This video shows you how to run an exploratory factor analysis in R and how to write it up Data and code can be found here drive.google.com/drive/folder...
Thanks for the wonderful lecture. I have two questions in EFA: First question: The assumption in EFA, that the measurement errors are not correlated between them isn't valid in reality? Second question: Can you send me in Data under which if I apply EFA I find heywood cases.
Thank you very much for the great lecture. I have a question: I run the same test with my Likert-scale data (203 responses, 20 questions/ nfactors = 20). In this first test, there were 6 factors with the eigenvalue 1.00 or more. Then, I did the test again just like you (nfactors = 6), and this time there were 5 factors with the eigenvalue 1.00 or more, and not 6 (one of the factors was 0.95 now). In this case, should I take these 5 factors with the eigenvalue 1.00 or more, or should I take 6 of them?
It's not particularly convenient to do it as the package doesn't lend itself to exporting but I have made this for you which is the best possible way I have found hope it helps! ruclips.net/video/SdG061mpfFw/видео.html
Thank you. This is really useful.
Thanks for the wonderful lecture.
I have two questions in EFA:
First question:
The assumption in EFA, that the measurement errors are not correlated between them isn't valid in reality?
Second question:
Can you send me in Data under which if I apply EFA I find heywood cases.
Thank you very much for the great lecture. I have a question:
I run the same test with my Likert-scale data (203 responses, 20 questions/ nfactors = 20). In this first test, there were 6 factors with the eigenvalue 1.00 or more. Then, I did the test again just like you (nfactors = 6), and this time there were 5 factors with the eigenvalue 1.00 or more, and not 6 (one of the factors was 0.95 now). In this case, should I take these 5 factors with the eigenvalue 1.00 or more, or should I take 6 of them?
Don’t know if you’re still wondering, but parallel analysis is likely to be a more accurate way of determining the number of factors to be extracted
Hi! Thanks a lot. Can you try that for dichotomous data set with using polychoric / tetrachoric correlation?
Yes you can easily do the same thing with polychoric / Tetrachoric correlation (the former for ordinal, the latter for dichotomous)
eg
Poly_DATA
@@DrPC_statistics_guides can you share how to do an EFA bifactor analysis?
thank you for this video.
i have a question please
in EFA we have to determine the factoring method , but you didn't , why ? . thank you again
Sorry Im not clear what you are referring to- the actual method e.g. ML, OLS, or Minres or do you ascertaining the number of factors?
very useful. Can you also explain how to export the EFA output into excel?
It's not particularly convenient to do it as the package doesn't lend itself to exporting but I have made this for you which is the best possible way I have found hope it helps! ruclips.net/video/SdG061mpfFw/видео.html
I tried this in R studio (ver 2022) and there was no "fa" function... did it get updated?
Have you got the psych package installed and activated from the library?
MR1, MR2 .... what is the official name for this? you said are Eigen factor value, right?
they are often called Eigenvalues
@@DrPC_statistics_guides thanks a lot! Now I am thinking about go for EFA or CFA to build my model in the master thesis. but this video helps a lot!