|
Message/Author |
|
Anonymous posted on Thursday, December 01, 2005 - 1:30 pm
|
|
|
I'm performing ggmm. When I performed my missing data analysis, I noticed that I had 30 percent missing data on some cases. This reduces my usable data to about 700 cases. How badly will this harm my results? Also, Allison (2004) wrote a paper about FIMLs as a method to alleviate problems with missing data in SEMs. Would this be a viable method for my ggmm runs with these missing data? Thank you in advance |
|
|
Yes. Use TYPE=MISSING MIXTURE; |
|
Back to top |
|
|