Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10 5/3/83; site gatech.UUCP
Path: utzoo!linus!wivax!decvax!microsoft!uw-beaver!cornell!vax135!floyd!whuxlb!pyuxll!eisx!npoiv!npois!hogpc!houxm!hocda!spanky!burl!sb1!sb6!emory!gatech!spaf
From: spaf@gatech.UUCP
Newsgroups: net.misc
Subject: Re: Re: The Earth-Centered Universe
Message-ID: <307@gatech.UUCP>
Date: Tue, 19-Jul-83 21:36:17 EDT
Article-I.D.: gatech.307
Posted: Tue Jul 19 21:36:17 1983
Date-Received: Fri, 22-Jul-83 03:02:40 EDT
References: <341@houxz.UUCP>
Organization: Georgia Tech, School of ICS
Lines: 37

The implication was made that results which are significantly worse
than predicted by probability should not be interpreted in a
positive manner.  I disagree with this.  If you perform a carefully
organized experiment and get results far from what you expected
then you have gained information, although you must be careful
how you interpret that information.

As an example, consider a test of ESP phenomena.  Suppose you
go through a deck of cards and ask someone to guess the color
of each card in turn.  Suppose in 100 attempts the person
got only 10 correct.  Suppose this was a carefully controlled
experiment and the results were repeatable.  Would that indicate
that there is no such thing as ESP?  I don't believe so.
It indicates something, although maybe not exactly what you
wanted to prove.  (BTW, I believe something like this
has been shown to be the case.  Some very reputable institutions
doing ESP research have found "negative" receivers who show
an ability to detect what something hidden from view is *not*.)

Other examples would be like finding 90% less gravitational
attraction at some point, or less electrical charge, or
less free ions in solution, or whatever.  It simply proves
that something in your original theory was wrong, but you are
measuring something that can be classed as information.

Finding only 1 person out of 120 when you expected 10 is a 
possibly significant result, especially if you can extend
the results to 120,000 test subjects.  You just have to
be a little more careful of the conclusions you draw.

-- 
"The soapbox of Gene Spafford"

CSNet:		Spaf @ GATech		
Internet:	Spaf.GATech @ UDel-Relay
uucp:		...!{sb1,allegra}!gatech!spaf
		...!duke!mcnc!msdc!gatech!spaf