"normal distribution" PreviousNext
Mplus Discussion > Confirmatory Factor Analysis >
Message/Author
 Rules Comm posted on Friday, July 01, 2011 - 8:33 pm
This is the second question related to the one-factor CFA I am working on. Thanks for your patience.

I found the model fit is not good. So, I went back to data and found that among 4 observed variables, three have very skewed distribution (all of them have the scale 0 to 1, and skewed to 0, meaning they have a lot of 0s).

I wanted to transform the data, but I have not found good method. I cannot use ln, since many data are 0. Do you have any suggestion on this part? Thanks a lot, professors.
 Linda K. Muthen posted on Friday, July 01, 2011 - 9:36 pm
What percentage are zeroes? Are you using MLR or MLM?
 Rules Comm posted on Saturday, July 02, 2011 - 3:56 pm
I just compared the results of MLR and MLM with those without using them. The results did not change a lot.

I may have to re-identify model or just give up CFA.
 Linda K. Muthen posted on Sunday, July 03, 2011 - 4:01 pm
If you get the same results with ML, MLM, and MLR this implies that you don't have severe non-normality of your data.
Back to top
Add Your Message Here
Post:
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Password:
Options: Enable HTML code in message
Automatically activate URLs in message
Action: