home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: sci.math.stat
- Path: sparky!uunet!newsserver.pixel.kodak.com!psinntp!psinntp!isc-newsserver!jsvrc
- From: jsvrc@rc.rit.edu (J A Stephen Viggiano)
- Subject: R-Squared without Intercept
- Message-ID: <1992Sep11.195753.7900@ultb.isc.rit.edu>
- Summary: What should it be???
- Keywords: regression, r-squared
- Sender: news@ultb.isc.rit.edu (USENET News System)
- Nntp-Posting-Host: bruno.rc.rit.edu
- Organization: RIT Research Corp
- Date: Fri, 11 Sep 1992 19:57:53 GMT
- Lines: 14
-
- In a recent post on this group, somebody described a problem whose solution, in
- part, involved running a regression without an intercept term. Which leads to
- today's thought.
-
- One of the ubiquitous texts (maybe of the past -- was it Daniel & Wood?)
- related an anecdote of somebody who noticed that R-squared always was
- higher when no intercept term was included, so they wound up never including
- the intercept term.
-
- Does anybody have a citation for this anecdote? Does anybody have a related
- story to tell? Similarly, what suggestions do you have to offer in re: R-square
- from regressions without an intercept term? Thank you!
-
-
-