R: st: Formula of F test for Wilk's lambda in MANOVA

[<carlo.lazzaro@tiscalinet.it>]

Dear Nick,
thanks a lot for pointing this out.
Actually, S.S. Wilks' final "s" slept in my keyboard.
Apologies for this.
Best wishes,
Carlo

-----Messaggio originale-----
Da: owner-statalist@hsphsun2.harvard.edu
[mailto:owner-statalist@hsphsun2.harvard.edu] Per conto di Nick Cox
Inviato: giovedì 12 gennaio 2012 10:37
A: statalist@hsphsun2.harvard.edu
Oggetto: Re: st: Formula of F test for Wilk's lambda in MANOVA

I'll let StataCorp comment on this if only because I have no copy of
Rencher. But a secondary detail is that this measure is attributed to S.S.
WIlks, so is Wilks' lambda. S.S. Wilks is not to be confused with M.B. Wilk;
there are vignettes on both people in the Stata manuals.

Nick


On Thu, Jan 12, 2012 at 9:25 AM,  <carlo.lazzaro@tiscalinet.it> wrote:

> in the .pdf manual of Stata 12 (page 374), the following formula of F 
> test for Wilk's lambda is reported:
>
> F=((1-Lambda^1/t)*df1)/((Lambda^1/t)*df2).
>
> However, taking a look at the same test in Rencher AC. Methods of 
> Multivariate Analysis.2nd ed. Wiley, 2002: 163, the formula turns the 
> degrees of freedom upside down:
>
> F=((1-Lambda^1/t)*df2)/((Lambda^1/t)*df1).
>
> Where's the bug, if any?
>
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/