**Restricted model**:
A mixed model (i.e., with random and fixed factors) is termed restricted if a
random factor is not allowed to have fixed cross factors amongst its components
of variation estimated in the population. This restriction influences the
choice of error Mean Square for random effects. It is generally used for
balanced and orthogonal
designs; non-orthogonal designs requiring analysis by
General Linear Model use the unrestricted
model.

Doncaster, C. P. & Davey, A. J. H. (2007) *Analysis of Variance and Covariance: How to
Choose and Construct Models for the Life Sciences*. Cambridge: Cambridge
University Press.

http://www.southampton.ac.uk/~cpd/anovas/datasets/