Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!rutgers!sri-spam!sri-unix!hplabs!sdcrdcf!randvax!edhall
From: edhall@randvax.UUCP (Ed Hall)
Newsgroups: soc.motss,sci.med
Subject: Re: AIDS and Drug Usage
Message-ID: <628@randvax.UUCP>
Date: Thu, 23-Oct-86 04:14:40 EST
Article-I.D.: randvax.628
Posted: Thu Oct 23 04:14:40 1986
Date-Received: Sun, 26-Oct-86 02:36:28 EST
References: <1966@bu-cs.bu-cs.BU.EDU>
Reply-To: edhall@rand-unix.UUCP (Ed Hall)
Organization: Rand Corp., Santa Monica
Lines: 20
Xref: watmath soc.motss:179 sci.med:96

Repeat after me:

    Correlation does NOT imply causation.
    Correlation does NOT imply causation.
    Correlation does NOT imply causation.

As Barry Shein so aptly pointed out, the correlation between drug
use and AIDS can be explained by a third factor: life-style.

Many--I dare say a majority--of epidemiological studies are flawed by a
failure to rule out all significant confounding factors.  Unfortunately,
it is in the very situation where cooler heads should prevail that
statistical safeguards tend to be forgotten.  And, sad to say, some
otherwise brilliant researchers tend to be ignorant when it comes to the
proper use of statistics, leaving it to others to puzzle out just how
meaningful--or meaningless--the results really are.

		-Ed Hall
		decvax!randvax!edhall
		edhall@rand-unix.ARPA