Hi to statistics people with knowledge

I am using 5 measures in a factor analysis to look for a latent variable. I have run the principal components default in SPSS however on reading up about factor analysis it suggests that principal components is different to factor analysis in that it reduces the data but it does not search for latent factors. Can anyone guide which method of extraction is used for searching for latent factors, rather than components? i.e. 'unweighted least squares' 'generalised least squares' 'maximum likelihood etc. these are the options given in SPSS.

Does that make sense?

My understanding is that PCA is exploratory FA, so it will look for latent factors - which will be the ones produced in your rotated matrix at the end.

What SPSS doesn't do is confirmatory FA, this is used once you have your factors and want to confirm they still apply/apply to a different sample. You can do this using Mplus, AMOS or Lisreal, but not SPSS.

(I think anyway!)

Within SPSS, the only option for principal component analysis is 'principal component analysis' (PCA). All other options are factor analysis. Principal Axis factor is probably the most commonly used factor analysis method used in SPSS, however, maximum likelihood is probably the best alternative, given the power of computers these days.

You might want to check out these videos that do a PCA in SPSS and a maximum likelihood FA in SPSS.

http://www.youtube.com/watch?v=qu4la8K212M

http://www.youtube.com/watch?v=iDT0tjU1TWo