RE: ANOVA, ANCOVA significance

[morphmet]

-------- Original Message --------
Subject: RE: ANOVA, ANCOVA significance
Date: Fri, 28 Aug 2009 15:42:32 -0700 (PDT)
From: F. James Rohlf <ro...@life.bio.sunysb.edu>
Reply-To: ro...@life.bio.sunysb.edu
Organization: Stony Brook University
To: morphmet@morphometrics.org
References: <4a980e04.9080...@morphometrics.org>

That difference is to be expected. Doing the overall anova test has less
sensitivity for particular differences such as between a and b. Testing 
the various pairs as you did has more power but it assumes that these 
are of particular a priori interest. If you are going to make all 
pairwise comparisons then you have to take into account the fact that 
you are making multiple comparisons and you need to use something like a 
Bonferroni adjustment or another method to take into account that you 
are making multiple post hoc tests.

------------------------
F. James Rohlf, Distinguished Professor
Ecology & Evolution, Stony Brook University
www: http://life.bio.sunysb.edu/ee/rohlf


-----Original Message-----
From: morphmet [mailto:morphmet_modera...@morphometrics.org]
Sent: Friday, August 28, 2009 1:04 PM
To: morphmet
Subject: ANOVA, ANCOVA significance



-------- Original Message --------
Subject:        ANOVA, ANCOVA significance
Date:   Fri, 28 Aug 2009 10:02:29 -0700 (PDT)
From:   Javier Santos <javiersant...@hotmail.com>
To:     <morphmet@morphometrics.org>



Hello all,

I had a question conserning the ANOVA analysis between three groups. In
the overall ANOVA testing for differences between groups (F-statistic)
I
get unsignificant differences, while in pairwise t-testing significant
differences between groups do appear for variables with unsignficant
F-statistics. Should the significant t-tests be taken as valid, or only
those which have first been significant under the F-statistic?
I also extend this question to ANCOVA analysis, the only difference
here
being that the F-statistic is always used (at least in SPSS). What I
changed is the independent factor used from "species" to a pairwise
factor: "a-b", "a-c", "b-c".  By doing so, I also come across variables
that are unsignificant when using "species", but that are significant
in
one of the pairwise tests i.e. "a-b".
Does anyone know how I should interpret these results?

Many thanks,
Javier Santos
javiersant...@hotmail.com <mailto:javiersant...@hotmail.com>
Ugent / USAL

-----------------------------------------------------------------------
-
Share your memories online with anyone you want anyone you want.
<http://www.microsoft.com/middleeast/windows/windowslive/products/photo
s-share.aspx?tab=1>

--
Replies will be sent to the list.
For more information visit http://www.morphometrics.org




--
Replies will be sent to the list.
For more information visit http://www.morphometrics.org