Date:         Wed, 30 Jul 1997 11:06:16 -0700
Reply-To:     isabella mori <ivonhuen@SFU.CA>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@UGA.CC.UGA.EDU>
From:         isabella mori <ivonhuen@SFU.CA>
Subject:      Re: Principle components extraction
Content-Type: text/plain; charset="us-ascii"

At 01:13 PM 7/28/97 EST5EDT, bill dudley wrote:
>Factor anlaysis texts make is clear that there are both philosophical
>and statistical differences between a factor analysis and a principle
>component analysis.  SPSS contains a single factor anaysis module
>that has as the defulate a  Principle components extraction.  Is this the
same as a
>Priciple component analysis (as opposed to a  factor analysis?).  Is it
reasonable to
>rotate this result?
 
 
from what i understand, rotation IS used in principle components analysis.
"the method of component analysis involves the rotation of coordinate axes
to a new frame of reference in the total variable space - an orthogonal
transformation wherein each of the n original variables is describable
in terms of the n new principal components."  (harman: modern factor analysis,
p. 136).  the idea is to create new variables that are uncorrelated - that's
what the rotation is for.  if you only want to extract ONE principal
component,
there seem to be differing ideas as to whether to do that with or without
rotation (my statistics advisor prefers without).
 
on the subject of factor analysis in general:  i am baffled that the subject
of factor naming never seems to be treated at any length at all, at least in
the literature that i have seen and searched.  i have heard reference to
an article called "factor naming and the dangers of a classical education"
(or something similar) - anyone know about it?  or any other more than 2-
paragraph treatment of factor naming?
 
 
 
 
 
 
isabella mori
simon fraser university
vancouver, canada
email address: ivonhuen@sfu.ca
web site:
http://www.sfu.ca/~imori/index.htm