Bookmark and Share

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

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

Re: st: Changing the reference category

From   Maarten buis <>
Subject   Re: st: Changing the reference category
Date   Wed, 20 Apr 2011 09:41:41 +0100 (BST)

--- On Wed, 20/4/11, Julia Martin wrote:
> I have created a reference category using the char varname[omit]#
> for regression analysis. I now do not want this reference category 
> - how do I get rid of it so it reverts back to the original 
> variable with NO reference categories. 

The default is not the absence of a reference category, as that 
does not make sense in most models under default conditions. Many
sources are available that explain this issue, here is one:

There are ways to make such models work, and they have their uses,
but they are not appropriate as a default. Below is an example.

*---------------- begin example ----------------------
sysuse auto, clear
recode rep78 1/2=3

// change the default reference catogory
char rep78[omit] 4
xi: reg mpg i.rep78

// restore the default
// note that there is still a reference category
char rep78[omit] 
xi: reg mpg i.rep78

// try to omit the reference category but
// that does not work as Stata will drop
// one category due to multicolinearity
xi, noomit : reg mpg i.rep78

// We need to remove the constant to 
// allow for a model without reference 
// category
xi, noomit : reg mpg i.rep78, nocons 
*------------------- end example -------------------
(For more on examples I sent to the Statalist see: )

Hope this helps,

Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen

*   For searches and help try:

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