|
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: st: Correspondence analysis for crossed - stacked variables
Not sure if this helps but I would check to see if the solutions really are different. Correspondence analysis is notorious for indeterminacies and for fits that appear different but give the same information. There are rotational indeterminacies, for instance and depending on how the loss function is defined it can be equivalent but not appear so. Changes in the normalization method will also alter the appearance of graphs without altering the fit (if I recall correctly).
I believe Michael Greenacre's newest book might be helpful:
MJ Greenacre. 2006. Multiple correspondence analysis and related methods. CRC Press.
-----Original Message-----
From: "Ina Krause" <[email protected]>
To: [email protected]
Sent: 1/27/2009 6:19 AM
Subject: st: Correspondence analysis for crossed - stacked variables
Dear Stata-User ,
I'd like do run a correspondence analysis for stacked-cross tables with
Stata 9.
May anyone can give me a hind what comand will be the best: "coranal"
with fweighting variable or the "camat" - comand with the
crosses-stacked table as matix input.
The answer would be very helpful for me. Because I get different
outcomes with these two analysis methods.
Yours
Ina Krause
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/