Path: utzoo!attcan!uunet!decwrl!limbo!taylor
From: eugene@wilbur.nas.nasa.gov (Eugene N. Miya)
Newsgroups: comp.society
Subject: Re: Student Writing: Can the Machine Maim the Message?
Message-ID: <986@limbo.Intuitive.Com>
Date: 13 Jul 90 16:09:44 GMT
Sender: taylor@limbo.Intuitive.Com
Organization: NASA Ames Research Center, Moffett Field, CA
Lines: 12
Approved: taylor@Limbo.Intuitive.Com

Rather than flame the article, the scientific thing to do is replicate
the experiment changing a variable or two (factorial design) and
invalidate the earlier study.  If some serious professional seeks to do
such I suggest getting a copy Campbell and Stanley's "Experimental and
Quasi-Experimental Designs for Research" [Rand-Mcnally] to seek out
sources of internal and external invalidity.

But I am skeptical anyone would do this.  As Peter Denning noted, few
studies ever get replicated in computer science.  Make my day, prove me
wrong.

e.  nobuo miya