From: Damien Broderick (d.broderick@english.unimelb.edu.au)
Date: Thu Jul 13 2000 - 20:06:57 MDT
At 09:30 PM 13/07/00 -0400, Eliezer wrote:
>Is it just me, or does this totally fail to even begin to address my
>question? <which had to do with the claimed effect size >
It's just you. :)
Mr McMoneagle's somewhat rambling response does at least *begin* to address
the question of how his RV results were assessed and evaluated for
significance.
The elements of his responses, he tells us, were compared by judges to the
elements of a target, and recorded as hit or miss. (The determination of
what is taken to be an element looks somewhat subjective, it's true, but
since the judges are blind to the target this will hinder as often as it
helps; moreover, I suspect the intel-style protocol declares assessable
binary elements in advance.) Targets were usually drawn from a prepared set
of alternatives, or `target pool', and were completely unrestricted in
subject matter. This was done under lab conditions to prevent cueing or
sensory leakage.
To test how surprising McMoneagle's results were (a grossly informal test
but arguably not without merit), one might make copies of the set of
targets posted on his site and arrange some blind judging/matching of his
posted RV reports. How ambiguous are the drawings/descriptions? How many of
the targets might be matched to any given report?
Damien Broderick
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 15:29:55 MST