Hello
I would like to calculate the covariance between the terms MS1+MSE1 and MS2+MSE2, where MS1 and MS2 are the mean squares for persons in models 1 and 2 and MSE1 and MSE2 are the corresponding mean square errors for these models. The two models are two-way mixed models with an interaction. However, the following pairs of variables are NOT independent:
1) MS1 and MSE2,
2) MS2 and MSE1
and
3) MSE1 and MSE2.
I would be most grateful for advice on how to approximate Cov(MS1 + MSE1, MS2+MSE2) when assumptions of independence are violated in the ways described above.
Many thanks
Best wishes
Margaret
---------------------------------
Inbox full of unwanted email? Get leading protection and 1GB storage with All New Yahoo! Mail.
|