Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: gologit vs logit

From   Richard Williams <[email protected]>
To   [email protected]
Subject   Re: st: gologit vs logit
Date   Mon, 06 Jun 2005 12:30:13 -0500

At 01:07 PM 6/6/2005 -0400, Zhu, Carolyn wrote:

Hi Statalister,

We estimated a gologit model where the depedent variable has 4 categories from best to worst. We compared coefficients from the generalized ordered logit model with the three binary logit models. We had thought these results might be similar. But they are very different (e.g., some coefficients are twice as large) and tell a very different stody. We are puzzled as to what might be going on and wondering if the experts on the list would tell us what we might be doing wrong.
First, are you sure you did the binary logits right? I suggest using the brant command that is part of the spost package of J. Scott Long, e.g. do something like

ologit y x1 x2
brant, detail
gologit y x1 x2

Second, you can try my gologit2 program, which does everything gologit does and a lot more. (See especially the -autofit- option). From within Stata, type

net from

If things still look weird, I'd be interested in seeing your output. My experience is that -brant, detail- and -gologit- tend to produce very similar results. (And if you have any feedback on gologit2, please let me know - I'll probably send it to ssc soon unless somebody tells me they have found horrible mistakes!)

Richard Williams, Notre Dame Dept of Sociology
OFFICE: (574)631-6668, (574)631-6463
FAX: (574)288-4373
HOME: (574)289-5227
EMAIL: [email protected]
WWW (personal):
WWW (department):

* For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index