variance matrix

[amusante]

In known fate model I am generating seasonal survival rates from weekly rates and will use the delta method to calculate variances. But the variance-covariance matrix has many values of 0.0 most likely due to a majority of survival rates close to 1.0 and small error involved? I still need to know values to calculate variances for seasonal survival rates. How can I do this?? Thanks for any responses..........

Tony

[cooch]

In known fate model I am generating seasonal survival rates from weekly rates and will use the delta method to calculate variances. But the variance-covariance matrix has many values of 0.0 most likely due to a majority of survival rates close to 1.0 and small error involved? I still need to know values to calculate variances for seasonal survival rates. How can I do this?? Thanks for any responses..........

Tony

You say most of the values are 0 - presumably, you mean most of the covariances are 0 (not the variances)? Also, make sure you are not simply looking at the V-C matrix output in the Notepad - it rounds to too few digits to be useful. Look at the 'full V-C' matrix, either by outputting it to a dBase file, or into the clipboard, and then pasting into Excel (or some such). You may find fewer 0's than you thought.

But, on the assumption that all your covariances are 0, no problem - simply means your V-C matrix is diagonal. This changes nothing in particular - you simply use the diagonal matrix in the calculation, as described in Appendix B.