Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Kaiser-Meyer-Olkin (KMO) statistic in SPSS

72 views
Skip to first unread message

aki...@research.baycrest.org

unread,
Feb 13, 2016, 6:48:05 PM2/13/16
to

I am performing PCA in SPPS 22. I get all of the statistics but I don't get Kaiser-Meyer-Olkin (KMO) statistic or Bartlett's test table in the SPSS output. I selected these options from gui. Does anyone know why they don't show up in the output?

Rich Ulrich

unread,
Feb 14, 2016, 1:09:10 PM2/14/16
to
On Sat, 13 Feb 2016 15:48:02 -0800 (PST),
aki...@research.baycrest.org wrote:

>
>I am performing PCA in SPPS 22. I get all of the statistics but I don't get Kaiser-Meyer-Olkin (KMO) statistic or Bartlett's test table in the SPSS output. I selected these options from gui. Does anyone know why they don't show up in the output?

How many cases do you have?
How many variables?

Was the output otherwise complete?
Did the output give you any notices or warnings?

Using the comments at
http://www-01.ibm.com/support/docview.wss?uid=swg21479963

- can you say, from inspection of results, what the KMO statistics
should have been?

--
Rich

David Marso

unread,
Feb 16, 2016, 7:02:31 PM2/16/16
to
Browsing online (everyone should try it) it seems KMO requires R inverse.
If the correlation matrix is singular then that may be one reason why KMO is not reported. If the concept of singular is Greek then I strongly suggest you do additional reading prior to pursuing any more FACTOR analytic work.
0 new messages