[R] Wilks lamda and single discriminant function

[Joël Meunier]

Dear R-users,

I'm wondering how to obtain Wilks-lambda values when discriminant 
analyses have only one discriminant function (i.e. 2 categories to 
discriminate between).

The use of manova(predictions~groups, test="Wilks") asks for multiple 
response and the use of anova(lm(predictions~groups), test="Wilks") 
simply does not consider the last term.

(The alternative question is whether such Wilks-lambda value is relevant 
for those simple comparisons between two groups of predicted values or 
whether t-test like are enough?)

Thanks a lot,

Joël



-- 

Joël Meunier, PhD

University of Basel
Zoological Institute
Evolutionary Biology
Vesalgasse 1, 4051 Basel - Switzerland