home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!stanford.edu!bu.edu!wupost!darwin.sura.net!paladin.american.edu!auvm!UHCCVX.UHCC.HAWAII.EDU!WENDELL
- From: WENDELL@UHCCVX.UHCC.HAWAII.EDU (John Wendell)
- Newsgroups: bit.listserv.stat-l
- Subject: Differences between correlation coeffients
- Message-ID: <721083383.910000.WENDELL@UHCCVX.UHCC.HAWAII.EDU>
- Date: 6 Nov 92 21:01:07 GMT
- Sender: "STATISTICAL CONSULTING" <STAT-L@MCGILL1.BITNET>
- Lines: 25
- Comments: Gated by NETNEWS@AUVM.AMERICAN.EDU
- Mail-System-Version: <VAX-MM(235)+TOPSLIB(133)+PONY(228)@UHCCVX.UHCC.HAWAII.EDU>
-
- I received several requests from people for a summary of what the
- replies to my question on how to calculate the significance between
- correlation coefficients calculated between a dependent variable and a
- set of two or more independent variables (from the same sample), so I
- will share the results for the whole list. I found the most useful
- source for my specific problem to be Xiao-Li Meng, Robert Rosenthal,
- and Donald B. Rubin, "Comparing Correlated Correlation Coefficients,"
- _Psychological Bulletin_, 1992, vol. 111, No. 1, 172-175. It provides
- a relatively easy to calculate approximation that seems to work well
- with most data likely to be encountered in behavioral studies and it
- provides numerical examples with intermediate results. They also
- provide a more accurate but more computationally intensive method in
- an Appendix. A very comprehensive treatment of the whole issue of
- testing correlation coefficients is James H. Steiger, "Tests for
- Comparing Elements of a Correlation Matrix," _Psychological Bulletin_,
- 1980, Vol. 87, No. 2, 245-251. Meng et al. and Steiger both make the
- point that the "traditional" method for comparing two correlation
- coefficients based on Hotelling's t-test has some serious drawbacks in
- most practical situations. Mahalo to James@ucrac1.ucr.edu, Dennis
- Roberts, Allan Reese, Karl Wuensch, and David Howell for their
- comments.
- Aloha, - John WENDELL@UHCCVX.UHCC.HAWAII.EDU.internet
- or WENDELL@UHCCVX.bitnet
- P.S. oops, missed Jerry Dallal in above. Mahalo Jerry!
- -------
-