Message/Author |
|
|
Dear Dr. Muthen, I'm confused about the basic assumption of EFA in dealing with the ordinal data. Is that based on the normal assumption? Thank you very much. |
|
|
It is assumed that the factor is normal with a probit or logit link. |
|
|
Thanks for your answer. Is that reasonable to assume latent normality if our data are skewed ordinal? |
|
|
Yes. See slide 133 of the Topic 2 course handout which in on the website at: http://www.statmodel.com/trainhandouts.shtml |
|
Back to top |