Newsgroups: comp.ai.philosophy,sci.cognitive
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news4.ner.bbnplanet.net!news3.near.net!news2.near.net!howland.reston.ans.net!news.sprintlink.net!dispatch.news.demon.net!demon!btnet!uunet!in1.uu.net!ncrgw2.ncr.com!ncrhub6!daynews!intruder!news
From: David E. Weldon, Ph.D. <David.E.Weldon@DaytonOH.ATTGIS.COM>
Subject: Re: FIRST order? was: why Ginsberg grouses
X-Nntp-Posting-Host: 149.25.61.42
Message-ID: <DCKB3w.81B@intruder.daytonoh.attgis.com>
Sender: news@intruder.daytonoh.attgis.com (News administrative Login)
Reply-To: David.E.Weldon@DaytonOH.ATTGIS.COM (WELDOD)
Organization: AT&T Global Info Solutions
X-Newsreader: DiscussIT 2.0.1.2 for MS Windows [AT&T Software Products Division]
References: <806807970snz@longley.demon.co.uk>
Date: Mon, 31 Jul 1995 03:59:56 GMT
Lines: 66
Xref: glinda.oz.cs.cmu.edu comp.ai.philosophy:31103 sci.cognitive:8709


}==========David Longley, 7/26/95==========
}
}In article <3v69ig$6u@nntp5.u.washington.edu>
}           mounce@u.washington.edu "R. Mounce" writes:
}
}> David Longley  <David@longley.demon.co.uk> wrote:
}> 
}> >    The   natural  assessments  of  representativeness   and 
}> >    availability do not conform to the extensional logic  of 
}> >    probability theory.'
}> 
}>         Well, they might have something here, but these are big
}> categories. 
}> 
}Well, Tversky and Kahneman are 'big researchers', but they are 
}only  part 
}of a wave of research which has its roots in work beginning with P. 
}Meehl 
}on Clinical vs Actuarial judgement in the early 1950s,  and  H. 
}Kelley on
}the attribution of causality at the same time (beginning  of  
}Attribution
}Theory - which is largely phenomenological/methodologically 
}solipsistic).
}
}Again - see 'Fragments of Behaviour: The Extensional Stance' 
}which can be
}found in sci.psychology.theory  22/7/95 and which partially 
}prompted this
}thread originally (Quine and FOL). 
}-- 
}David Longley
}
I can't let this go on.  The entire literature on clinical judgment is
terribly flawed and hence has no bearing here.  It should not be used to
support any argument of the type put forth by David Longley.

What are the flaws?
They are numerous, but the following will suffice:
1.  Medical Model II (or whatever level it is now) classifies mental
pathologies.  But these categories are at best vague and controversial.  No
one should use them to study the human judgment process.
2.  The criterion used in clinical judgment studies represents the consensus
judgement of "trained professionals," but the truth is, no two phychologists
can agree on what the cases represent, so the researchers must use some
measure of central tendency to assign the "cases" to a clinical category. 
Once this is done, it is tacitly assumed by the researchers that the
"consensus" judgments correctly classify the cases.
3.  Then the researchers ask volunteers to classify the cases, but instead of
aggregating the volunteer judges responses, each volunteer's judgements are
compared to the "correct" judgements.  Lo, and behold, the volunteer's
judgements do not match the "correct" judgments very well (the correlation
between the two ranges from .40 to .80) even though the results can be
predicted from an elementary knowledge of statistics and understanding of the
impact of using a measure of central tendency for one set of measurements and
raw judgments for the other set.

A far better test of David Longley's assertions is found in the social
perception area of social psychology.  There the technique is to average the
judgments of the volunteers as well.  In this variation, there is no
statistical artifact and the correlations between the two sets of judgments is
usually between .95 and .995 with the distribution skewed toward the higher
number.


