General Random Effects Model
A random effects model for data from a completely
nested design with n factors has the general form
![y_{i_1 i_2 ... i_n r} = \mu + \alpha_{i_1} +
\beta_{i_1 i_2} + ... + \epsilon_{i_1 i_2 ... i_n r}](images/neseq8.gif)
where
- yi1 i2 ... in r
- is the value of the dependent variable observed at the rth
replication with factor j at level ij, for j = 1, ... ,n.
![\mu](images/neseq2.gif)
- is the overall (fixed) mean of the sampled population.
![\alpha_{i_1},\beta_{i_1 i_2}, ... ,
\epsilon_{i_1 i_2 ... i_n r}](images/neseq9.gif)
- are mutually uncorrelated random effects with zero means and
respective variances
,
, ...,
.
Copyright © 1999 by SAS Institute Inc., Cary, NC, USA. All rights reserved.