s-news
[Top] [All Lists]

Which robust result is correct?

To: s-news@lists.biostat.wustl.edu
Subject: Which robust result is correct?
From: tangles@fastmail.fm
Date: Wed, 23 Jun 2004 16:39:20 -0500
Hello List,

I have been running some robust regressions (MM) and loaded the lmRob
library. When I do that a new option is added to the menus at the top of
the main window called Robust. Now when I run a robust linear regression
using that option I get different results from when I run the same
robust regression by selecting STATISTICS, REGRESSION, ROBUST MM. As far
as I can tell they should produce the same results but I must be missing
something because the R-square for the robust library I loaded is .2635
and the R-square for the robust method (in S-PLUS® 6.2 for Windows
professional edition) accessed by the menus is .4546. That's a big
difference in my book. Any ideas as to what's going on? Which should I
use? Thanks!
Dan

<Prev in Thread] Current Thread [Next in Thread>
  • Which robust result is correct?, tangles <=