916 0.807 0.993
1.69
0.859 0.397 0.514 0.535 0.813 0.821 0.724 0.506
0.606 1.242
I would be very grateful to anyone who can help with this, I am learning
matrix algebra but still have a way to go.
Last, I am certainly aware that I can do this in LISREL, AMOS, EQS, and
that C.Enders does have some SAS macros that can perform this etc etc but I
am after an SPSS programming solution for one reason or another.
Many thanks,
Jonathon Little
Postgraduate student
Curtin University, Perth, Australia
|
Locked
