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, statalist.org is already up and running.


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

RE: st: Including constant?


From   Nick Cox <n.j.cox@durham.ac.uk>
To   "'statalist@hsphsun2.harvard.edu'" <statalist@hsphsun2.harvard.edu>
Subject   RE: st: Including constant?
Date   Wed, 1 Jun 2011 17:28:58 +0100

In various Earth and environmental sciences in which I write or review papers leaving out the constant would be generally regarded as gross statistical illiteracy.  

Nick 
n.j.cox@durham.ac.uk 

-----Original Message-----
From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Richard Williams

At 07:23 AM 6/1/2011, lreine ycenna wrote:
>Do we typically not have to include constants in regression tables
>when presenting them?

You can probably say most or all of what you want to say without 
including the constant. After all, how many times do you see a paper 
discussing the values of the constants?

But, I personally always include them. It is particularly useful if 
somebody wants to calculate a yhat value for some combination of 
values for the Xs. How useful the constant is as a standalone value 
depends on the coding of the Xs. The constant is the value someone 
would have if they had a value of 0 for every X. Often such a person 
cannot exist because one or more variables cannot take on a value of 
0, e.g. nobody has 0 height, and nobody gets a score of 0 on a test 
scaled to range between 400 and 1600. However, if you center the Xs 
(subtract the mean from each case for each X) then the constant 
becomes the predicted score for a person who has average values on 
every X. Such a person (or someone close to it) may actually exist, 
making the constant more interpretable as a stand-alone number, i.e. 
the constant is the score a totally average person would be expected 
to have. Or, if the model just has dummy variables (female, black) 
then the constant could be the average score for, say, a male white.


*
*   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 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index