Wed Feb 12 23:09:04 1997

From owner-tc-list  Wed Feb 12 23:09:04 1997
Return-Path: 
Received: by shemesh.scholar.emory.edu (SMI-8.6/SMI-SVR4)
	id XAA20606; Wed, 12 Feb 1997 23:09:04 -0500
Message-Id: <199702130408.UAA09751@m8.sprynet.com>
Comments: Authenticated sender is 
From: "Lewis Reich" 
To: tc-list@shemesh.scholar.emory.edu
Date: Wed, 12 Feb 1997 23:08:51 -500
MIME-Version: 1.0
Content-type: text/plain; charset=US-ASCII
Content-transfer-encoding: 7BIT
Subject: Re: Koren vs BHS text
Priority: normal
References: <199702121332.FAA08487@m8.sprynet.com>
In-reply-to: 
X-mailer: Pegasus Mail for Win32 (v2.50)
Sender: owner-tc-list@shemesh.scholar.emory.edu
Precedence: bulk
Reply-To: tc-list@scholar.cc.emory.edu
content-length: 1671

On 12 Feb 97 at 14:46, Nichael Cramer wrote:

> On Wed, 12 Feb 1997, Lewis Reich wrote:

> > I wonder if it is appropriate to describe the equidistant letters 
> > matter as a theory, since what is being reported is a result, and as 
> > far as I know, no particular theory was advanced in the statistical 
> > articles to account for the result.  

> While it is certainly true that "theory" is not the correct term to use 
> in this case, it is equally true that the "fact" of these claims has yet 
> to be established.

Part of my point was that as far as I can tell from the secondary 
accounts I have read of the statistical articles, no claims have been 
made.  An algorithm has been run on a text in a manner 
reasonably easily reproducible and has yielded certain results.  
Those results should be easy to verify by anyone who cares to 
reproduce the experiment.  Statistical analysis was then performed to 
test the likelihood that the results could have been obtained by 
chance.   In the social sciences, null hypotheses generally are 
accepted if the likelihood the results could have been obtained by 
chance is less than 5%.  In the reported instances, the likelihood of 
the results being due to chance was several orders of magnitude lower 
than that.  The statistical analysis was subject to the rigorous 
scrutiny applied in all cases to reputable refereed journals.  These 
are indeed facts.  The claims reported to have been made in the 
statistical articles do not go beyond what I have stated above.  If 
those secondary accounts are accurate, I  think it is quite safe to 
say that those claims have been established.

Lewis Reich
lbr@sprynet.com

Back