Dear researchers
greeting
How much must the correlation between the extracted factors in "exploratory factor analysis" be in order to be able to use orthogonal rotations?
For example: we have extracted three factors that their correlations are more than 0.7.
Should we use orthogonal rotations or oblique rotation?
Sincerely yours
Hamid From Iran
Reasonable people may disagree, but I tend to always use oblique rotation as there is rarely a case in which it is defensible to expect 0 correlation between variables collected using standard approaches available to social scientists. Now, that being said, there may be a reason that you would like to have uncorrelated factors for subsequent research aims (i.e., including factor-derived scores in a regression).
In this case it seems hard to argue for orthogonal factors on empirical grounds seeing as how those factors appear to share ~50% or more of their variance. You may find if you take a confirmatory approach that a higher order factor exists, though the interpretation of higher order factors can sometimes be challenging.
Hope that helps. Best of luck!
Dear Hamid,
Another issue to consider is to compare the average correlation between variables in each factor and the average correlation between the factors as the former should be higher than the latter. This can indicate whether or not to use an orthogonal rotation. Please see my guide to EFA.
@Chalamalla
Thank you so much for your helping
The first article is excellent
Best wishes
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size