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

st: Re: Testing coefficients across equations


From   "Steven Stillman" <stillman@rand.org>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: Re: Testing coefficients across equations
Date   Fri, 7 Jun 2002 16:33:40 -0700

There is a faq on this somewhere.  In order to test differences in
coefficients across different groups (if done jointly for all coefficients
it is often called a Wald Test) you need to interact your sex variable
(coded 0/1) with all of your independent variables and then run the model
(logistic hyperlip rf1 rf1*sex rf2 rf2*sex etc).  The coef on rf1*sex will
then tell you if there is a difference in the effect of rf1 on hyperlip
between genders, what it is, and whether this is significant.

Steve

<><><><><><><><><><><><><><><><><><><><><><><><><>
WORK RELATED                              PERSONAL
Email: stillman@rand.org                       Steven Stillman
Tel: 310-393-0411 ext. 7719               Email: s.stillman@verizon.net
Fax: 310-451-7061                              Tel: 310-268-8434
RAND                                                 11411 Ohio Ave. #9
Mail Stop M-15A                                 Los Angeles, CA 90025
1700 Main Street
www.thestillmans.org/econ.html
P.O. Box 2138
Santa Monica, CA 90407-2138
<><><><><><><><><><><><><><><><><><><><><><><><><>
----- Original Message -----
From: "AJ Bostian" <ajbostian@sonomedica.com>
To: <statalist@hsphsun2.harvard.edu>
Sent: Friday, June 07, 2002 7:49 AM
Subject: st: Testing coefficients across equations


Dear Statalist,

I am trying to test the equality of coefficients across two equations in
the following setting.  I am running two logistic regressions (males and
females) on the presence of hyperlipidemia using identical risk factors
in each.  (i.e., "by sex: logistic hyperlip rf1 rf2 rf3") I want to test
whether the coefficients as a whole are different between males and
females, and also whether specific ones are.

I believe the -test- command is the way to do this, and, in fact, there
is a very good example for this type of thing in the -test- helpfile
using -mlogit-.  However, in the mlogit case, it appears that the
equations are automatically named 0,1,2,etc. (based on the number of
cases).  Could anyone tell me the syntax for naming an equation and also
how to access an estimate by equation name (i.e., "display _b[rf1]" for
equation 1,2,3,etc.)?

Regards,
AJ Bostian

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index