Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: Convergence time of logit with fixed effect


From   "Supnithadnaporn, Anupit" <gtg065t@mail.gatech.edu>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Convergence time of logit with fixed effect
Date   Mon, 9 Mar 2009 00:00:05 -0400 (EDT)

Hello,

Thanks for replying my post. 

For Glenn, 
I cannot use -clogit- because my data structure does not
allow me to do that. Basically, X3 is a group level information 
that I want to control. I thought that fixed effect would 
be enough for this purpose.

obsid	y	x1	x2	x3(stnid)
1	0	3	0	3
2	0	3	1	3
3	1	9	1	1
4	0	5	1	9
5	0	7	0	2
6	1	3	0	1
7	1	4	0	9
8	0	5	1	6
…	…	…	…	…


For Kyle, 
Thanks for the answers (1) and (2). In fact, I had used
-xtlogit- before doing this. And I got the following errors. 
So, I just try -xi:logit- and hope that it will give me something. 

Error message:
---------------------------
note: multiple positive outcomes within groups encountered.
note: 54 groups (2286 obs) dropped due to all positive or
      all negative outcomes.

<omit Iteration ... >


initial values not feasible
r(1400);

end of do-file
r(1400);
---------------------------

As you suggested that I can set the tolerance or the max number of
iteration, would the resulting estimates be correct? But, I will
try and see how they look like.


Thank you and best regards,
Anupit


*
*   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/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index