Dear Nripenda,

Please see my guide to EFA. The eigenvalue threshold of 1 often produces too many factors. I would recommend a different method to principal component analysis that focuses on shared variance rather than total variance.

It is acceptable to fix the number of factors but you need to assess the weights of the items on each factor and whether any items have high cross loadings.

If you are developing a new instrument, then based on theory or logic, you can specify the number of factors rather than depending on the eigenvalue. However, EFA need to be conducted with patience, the researcher has to try different methods and extractions until the best items distribution with meaning and highest variance.

Some examples are attached, you can also check the reference list.

Hi Dr. Ahmad,

Thanks for the quick reply and sharing your work. I highly appreciate it.

Still, I am not clear, what is the significance of using the restricted or forced EFA (where a researcher specifies the number of factors).

Please suggest,

Thank you,

Nripendra

Hi Peter,

Thank you very much for your providing your guide. It is very helpful.

Can you please provide me the correct APA citation for this guide?

Thank you,

Nripendra

Nripendra- good that you followed up on that point. Be sure to cite these other sources in the document.

Thank you Dr. Bob.

Hi Peter,

Thank you for the private message on my email. It is very helpful.