Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, is already up and running.

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

Re: st: oligit interpretation

From   Maarten Buis <>
Subject   Re: st: oligit interpretation
Date   Mon, 22 Oct 2012 11:14:07 +0200

You can define multiple dependent variables, and thus multiple
marginal effects, for this model. You seem to have three categories
for your dependent variable, so you can look at the probability of
getting category 1, getting category 2, or getting category 3. So that
is already three different outcomes you can look at. Moreover, you
have the latent variable. Than you can look at the odds of ending up
in category 2 versus 1 and the odds of ending up in category 3 versus
1 or 2.

So your first step needs to be to define more precisely what it is
exactly what you want to know. You'll probably will have to get access
to (and read) a book like:

-- Maarten

On Mon, Oct 22, 2012 at 10:09 AM, Lynn Lee <> wrote:
> Dear all,
> I am not sure how to interpret the marginal effect in ordered logit model.
> For example, in a ologit model, the coefficient of age is -0.03, (z=-17.43),
> the two threshold parameters are "/cut1=-1.4 ,  /cut2=0.95". I know that the
> latent dependent variable is decreasing with age (indicated by coefficient),
> but I do not know the exact marginal effect of age on dependent variable.
> Like in logit model, the maximum value of density is .25, so, it is easy to
> interpret the marginal effect of continuous regressor. But in ologit, there
> are several threshold parameters, is it still possible to use maximum value
> of density to interpret the marginal effect of continuous regressor?
> Any suggestion is appreciated.
> Best Regards,
> Lynn Lee
> *
> *   For searches and help try:
> *
> *
> *

Maarten L. Buis
Reichpietschufer 50
10785 Berlin
*   For searches and help try:

© Copyright 1996–2016 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index