[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: RE: ologit data handling

From   "Clive Nicholas" <>
Subject   Re: st: RE: ologit data handling
Date   Tue, 20 Nov 2007 05:11:21 +0000

Liren Zhou replied to Maarten Buis:

> What are you talking about? Yes, i know there are other
> models(softwares) can handle the ordered dependent variable, now i
> want to try ologit. clogit can not handle the ordered outcome. Can you
> tell me with ologit apply my case? As for the multiple row, it is not
> multiple records.  It is just multiple rows for each records.

With all due respect, you're not paying _any_ attention to what
Maarten correctly advised you (nor Nick, for that matter).

I understand what you're saying, but disabuse yourself of this
obsession with -ologit-: -clogit- really is a better solution to your
research question. You need to think of it as a choice between five
alternatives. As you correctly stated, the choice is _not_ ordered, so
why can't you see that this logically invalidates the use of -ologit-?
-clogit- is designed precisely to fit models to the kind of data you

If you need help in how to set up your data for the use of -clogit-
(or perhaps even -asclogit- if necessary), then try reading Chapter 7
of Long and Freese's excellent textbook (reference below), where you
may find their -case2alt- command useful (downloadable from SSC).

Hope all that helps: and don't bite the hand that feeds you! :)

Clive Nicholas

[Please DO NOT mail me personally here, but at
<>. Thanks!]

Long JS and Freese J (2006) "Regression Models For Categorical
Dependent Variables Using Stata" (2nd ed), College Station, TX: Stata
*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index