Nesting: One factor is nested within another when each of its levels are tested in (or belong to) only one level of the other. For example a response measured per leaf for a treatment factor applied across replicate bushes must declare the bushes as a random factor nested in the treatment levels. The sampling unit of Leaf, L, is then correctly nested in Bush, B, nested in Treatment, T. The model is: Y = B(T) + ε, where the residual error term ε refers to L(B(T)). This model is called in a statistics package by requesting the terms: T + B(T) and declaring B as a random factor.

 

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/