> Just one question: How valid are the REML estimation of variance
> components when the data don't follow a normal distribution ?
What distribution does the data follow?
Does the central limit theorem hold?
ASREML provides other distributions such as the Poisson, Gamma than the
Normal and other link functions than the identity.
Asreml mailinglist archive: http://www.chiswick.anprod.csiro.au/lists/asreml