NetNews Usenet Archive 1992 #19/NN_1992_19.iso/spool/bit/listserv/statl/1408

Text File | 1992-08-26 | 1.6 KB | 34 lines

Comments: Gated by [e-mail]
Path: sparky!uunet!paladin.american.edu!auvm!URIACC.BITNET!MARSH
Message-ID: <[e-mail]>
Newsgroups: bit.listserv.stat-l
Date:         Wed, 26 Aug 1992 22:34:12 EDT
Sender:       "STATISTICAL CONSULTING" <[e-mail]>
From:         Marshall Feldman <[e-mail]>
Subject:      Philosophy and Stats
Lines: 23

While a good deal of stats does not have an immediately obvious relation to
philosophy, at a very fundamental and important level the two disciplines
are inseperable.  In virtually all applications, statistics is about
obtaining knowledge through empirical measurements (describing, making
inferences, etc.).  Hence, unlike most other branches of mathematics,
it is impossible to separate statistics from epistemology and the
philosophy of science.  Some forms of statistics, such as so-called
"causal models" presume distict ontological positions, while others,
such as hypothesis testing, presume distinct epistemological positions.

I still remember the question on my probability theory final examination
twenty-five years ago: "Probability is a branch of mathematics and therefore
is based on the axiomatic method; it is inherently axiomatic and deductive.
Yet we use it to enhance our understanding of the world.  Explain how this
is possible."

The question underlies all empirical applications of statistics.

Marsh Feldman
Community Planning                      Phone: [phone]
204 Rodman Hall                           FAX: [phone]
University of Rhode Island           Internet: [e-mail]
Kingston, RI 02881-0815