R help - Apr 2013 - (no subject)

Hi, I'm trying to decide between doing a FA or PCA and would appreciate some
pointers. I've got a questionnaire with latent items which the participants
answered on a Likert scale, and all I want to do at this point is to explore
the data and extract a number of factors/components. Would FA or PCA be most
appropriate in this case?
Cheers,
Hannah




--
View this message in context:
http://r.789695.n4.nabble.com/no-subject-tp4664248.html
Sent from the R help mailing list archive at Nabble.com.
	[[alternative HTML version deleted]]

On Apr 15, 2013, at 14:30 , ilovestats wrote:
> Hi, I'm trying to decide between doing a FA or PCA and would appreciate
some
> pointers. I've got a questionnaire with latent items which the
participants
> answered on a Likert scale, and all I want to do at this point is to
explore
> the data and extract a number of factors/components. Would FA or PCA be
most
> appropriate in this case?
> Cheers,
> Hannah
> 
> 

Not really an R question, is it? 

Stats.StackExchange.com is  -----> that way!

In terms of theory, PCA is essentially FA with the same residual variance in all
responses. With all-Likert scales, it is unlikely that there will be much of a
difference.

In practical terms:

- factanal can diverge (Heywood cases) which is a bit of a bother

on the other hand

- factor rotation is based on factanal() output; may require a little extra
diddling to work with prcomp().

I think I'd try factanal() first, and if it acts up, switch to prcomp().

-- 
Peter Dalgaard, Professor
Center for Statistics, Copenhagen Business School
Solbjerg Plads 3, 2000 Frederiksberg, Denmark
Phone: (+45)38153501
Email: pd.mes at cbs.dk  Priv: PDalgd at gmail.com