Recently I was suggested to do Parallel Analysis and compare with EV>1 to determine no of factors for my scale. However, the scale I have developed uses Principal Axis Factoring and not PCA. If I use the SPSS syntax for Common Factor Analysis as given by Connors, all the mean EVs are less than 1.If I am using PCA then I am getting some factors with mean values greater than 1. I am confused as how to proceed further as the constructs in my scale are correlated and not suitable for PCA theoretically. Do I still proceed with Parallel Analysis?
If the concepts in your scale are correlated, you should use an oblique rotation (e.g., OBLIMIN), which is quite common for Principal Factors.
O'Connor's syntax for parallel analysis does support principal axis factoring (PAF), as stated in the webpage:
https://oconnor-psych.ok.ubc.ca/nfactors/nfactors.html
I haven't used it for a while. As far as I recall, it prints the results for both principal component analysis and PAF by default. If it does not, it should be easy to request them.
By the way, there may be a misunderstanding of how parallel analysis works. It was developed to address a problem with the "EV (eigenvalues) >1" rule. Therefore, if we use parallel analysis, we do *not* check the number of EVs > 1. We check the number of EVs greater than the corresponding EVs based on random data, although dfiferent cutoff values have been proposed (e.g., means, percentiles), and there are different ways to generate the random data (e.g., multivariate normal distribution or random permutation).
You can find more about parallel analysis from O'Connor's paper, or the following two papers:
Glorfeld, L. W. (1995). An improvement on Horn’s parallel analysis methodology for selecting the correct number of factors to retain. Educational and Psychological Measurement, 55(3), 377–393. https://doi.org/10.1177/0013164495055003002
Horn, J. L. L. (1965). A rationale and test for the number of factors in factor analysis. Psychometrika, 30(2), 179–185.
Hope this helps
- Badges
- Science topic
- Similar topics
- Medicine
- Disease
- Infectious Diseases