home *** CD-ROM | disk | FTP | other *** search
- Comments: Gated by NETNEWS@AUVM.AMERICAN.EDU
- Path: sparky!uunet!paladin.american.edu!auvm!UKC.AC.UK!MFF
- Via: UK.AC.UKC; 22 DEC 92 11:58:15 GMT
- Message-ID: <EDSTAT-L%92122206591872@NCSUVM.CC.NCSU.EDU>
- Newsgroups: bit.listserv.edstat-l
- Date: Tue, 22 Dec 1992 11:55:10 GMT
- Reply-To: "M.F.Fuller" <mff@UKC.AC.UK>
- Sender: Statistics Education Discussion <EDSTAT-L@NCSUVM.BITNET>
- From: "M.F.Fuller" <mff@UKC.AC.UK>
- Subject: Re: quartiles
- Lines: 13
-
- Flavia Joliffe's suggestion about percentiles does not suffer from the problem
- suggested by Murray Jorgensen.
-
- If the signs of the data are all reversed, then the observation in position i
- out of N is now in position N+1-i (Numbers 1 to N would become -N to -1 for
- example). So the 100P th percentile of the original set is in (ordered)
- position PN+1/2 by the definition and, apart from the change of sign, equals the
- observation in the (1-P)N+1/2 th position of the reversed signs set, i.e.,
- the observation which is the 100(1-P) th percentile of the reversed signs set.
-
- Mike Fuller
- Canterbury Business School, University of Kent, Canterbury, Kent, CT2 7PE,
- England. Phone +44 227 764000 x.7729; fax +44 227 761187; email mff@ukc.ac.uk
-