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

st: Re: Overriding dropping of collinear variables

From   "Scott Merryman" <>
To   <>
Subject   st: Re: Overriding dropping of collinear variables
Date   Tue, 8 Jul 2003 21:09:23 -0500

----- Original Message ----- 
From: "James Valcour" <>
To: <>
Sent: Tuesday, July 08, 2003 10:45 AM
Subject: st: Overriding dropping of collinear variables

> Is there a way to tell Stata (either v7 or v8) not to drop collinear
>   Usually I wouldn't try and do this, but I'm trying to compare some
> produced from SAS's proc genmod with some glm output from Stata. SAS
> automatically drop collinear variables.  I'm trying to do this because all
> examples from a course I recently took were in SAS and I'm trying to see
if I
> can get the same results from Stata.
> I'm a regular Stata user and after using SAS again (I hadn't used it in a
> of years) I'm even more of a proponent of Stata.
> Thanks!
> James Valcour

I don't know if you can, however, I would be more concerned with SAS
estimating a model with perfectly collinear variables.  A standard
assumption in OLS is that the data matrix X is of full column rank.  If you
have collinear variables then this assumption is violated and the parameters
cannot be identified.

I believe that SAS will produce the following footnote in the output
regarding this problem:

"NOTE: The X'X matrix has been found to be singular and a generalized
inverse was used to solve the normal equations. Estimates followed by
the letter 'B' are biased, and are not unique estimators of the

Employing a generalized inverse as a solution to collinearity that SAS has
implemented strikes me a particularly dangerous.

Hope this helps,

*   For searches and help try:

© Copyright 1996–2022 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index