Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down at the end of May, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re:Re: Re: st: How do I run a 3-way repeated ANOVA?


From   Philip Ender <ender97@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re:Re: Re: st: How do I run a 3-way repeated ANOVA?
Date   Fri, 19 Mar 2010 09:03:47 -0700

"RAMPL Linn" <Linn.RAMPL@insead.edu> wrote:

>When I use this type of notation you suggested to run a 2-way repeated
>Anova
>
>anova y x1##x2 subject
>...
>I get different results for the F-statistics than using
>
>anova y subject x1 / subject#x1 x2 / subject#x2 x1#x2, repeated(x1 x2)
>
>(this command is taken from http://www.stata.com/support/faqs/stat/anova2.html "No between-subjects
>factors with two repeated variables")

This comes down to the way that Kirk conceptualizes error versus the
Winer, Brown, and Michels approach.  For Kirk the error term is
subjects nested in x1#x2 while Winer, Brown, and Michels partitions
the nested error component into a number of pieces.  This is also the
approach used by SPSS when analyzing the data in the wide form.

In the approach that I suggested, anova y x1##x2 subject, the
residual for the model is just s|x1#x2.  In the end, I don't think
this is a matter of right and wrong rather it reflects differences in
how the error is parameterized.

Just to note: the repeated option does not change the model it only
provides additional information should the assumption of compound
symmetry be violated.

Phil
-- 
Phil Ender
UCLA Statistical Consulting Group
*
*   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/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index