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

Re: st: What are the practical uses of the -offset- option?


From   John Hendrickx <john_hendrickx@yahoo.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: What are the practical uses of the -offset- option?
Date   Fri, 1 Oct 2004 04:23:56 -0700 (PDT)

An offset variable is a known component of the predicted value, in
effect it's given a coefficient of 1. It's routinely used in poisson
regression. 

I've used an offset variable for two other purposes. One was to
create a table with a given set of marginals and an association
pattern. If you fit a loglinear model to (artificial) data with the
marginals you want and the log pattern of association as an offset
variable, the predicted values constitute the table you're looking
for. A second use of an offset variable is to fix the effect of a
coefficient to a value other than 0.

Hope this clarifies things a bit,
John Hendrickx


--- Joseph Coveney <jcoveney@bigplanet.com> wrote:

> Is there a catalogue of the various practical uses of the
> -offset()- option
> (available in -logit-, -probit-, -ologit-, -oprobit-, . . .)?  I
> couldn't
> find any general description for its intended use(s) in the Stata
> User's
> Guide, Programming manual (didn't really expect to see anything
> there on it,
> anyway) or in the Reference manuals for the commands that have it
> as an
> option.  Its description under the Options sections of the various
> commands
> that offer it is pretty brief and doesn't lend any insight as to
> what it's
> for.
> 
> I knew of its use in profiling the log-likelihood to obtain
> likelihood-ratio
> confidence intervals--there's one user-written command (-logprof-)
> by Mark
> Pearce that automates the process for one estimation command
> (-logit-/-logistic-), and I've used Bill Gould's -bisect- for the
> same
> purpose for -probit- and -ologit-.  But I'm not aware of anything
> comprehensive written on -offset()-'s applications beyond that. 
> Today I
> stumbled across another application remeniscient to likelihood
> profiling,
> except that it doesn't search for log-likelihood values and doesn't
> calculate a deviance:  hypothesis testing of a regression
> coefficient at an
> alternative-hypothesis value different from its maximum likelihood
> estimate.
> 
> -findit offset- doesn't turn up anything on-target.  Is there a
> source for
> the beginner that I could refer to?
> 
> Joseph Coveney
> 
> 
> 
> 
> 
> *
> *   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/
> 



		
_______________________________
Do you Yahoo!?
Declare Yourself - Register online to vote today!
http://vote.yahoo.com
*
*   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