home *** CD-ROM | disk | FTP | other *** search
/ NetNews Usenet Archive 1992 #26 / NN_1992_26.iso / spool / bit / listserv / edstatl / 867 < prev    next >
Encoding:
Internet Message Format  |  1992-11-12  |  1023 b 
  1. Path: sparky!uunet!utcsri!torn!nott!dgbt!netfs!ub!zaphod.mps.ohio-state.edu!darwin.sura.net!paladin.american.edu!auvm!LSE.AC.UK!P.JACKSON
  2. From: P.Jackson@LSE.AC.UK (P Jackson)
  3. Newsgroups: bit.listserv.edstat-l
  4. Subject: Re: robustness and correlation
  5. Message-ID: <9211121701.A00217@smtplink.lse.ac.uk>
  6. Date: 12 Nov 92 16:01:06 GMT
  7. Sender: "Statistics Education Discussion" <EDSTAT-L@NCSUVM.BITNET>
  8. Reply-To: P Jackson <P.Jackson@LSE.AC.UK>
  9. Lines: 12
  10. Comments: Gated by NETNEWS@AUVM.AMERICAN.EDU
  11. Via:        UK.AC.LSE; 12 NOV 92 16:01:33 GMT
  12.  
  13.  
  14.           Surley it is when considering whether or not a correlation
  15.           coefficient is significantly different from  value (usually
  16.           zero) that  distributional assumptions are important.
  17.           The coefficient itself does carry information regarding the
  18.           linear component of the association between two variables,
  19.           whatever the distribution, does it not ?
  20.  
  21.           Paul Jackson
  22.           Dept Social Psychology
  23.           LSE
  24.           London
  25.