Bookmark and Share

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


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

st: robust regression


From   Michael Riechert <Michael.Riechert@the-klu.org>
To   "statalist@hsphsun2.harvard.edu" <statalist@hsphsun2.harvard.edu>
Subject   st: robust regression
Date   Wed, 19 Feb 2014 13:15:31 +0000

Dear STATA community,

I've run a robust regression by using the rreg command. I've compared the number of observation of an ordinary OLS (reg) and robust regression (rreg). In cases where both numbers are different (N rreg < N reg), I want to check which case has been deleted. Therefore, I used the "predict d1, cooksd" command to generate a variable containing the Cooks D which should be above 1 to be deleted in the robust regression (as far as I understood correctly). The robust regression weights all cases due to their Cooks D whereas observations with a low Cooks D are associated with a high weight. Cases with an Cooks D larger than 1 should be associated with a weight no wait at all or rather with a missing and will be excluded by the robust regression. 

Two things are strange to me. First, even if the difference in the number of observations is 1, two obs. have a Cooks D larger than 1. Second, I have far more obs. with a missing weight even if the associated Cooks D is below 1. Do you have any reason why this might be the case? Can I rely anyway on the results of the robust regression? Thank you!

Best,
Michael 

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/


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