Date: Tue, 8 Mar 2005 12:18:34 -0300
Reply-To: Hector Maletta <hmaletta@fibertel.com.ar>
Sender: "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From: Hector Maletta <hmaletta@fibertel.com.ar>
Subject: Re: Intrepretation from factor loading
In-Reply-To: <200503080301.j28314oI003644@listserv.cc.uga.edu>
Content-Type: text/plain; charset="us-ascii"
The loading being positive or negative depends on the way your variables are
coded. Suppose the variables are all about the same concept, such as living
standard or cognitive ability; suppose in all but one of them a higher value
of the observed variable means a higher IQ or a higher standard of living,
such as SAT scores or yearly income, but one of them is coded the opposite
way (lower value is better), such as "time to perform a task" or "number of
people per bedroom." A factor with high positive loadings for the other
variables may have a high but negative loading on this "contrarian"
variable. What is important is the absolute value of the loadings. Their
sign could be reversed by a different way of coding the variables. It is
advisable, but not really necessary, that all variables "point the same way"
in the sense that higher (or lower) scores are uniformly better in all of
them.
Hector
> -----Original Message-----
> From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU]
> On Behalf Of Raman
> Sent: Tuesday, March 08, 2005 12:01 AM
> To: SPSSX-L@LISTSERV.UGA.EDU
> Subject: Intrepretation from factor loading
>
>
> Hi all,
>
> After reading some journal, I am a bit confused in
> intrepretation the data.
>
> The first one is to take the absolue value of the factor
> loading. Then consider those variables with factor loading >
> 0.6 (for example) as the contributed variables in that factor.
>
> The second one is, in addition, to take the sign of the
> factor loading. e.g. factor loading of var1=0.8 and
> var2=-0.7, then both of them are regarded as important to
> that factor, but they have opposite effect.
>
> So, which one should I consider? And how to draw a
> conclusion on the factor?
>
> Thank you very much!
> Raman
>
| 
