I am struggling with understanding the relationship (or rather, the interaction) between different factor rotation algorithms and score extraction methods in Exploratory Factor Analysis. As far as I understand, oblique rotations (like direct oblimin) allow for correlations between the extracted factors, while orthogonal rotations do not. This much is clear. What I don't understand is how the choice of score extraction method (Anderson-Rubin, Bartlett or Regression) interacts with the choice of rotation. From what I understand, the Anderson-Rubin method ensures that the factor scores will be orthogonal while the Bartlett method allows the factors to correlate. Does this mean that the Anderson-Rubin method should not be used following an oblique rotation? Or do the two things apply to completely separate things and there is something I'm sorely missing?

On a related note: Am I correct in assuming that if one plans to regress the score of one factor on the other factors, only the oblique-Bartlett combination should be used?

Please, try to keep the explanation relatively simple, as I don't have an extensive mathematical background.

Thank you!

  • $\begingroup$ stats.stackexchange.com/a/126985/3277 speaks of that. There is no interaction, you may use any conbination of methods of rotation and scores computation. (These methods, Anderson-Rubin, Bartlett or Regression etc are not factor extraction methods but are factor score estimation methods.) $\endgroup$ – ttnphns Feb 7 at 16:27
  • $\begingroup$ Thanks a ton, I'll check that! $\endgroup$ – Bálint L. Tóth Feb 7 at 16:36

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.