home *** CD-ROM | disk | FTP | other *** search
- Comments: Gated by NETNEWS@AUVM.AMERICAN.EDU
- Path: sparky!uunet!uvaarpa!darwin.sura.net!paladin.american.edu!auvm!WUBIOS.WUSTL.EDU!PHIL
- Return-Path: <phil@wubios>
- X-Mailer: ELM [version 2.4 PL20A]
- Mime-Version: 1.0
- Content-Type: text/plain; charset=US-ASCII
- Content-Transfer-Encoding: 7bit
- Content-Length: 836
- Message-ID: <9301262202.AA29088@wubios>
- Newsgroups: bit.listserv.stat-l
- Date: Tue, 26 Jan 1993 16:02:51 -0600
- Sender: STATISTICAL CONSULTING <STAT-L@MCGILL1.BITNET>
- From: "J. Philip Miller" <phil@WUBIOS.WUSTL.EDU>
- Subject: t-test assumptions and SAS (fwd)
- Lines: 23
-
- > SAS tests the F ratio of the two samples and then computes t both
- > ways so you can choose the unequal t if F is significant.
- >
- That two step procedure is not an optimal strategy, see Miller & Wette, Proc
- of ASA Stat Computing Section, 1981 160-4 or Miller, Wette & Parks, SUGI
- Proceedings, 1981, 156-61.
-
- Having said that, those papers only talk about the case where both are normal.
- In most applications, the reason the variances are different is that they are
- not normally distributed and therefor some alternative procedure, such as a
- data transformation or a rank order test might be considered.
-
- -phil
-
-
- > Will@unh.edu
- >
-
-
- --
- J. Philip Miller, Professor, Division of Biostatistics, Box 8067
- Washington University Medical School, St. Louis MO 63110
- phil@wubios.WUstl.edu - (314) 362-3617 [362-2694(FAX)]
-