I am currently trying to find the meaning of principal components of my data. I find the raw loading very hard to interpret, because the first components drags a large part of the total variance (50 %).

So I decide to do a varimax rotation for a easier interpretation of my components.

I work in R, and with the psych packages, I can have to access to rotated scores too. Does this scores, have the same meaning of unrotated raw scores ? Should I just do a rotation for interpretation of components loadings, and project my observations in the unrotated space ?

Thanks !

  • 1
    $\begingroup$ Well, what are you going to do with the scores? $\endgroup$ – amoeba Jan 16 '18 at 9:58
  • $\begingroup$ @amoeba : Thanks for the quick response, I compute the PCA on language test scores, I study the impact of stroke in infants on language perfomance. I planned to use those scores in statistical analysis, where those scores can be entered as "predictors" for resting state functional fmri. data. $\endgroup$ – Dhaif Jan 16 '18 at 10:10
  • 1
    $\begingroup$ Then you probably want to use the interpretation of the rotated loadings when dealing with these new predictors. If so, you clearly have to use ROTATED scores. Unrotated scores do not correspond to the rotated loadings. $\endgroup$ – amoeba Jan 16 '18 at 10:16
  • $\begingroup$ @amoeba : So the interpretation of the projection of the individuals in the rotated space is not the same as in the unrotated space ? I will edit my answer to illustrate my question . $\endgroup$ – Dhaif Jan 16 '18 at 12:47

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.