The example below shows a very small dataset, much smaller than one would normally care to do an analysis of variance for.

I am trying to understand the meaning of the ANOVA table in LinearModelFit when the ANOVA package says that there isn't enough data to do an analysis of variance.

This produces an error message to the effect that there is not enough data. It does produce an ANOVA table, but there is a complex infinity in it and a very small number for the error sum of squares (~10^-30).

Needs["ANOVA`"]
data = {{1, 2}, {2, 4}, {5, 2}};
ANOVA[data]

On the other hand, this produces a perfectly reasonable ANOVA table with an error sum of squares of about 2.5 (the value of data has NOT been changed).

lm = LinearModelFit[data, x, x]
lm["ANOVATable"]

Any ideas will be appreciated. Thanks,

OL.