In article <[EMAIL PROTECTED]>,
Bob Hayden <[EMAIL PROTECTED]> wrote:
>----- Forwarded message from Eric Scharin -----
>I'm looking for a good introduction to statistics textbook. I'm
>interested in reading it mostly for fun and to become more "well
>rounded" mathematically. I have a B.S. in mathematics and took a
>single senior level statistics course in college (but have since
>probably forgotten a lot of it). I routinely read mathematically
>rigorous texts, so a rigorous (introduction) to statistics shouldn't
>bother me a bit.
>----- End of forwarded message from Eric Scharin -----
>First, thanks for that great quote, which I took the liberty of
>appending to my own sig file. Seemed appropriate for one who wears
>overalls!-)
>I was trained in math. Around 1982 I was asked to teach statistics
>and returned to Iowa State to learn some. Dean Issacson (with a joint
>appointment in math. and statistics) was temporary chair of the
>Stats. Dept. that summer. He said he would not let me take any
>math. courses and that what I needed was a chance to get my hands
>dirty with real data. For that purpose I would recommend thinks like
>David Moore's _Statistics: Concepts and Controversies_, Ed Tufte's
>Prentice-Hall book on statistics in public policy, working through
>_The Minitab Handbook_, etc. That way you find out what the questions
>are. You can always go back and look up the mathematical answers to
>those questions.
If you want to understand statistics, not how to carry out
the statistical manipulations commonly done by those who know
not what they are doing, you will avoid such like the plague
until you have the concepts down. Those books are recipes
for those who cannot possibly understand what is happening
during cooking, and are misused as the users cannot tell
whether they are cooking fish or fowl. For those not used
to cooking, I assure you that the same instructions are not
used for both.
David Moore's book does not even mention many of the important
concepts at all. In statistics one takes actions; these
actions have consequences in the various states of nature,
and what separates statistics from probability is that one
cannot assume knowledge of the state of nature. So
It is necessary to simulataneously consider
all consequences of the proposed action in
all states of nature.
You will not find this at all in most statistics books,
as it directly contradicts much which is maintained as
gospel by most of the expounders. Among these are the
use of fixed-level significance tests, p values, and
confidence regions; Bayesian confidence regions also
fail to meet this.
De Groot's book gives the Bayesian approach, but without
considering the consequences. There is a low-level book
by Clemen which does consider all of this, the title is
_Making Hard Decisions_.
--
This address is for information only. I do not claim that these views
are those of the Statistics Department or of Purdue University.
Herman Rubin, Dept. of Statistics, Purdue Univ., West Lafayette IN47907-1399
[EMAIL PROTECTED] Phone: (765)494-6054 FAX: (765)494-0558
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================
