Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10 5/3/83; site houxz.UUCP
Path: utzoo!linus!decvax!microsoft!uw-beaver!cornell!vax135!ariel!houti!hogpc!houxm!houxz!halle1
From: halle1@houxz.UUCP
Newsgroups: net.misc
Subject: Re: The Earth-Centered Universe
Message-ID: <341@houxz.UUCP>
Date: Tue, 19-Jul-83 10:54:18 EDT
Article-I.D.: houxz.341
Posted: Tue Jul 19 10:54:18 1983
Date-Received: Wed, 20-Jul-83 01:17:28 EDT
References: <5567@unc.UUCP>
Organization: Bell Labs, Holmdel NJ
Lines: 14

Tim, your experimental idea seems sound, but there is a potential problem
with the interpretation of the results.  Suppose out of 120 subjects, only
1 was identified properly.  Pure chance would say 10 should be identified
correctly.  Then some might say, "Astrologers results are significantly
different than what pure chance would predict," ignoring the fact that they
were WORSE than chance.  If results were not given, they'd get away with it.

Unlikely, you say?  Well, this has already happened.  One of the astronauts
to the moon, Buzz Aldren, I think, did an ESP experiment on the way with
"psychics" on earth.  The results were significantly worse than chance.
Besides, the experiment was invalid for several other reasons.  Yet, the
results were reported as I stated above.  "These results could occur by
chance only once out of a million times."  So the hype went.
So beware of tests of this sort.