I have used PCA in my thesis and would like to argue (in best way) during viva the choice in addition to the fact that PCA analyses the variance of the observed items whereas FA analyses covariance.

  • 1
    $\begingroup$ Is not this something you should have been thinking when you decided to apply that method? Also, the question does not fit this site, it is about comparing two statistical/mathematical methods. $\endgroup$ – mmh May 7 '15 at 14:04
  • $\begingroup$ It's common for people do to PCA when they really should have done EFA (and to call the components that they extract 'factors'). The other way around is not so common. I'd suggest your best argument is "They are similar enough that it usually doesn't matter". Depending on your subject, it's likely that the examiners don't know though. $\endgroup$ – Jeremy Miles May 7 '15 at 16:35

Browse other questions tagged or ask your own question.