Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.


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

Re: st: gologit2 and mlogit coefficients do not agree


From   Richard Williams <[email protected]>
To   [email protected], [email protected]
Subject   Re: st: gologit2 and mlogit coefficients do not agree
Date   Sun, 12 Feb 2012 23:14:30 -0500

At 07:54 PM 2/12/2012, Rauscher, Garth wrote:
Dear listservers,

I am unable to reproduce the coefficients that I obtain from mlogit when I
attempt to run the same model in gologit2. As a simplified example of the
problem, my dependent variable (Y) has 3 categories (0,1,2) and I have a
single binary independent variable X (0,1). Mlogit gave me the same result
I obtained when I ran separate logistic regressions comparing Y=1 and Y=2
separately with Y=0, but gologit2 did not. My results are below. At first
I thought that gologit2 might be giving the inverse of mlogit but that is
not the case.  I like the flexibility of gologit2 but am not sure how to
interpret it's results.

Thanks for listening, Garth


. mlogit   y x , rrr baseoutcome(2)

Multinomial logistic          Number of obs   =    730
                              LR chi2(2)      =  25.52
                              Prob > chi2     = 0.0000
Log likelihood = -754.39125   Pseudo R2       = 0.0166

-------------------------------------------------------
            y |        RRR   Std. Err.      z    P>|z|
-------------+-----------------------------------------
0           x |   .3853242   .1040091    -3.53   0.000
1           x |   .3950005   .0858599    -4.27   0.000
2             |  (base outcome)
-------------------------------------------------------


. gologit2 y x, npl or

Generalized Ordered Logit    Number of obs   =    730
                             LR chi2(2)      =  25.52
                             Prob > chi2     = 0.0000
Log likelihood = -754.39125  Pseudo R2       = 0.0166

-------------------------------------------------------
           y | Odds Ratio   Std. Err.      z    P>|z|
-------------+-----------------------------------------
0           x |   1.822296   .4744057     2.31   0.021
1           x |   2.554348   .4826326     4.96   0.000
-------------------------------------------------------

To elaborate on my earlier message -- mlogit is basically 0 vs 2 and 1 vs 2. But gologit2 is like 0 versus 1 and 2 followed by 0 and 1 versus 2. With unconstrained models like this the fits are often identical or nearly identical, but the parameterizations are different.


-------------------------------------------
Richard Williams, Notre Dame Dept of Sociology
OFFICE: (574)631-6668, (574)631-6463
HOME:   (574)289-5227
EMAIL:  [email protected]
WWW:    http://www.nd.edu/~rwilliam

*
*   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 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index