[Limdep Nlogit List] How to fix certain coefficients at zero for Latent Class Model?

William Greene wgreene at stern.nyu.edu
Tue Jul 29 02:21:45 EST 2008

M. Xie. Your base MNL model will contain 12 parameters, so the LCM
contains 24 plus the two for the class probabilities.  You can restrict
coefficients to equal zero or other fixed values, or to equal each 
other, by using
;RST = *,*,*,..... *
where the "*" is either a symbol for a free parameter, zero for fixing
the parameter to zero, or some other fixed value. There is a section
in the reference guide on ;RST for fixing coefficients.  Your list
of items in this specification will contain exactly 24 things, for the
12 parameters in each of the two class specific models. Do not include
any specifications for the probabilities (the 25th and 26th parameters).

/Bill Greene

----- Original Message -----
From: "yuanchang xie" <xieyc79 at gmail.com>
To: limdep at limdep.itls.usyd.edu.au
Sent: Monday, July 28, 2008 8:58:30 AM (GMT-0500) America/New_York
Subject: [Limdep Nlogit List] How to fix certain coefficients at zero for Latent Class Model?

Dear All,

I am trying to fit a latent class model with 2 classes.  Part of the
code is listed below

Rh2=one, Age, Male, Alcohol, Belt,Speed;

The p-value results show that some coefficients are not significantly
different from zero. So I want to fix those coefficients at zero, for
example, fixing the coefficient for Class 1 and Choice B at zero. Can
anyone kindly tell me how to do this? Also, in latent class model, how
can I set the base alternative? Thanks.

Yuanchang Xie
Limdep site list
Limdep at limdep.itls.usyd.edu.au

Professor William Greene
Department of Economics
Stern School of Business
New York University
44 West 4th St., Rm. 7-78
New York, NY   10012

More information about the Limdep mailing list