Mauchly's sphericity test

[la_kennedy]

Hey,
I am running analysis to find how the manipulation of mood affects sporting performance. I have done pre and post mood manipulation tests on 2 sets of participants (positive and control), from this I get 6 lots of data (performance scores, positive mood scores, negative mood scores both pre and post).
I used a repeated measures anova on spss but the sig section in the Mauchly's sphericity test table is empty.
What am I doing wrong?
Any advice would be appreciated.
Thank you.
with regards,
la_kennedy

[Romuald Litwin]

Please be informed that Friday, July 16, 2010, was the last day I worked in the capacity of the Operations Manager at Baruch College Survey Research. If you have any questions about our current projects or you would like to talk about any BCSR-related business, kindly contact Ms. Mickey Blum at her business number [phone] or via [hidden email]. If you would like to reach me, please send an e-mail message to [hidden email] and I will respond as quickly as possible. Best regards, - Rom Litwin

[Bruce Weaver]

If you mean that you run a 2x2 between-within ANOVA on each outcome variable, then there can be no lack of sphericity--it can only be a problem when you have 3 or more levels for a repeated measures factor.

Also, if you are running 2x2 ANOVAs, you might consider ANCOVA instead, with Y = post, covariate = pre, fixed factor = group (positive v control).  For experiments (with random assignment to groups), this is usually the preferred model.

HTH.

la_kennedy wrote

Hey,
I am running analysis to find how the manipulation of mood affects sporting performance. I have done pre and post mood manipulation tests on 2 sets of participants (positive and control), from this I get 6 lots of data (performance scores, positive mood scores, negative mood scores both pre and post).
I used a repeated measures anova on spss but the sig section in the Mauchly's sphericity test table is empty.
What am I doing wrong?
Any advice would be appreciated.
Thank you.
with regards,
la_kennedy

[Mike]

Mauchley's test is inappropriate for a within-subject design with only
two time points/repeated measures.  Sphericity is automatically met
in this condition.  Jason Newsom explain this point in his notes for his
graduate data analysis class; see:
http://www.upa.pdx.edu/IOA/newsom/da1/ho_wsassump.pdf

In the SPSS repeated measures ANOVA there are several results
that are output which most people will routinely ignore if all they
want is a simple repeated measures ANOVA, results such as
Multivariate results and polynomial contrasts for within-subject
factors that purely categorical (i.e., they should only be used if
the levels of the within-subject factor can be order, best on a ratio
scale and not a factor such as different categories of words/stimuli).
Historically, SPSS came late to doing repeated measures ANOVA
and still make these types of "errors" when (1) is should know
whether the Mauchley test is relevant or not and omit it if it is
irrelevant, (2) have the analyst request multivariate results if they
want it, and (3) have the analyst request polynomial tests for
within-subject factors.

So, the Mauchley test results should be ignored, as should the
corrections in the ANOVA table for the within-subject factor
(why does SPSS print 3 additional lines of exactly the same
numbers???).

-Mike Palij
New York University
[hidden email]

----- Original Message -----
From: "la_kennedy" <[hidden email]>
To: <[hidden email]>
Sent: Tuesday, March 22, 2011 8:41 PM
Subject: Mauchly's sphericity test

> Hey,
> I am running analysis to find how the manipulation of mood affects sporting
> performance. I have done pre and post mood manipulation tests on 2 sets of
> participants (positive and control), from this I get 6 lots of data
> (performance scores, positive mood scores, negative mood scores both pre and
> post).
> I used a repeated measures anova on spss but the sig section in the
> Mauchly's sphericity test table is empty.
> What am I doing wrong?
> Any advice would be appreciated.
> Thank you.
> with regards,
> la_kennedy
>
> --
> View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Mauchly-s-sphericity-test-tp4258003p4258003.html
> Sent from the SPSSX Discussion mailing list archive at Nabble.com.
>
> =====================
> To manage your subscription to SPSSX-L, send a message to
> [hidden email] (not to SPSSX-L), with no body text except the
> command. To leave the list, send the command
> SIGNOFF SPSSX-L
> For a list of commands to manage subscriptions, send the command
> INFO REFCARD

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Bruce Weaver]

I agree with Mike's 3 suggestions below.  This would make it a lot easier for students (especially) to deal with the output.

Mike Palij wrote

Mauchley's test is inappropriate for a within-subject design with only
two time points/repeated measures.  Sphericity is automatically met
in this condition.  Jason Newsom explain this point in his notes for his
graduate data analysis class; see:
http://www.upa.pdx.edu/IOA/newsom/da1/ho_wsassump.pdf

In the SPSS repeated measures ANOVA there are several results
that are output which most people will routinely ignore if all they
want is a simple repeated measures ANOVA, results such as
Multivariate results and polynomial contrasts for within-subject
factors that purely categorical (i.e., they should only be used if
the levels of the within-subject factor can be order, best on a ratio
scale and not a factor such as different categories of words/stimuli).
Historically, SPSS came late to doing repeated measures ANOVA
and still make these types of "errors" when (1) is should know
whether the Mauchley test is relevant or not and omit it if it is
irrelevant, (2) have the analyst request multivariate results if they
want it, and (3) have the analyst request polynomial tests for
within-subject factors.

So, the Mauchley test results should be ignored, as should the
corrections in the ANOVA table for the within-subject factor
(why does SPSS print 3 additional lines of exactly the same
numbers???).

-Mike Palij
New York University
[e-mail]

----- Original Message -----
From: "la_kennedy" <[e-mail]>
To: <[e-mail]>
Sent: Tuesday, March 22, 2011 8:41 PM
Subject: Mauchly's sphericity test


> Hey,
> I am running analysis to find how the manipulation of mood affects sporting
> performance. I have done pre and post mood manipulation tests on 2 sets of
> participants (positive and control), from this I get 6 lots of data
> (performance scores, positive mood scores, negative mood scores both pre and
> post).
> I used a repeated measures anova on spss but the sig section in the
> Mauchly's sphericity test table is empty.
> What am I doing wrong?
> Any advice would be appreciated.
> Thank you.
> with regards,
> la_kennedy
>
> --
> View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Mauchly-s-sphericity-test-tp4258003p4258003.html
> Sent from the SPSSX Discussion mailing list archive at Nabble.com.
>
> =====================
> To manage your subscription to SPSSX-L, send a message to
> [e-mail] (not to SPSSX-L), with no body text except the
> command. To leave the list, send the command
> SIGNOFF SPSSX-L
> For a list of commands to manage subscriptions, send the command
> INFO REFCARD

=====================
To manage your subscription to SPSSX-L, send a message to
[e-mail] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD