home *** CD-ROM | disk | FTP | other *** search
- Comments: Gated by NETNEWS@AUVM.AMERICAN.EDU
- Path: sparky!uunet!paladin.american.edu!auvm!MAIL.SWIP.NET!MASTFOH
- Message-ID: <199301210955.AA08983@mail.swip.net>
- Newsgroups: bit.listserv.stat-l
- Date: Thu, 21 Jan 1993 10:57:11 -0100
- Sender: STATISTICAL CONSULTING <STAT-L@MCGILL1.BITNET>
- From: MAGNUS STENBECK <mastfoh%tellus.sos.se%sos@MAIL.SWIP.NET>
- Subject: Re: Nominal, Ordinal, Interval,
- and Ratio Typologies are Misleading
- Lines: 47
-
- > As the topic has frequently been raised on this list, I think
- that most
- > subscribers should read a recently published article by Paul
- Velleman and
- > Leland Wilkinson (both names which should be familiar to readers
- here) in the
- > February American Statistician entitled Nominal, Ordinal,
- Interval and Ratio
- > Typologies are Misleading. Just a single quote
- >
- > Unfortunately, the use of Stevens's categories in
- selecting or
- > recommending statistical analysis methods is
- inappropriate and can
- > often be wrong. They do not describe the attributes of
- real data that
- > are essential to good statistical analysis. Nor do they
- provide a
- > classification scheme appropriate for modern data
- analysis methods.
- >
- > Before critics start responding here, I suggest they go and read
- Paul and
- > Leland's thoughtful commentary.
- > -phil
-
- > --
- > J. Philip Miller, Professor, Division of Biostatistics, Box
- 8067
-
- I have not yet read this article, but I suggest that anyone
- interested in this topic also looks at Chapter 4: "On Scales of
- Measurement" from Otis Dudley Duncan's book "Notes on Social
- Measurement: Historical and Critical" (New York, Russel Sage,
- 1984). I cite from Duncan's own description of the contents of the
- chapter:
- "The theory of scale types proposed in 1946 by S.S. Stevens
- focused on nominal, ordinal, interval, and ratio scales of
- measurement. Some of his examples of these types - notably those
- concerning psychological test scores - are misleading. Steven's
- equating of scientific classification with measurement on a
- 'nominal scale' and his consequent underemphasis on counting (which
- actually employs an absolute scale) has mischievous consequences
- when taken seriously by population scientists. There is no clear
- place in his theory for the probability scale which, like counting,
- is central to the population sciences as well as important for some
- parts of physics."
-