Yes thanks. But with 150 or so variables to eliminate, one at a time is a
bit too slow!
Steve
-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Martin Weiss
Sent: 30 January 2009 15:47
To: [email protected]
Subject: AW: st: AW: estimation of series of OLS regressions based on
t-values from previous regression
<>
Sorry about that, I saw my own reply after about a minute. Anyway, are you
aware of this FAQ: http://www.stata.com/support/faqs/stat/stepwise.html ?
HTH
Martin
-----Ursprüngliche Nachricht-----
Von: [email protected]
[mailto:[email protected]] Im Auftrag von sdm1
Gesendet: Freitag, 30. Januar 2009 14:35
An: [email protected]
Betreff: RE: st: AW: estimation of series of OLS regressions based on
t-values from previous regression
Thanks for your reply Martin (but it didn't get through to me...I saw it as
part of Richard's reply).
Yes Richard, the problem is that with so many regressors dropping one term
at a time will be very slow. That's why I wanted to drop groups of
regressors based on their t-values.
Steve
-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Richard Williams
Sent: 30 January 2009 13:25
To: [email protected]; [email protected]
Subject: Re: st: AW: estimation of series of OLS regressions based on
t-values from previous regression
At 03:31 AM 1/30/2009, Martin Weiss wrote:
><>
>
>sdm,
>
>you could try -stepwise- with the -lockterm1- option. This option
>forces
>-stepwise- to hold on to the first term no matter what. You can specify
>this first term with parantheses so it can include several covariates,
>in your case x1- x200...
I almost suggested the same thing. One problem is that it might take an
eternity to run. Stepwise is going to drop one variable at a time, whereas
sdm is willing to drop huge chunks at a time that don't meet the desired
criteria. Still, it might be worth trying using the
lockterm1 and pr(.05) options.
My guess is that if somebody really wanted to, they could write a routine
that looked at the t-values and then only kept those Xs that met the desired
criteria. (But I don't really want to so I'm not going to bother!)
Of course, there are qualms about stepwise regression in general, let alone
stepwise regression involving 200 variables.
-------------------------------------------
Richard Williams, Notre Dame Dept of Sociology
OFFICE: (574)631-6668, (574)631-6463
HOME: (574)289-5227
EMAIL: [email protected]
WWW: http://www.nd.edu/~rwilliam
*
* 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/
*
* 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/
*
* 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/
*
* 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/
