| Date: | Wed, 28 May 2003 10:14:28 +0000 |
|---|
| Reply-To: | Patsy Chau <patsychau@hotmail.com> |
|---|
| Sender: | "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU> |
|---|
| From: | Patsy Chau <patsychau@hotmail.com> |
|---|
| Subject: | Regression--Multinomial |
|---|
| Content-Type: | text/plain; format=flowed |
|---|
I would like to ask about the procedure regression-multinomial. The first
step that I use is to weight the data set by count. But I find that the
p-value of goodness-of-fit test is different when I weight the data set by
proportion. The p-value is large when I use proportion, and small for count.
May I know why there is such a difference? Also, which setting shall I use?
Thanks a lot!
_________________________________________________________________
No masks required! Use MSN Messenger to chat with friends and family.
http://go.msnserver.com/HK/25382.asp
| 
