Date:         Wed, 17 Sep 2003 10:33:04 +0100
Reply-To:     jmcconnell@sigdiff.com
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         John McConnell <jmcconnell@sigdiff.com>
Subject:      Re: SPSS chokes, SAS computes principal components
Comments: To: paulandpen@optusnet.com.au
In-Reply-To:  <200309170110.h8H1At504954@mail015.syd.optusnet.com.au>
Content-Type: text/plain; charset="us-ascii"

Hi Sally,

I think Paul is probably spot on w.r.t. missing values.

SPSS Answernet reiterates and adds further references...

http://support.spss.com/answernet/details.asp?ID=24433

John

Significant Difference
Building 3
Chiswick Park
566 Chiswick High Road
London, W4 5YA

Mobile:       +44 (0) 7711 786600
Office:       +44 (0) 208 899-6029




> -----Original Message-----
> From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] On Behalf
Of
> Paul Dickson
> Sent: 17 September 2003 02:11
> To: SPSSX-L@LISTSERV.UGA.EDU
> Subject: Re: SPSS chokes, SAS computes principal components
>
> Hi Sally,
>
> On further thought, it is possible that this is occurring purely
because
> you are
> using pairwise deletion in SPSS.  The matrices in this case are being
> computed
> from a different set of cases each time.  Try listwise deletion to
start
> with, though I
> would never use this in my analyses and see if that makes any
difference.
>
> By trying a different method of handling missing data you should see
> whether
> your choice of treatment of missing data is the cause
>
> Cheers again Paul
>
>
> > Sally Zitzer <sallyz@u.washington.edu> wrote:
> >
> >   I have a client who is trying to run factor analysis (principal
> > components) in SPSS using pairwise deletion but gets the message
> > "nonpositive definite matrix".   When the same data is run in SAS
> > proc princomp SAS proceeds to do the analysis with no problem.
> > I just wondered if anyone on the list knows why SPSS comes to a
> > complete halt?   (my client is unhappy because she has only SPSS
> > and can't get SAS, she is in a university in Turkey).
> >
> >   Thanks,
> >
> >      Sally in Seattle
> >
> > Sally Zitzer, C&C, Senior Computing Specialist, Statistical Software
> > sallyz@cac.washington.edu or sallyz@u.washington.edu  (206)-543-8515