Date: Tue, 27 Jun 95 23:17:54 CDT
From: rusin (Dave Rusin)
To: johnd@deltec.co.nz
Subject: Re: Numerical Paradox - any collections of ?
Newsgroups: sci.math.num-analysis

In article <1995Jun26.013512.29655@deltec.co.nz> you write:
>Any suggested reading for the not-so-numerate, with
>regard to any numerical or statistical paradox ?

"How to Lie with Statistics". I think it was in that book that I was
first introduced to this paradox:

	Imagine an idealized university with two departments, math and
literature.  A complaint leads to the following discovery: the
university is accepting a smaller portion of female applicants as
students than male applicants. Naturally, the university seeks to fix
the problem, beginning with the determination of which department is
in the wrong. The following data result.

Number of students applying and accepted
              Male              Female
Math         80, 30 (37.5%)    20, 10 (50%)
Literature   20,  5 (25%)      80, 22 (27.5%)
Total       100, 35           100, 32

How, then, does one assign blame in the university or (more constructively)
ameliorate the situation with fairness to all?

The classic "Mathematics made difficult" is often mentioned in the math
newsgroup. It includes some paradoxical material. Outside of logic/set theory
and statistics/probability, I don't think you will find many real
"paradoxes", but there are certainly some surprising results in mathematics.

dave

PS - I forgot to mention things like the paint can which only holds a
finite volume of paint but whose inner surface requires an infinite
amount of paint to coat it. Must be one of the surfaces z=-r^k in
cylindrical coordinates.