home *** CD-ROM | disk | FTP | other *** search
- Comments: Gated by NETNEWS@AUVM.AMERICAN.EDU
- Path: sparky!uunet!europa.asd.contel.com!paladin.american.edu!auvm!UCLAMVS.BITNET!IARSPAL
- Message-ID: <EDSTAT-L%92111001570111@NCSUVM.CC.NCSU.EDU>
- Newsgroups: bit.listserv.edstat-l
- Date: Mon, 9 Nov 1992 22:53:00 PST
- Reply-To: "Peter A. Lachenbruch" <IARSPAL@UCLAMVS.BITNET>
- Sender: "Statistics Education Discussion" <EDSTAT-L@NCSUVM.BITNET>
- From: "Peter A. Lachenbruch" <IARSPAL@UCLAMVS.BITNET>
- Subject: fisher's z and r
- Lines: 9
-
- A recent comment on the Fisher z transformation reraised a question in my mind.
- It has long been known that the distribution of the correlation coefficient is
- very NONROBUST (shout from rooftops). An article by E. Pearson and another by
- C. Kowalski (roughly in the early 1970s) showed this. My question is "how
- robust are analyses which are based on the Fisher z transformation?" I've never
- seen anything on it, but maybe the literature I read doesn't give it much play
- - does anybody on the network have any information?
-
- Tony Lachenbruch
-