home *** CD-ROM | disk | FTP | other *** search
- Comments: Gated by NETNEWS@AUVM.AMERICAN.EDU
- Path: sparky!uunet!gatech!paladin.american.edu!auvm!VTVM1.BITNET!WLTPIRIE
- Message-ID: <STAT-L%92111609274205@VM1.MCGILL.CA>
- Newsgroups: bit.listserv.stat-l
- Date: Mon, 16 Nov 1992 08:59:37 EST
- Sender: "STATISTICAL CONSULTING" <STAT-L@MCGILL1.BITNET>
- From: Walt Pirie <WLTPIRIE@VTVM1.BITNET>
- Subject: Testing for Variances
- Lines: 35
-
- Some imply it almost doesn't matter, everything is robust, others say throw
- a dandelion seed in the wind, yet others say let's vote on it, and finally many
- have their own (miriad) of favorite folk-lore remedies. What's going on here?
- This is all just mirroring the inadequacy of virtually all text-book treatments
- of this topic.
- The common thread through all of this is the idea that Location (means) is the
- only worthwhile subject of statistical inference (testing if that's your thing)
- , and questions about dispersion (variance) only arise in the context of assump
- tions underlying location procedures. Bunk.
- There are many situations where questions of location (accuracy) are relativel
- y minor and easily dealt with, but questions of dispersion (precision) are seri
- ous and difficult to deal with. If formal statistical procedures are important
- for location problems, then they must be also for dispersion problems. When we
- face up to that issue, then we'll be properly equipped to discuss it in the
- context of assumptions.
- And I'm not eschewing the use of visual techniques. I teach all my students tha
- t all statistical analysis is unacceptable unless accompanied by
- insightful visual displays.
- I can't help but think the reason problems of dispersion are so universally
- ignored is that good analytical solutions are hard to come by, so it's easier
- to pretend they aren't important.
- Walt.
-
- |==================================|============================|
- | Walter R. Pirie | |
- | Department of Statistics | |
- | Virginia Tech | |
- | Blacksburg, VA 24061-0439 | |
- | | |
- | Tel. 703-231-5441 | |
- | Bitnet WLTPIRIE@VTFVM1 | |
- | Telnet WLTPIRIE@VTVM1.CC.VT.EDU | |
- | | |
- | Fax 703-231-3863 | |
- |==================================|============================|
-