=========================================================================
Date: Thu, 20 Jul 2006 11:18:51 +0200
Reply-To: Samir Omeroviæ <samir.omerovic@gfk.ba>
Sender: "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From: Samir Omeroviæ <samir.omerovic@gfk.ba>
Subject: Re: Sphericity test
In-Reply-To: <1F6CEC6E7A3BDE459B06BECA190CB0C4E52200@ex-001.twentysixlondon.com>
Content-Type: text/plain; charset="iso-8859-2"
Thanks Michael,
One more question. I have checked the normality and it is ok. But what should I use to test
sphercity if I have small sample?
Thanks once more...
Samir
-----Original Message-----
From: Michael Pearmain [mailto:Michael.Pearmain@tangozebra.com]
Sent: Thursday, July 20, 2006 11:12 AM
To: Samir Omeroviæ; SPSSX-L@LISTSERV.UGA.EDU
Subject: Sphericity test
Hi Samir,
In addition to standard ANOVA assumptions, there is one specific to repeated measures when there are
more than two levels to a repeated measures factor. If a repeated measures factor contains only two
levels, there is only one difference variable that can be calculated and you need not be concerned
about the assumption. However, if a repeated measures factor has more than two levels, you generally
want an overall test of differences (main effect). Pooling the results of the contrasts between
conditions creates the test statistic (F). The assumption called sphericity deals with when such
pooling is appropriate. The basic idea is that if the results of two or more contrasts (the sums of
squares) are to be pooled, then they should be equally weighted and uncorrelated.
More over
If the sphericity assumption is met then the usual F test (pooling the results from each contrast)
is the most powerful test. Also, the test for sphericity itself is not sensitive when the sample
size is small, and the sphericity test is sensitive to lack of normality.
HTH
Mike
-----Original Message-----
From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] On Behalf Of Samir Omerovic
Sent: 20 July 2006 08:57
To: SPSSX-L@LISTSERV.UGA.EDU
Subject: Sphericity test
Hi to all,
I have one probably simple question for most of you. It was a long time I have studied and used
this so.
I have results from Mauchly's Test of Sphericity and I do not know what they mean. It is done in
SPSS.
Mauchley-s W - 0.477
Approx Chi Square - 10.685
DF - 9
Sig - 0.301
Epsilon
Greenhouse - 0.717
Huyhn - 0.891
Lower bound - 0.250
SUm of sq DF Mean SQ
F Sig
hemog Sphericity Assumed 14018,779 4 3504,695 47,180
,000
Greenhouse-Geisser 14018,779 2,869 4885,711 47,180
,000
Huynh-Feldt 14018,779 3,563 3935,023
47,180 ,000
Lower-bound 14018,779 1,000 14018,779
47,180 ,000
Error(hemog) Sphericity Assumed 4754,109 64 74,283
Greenhouse-Geisser 4754,109 45,909 103,554
Huynh-Feldt 4754,109 57,001 83,404
Lower-bound 4754,109 16,000 297,132
And later on i have Pairwise Comparinsons which I do know to explain. But I was wondering why did I
need the Sphericity test?
Any help on this?
Samir
________________________________________________________________________
This e-mail has been scanned for all viruses by Star. The service is powered by MessageLabs. For
more information on a proactive anti-virus service working around the clock, around the globe,
visit:
http://www.star.net.uk
________________________________________________________________________
______________________________________________________________________
This email has been scanned by the MessageLabs Email Security System.
For more information please visit http://www.messagelabs.com/email
______________________________________________________________________
__________ NOD32 1.1668 (20060719) Information __________
This message was checked by NOD32 antivirus system.
