[net.sci] Repeatability of Human based experiments

rb@cci632.UUCP (07/03/86)

In article <1292@psivax.UUCP> friesen@psivax.UUCP (Stanley Friesen) writes:
>In article <3719@decwrl.DEC.COM> cooper@pbsvax.dec.com (Topher Cooper) writes:
>>>In general, parapsychologists are unwilling to accept *any* results until
>>>they have been replicated.  This is in stark contrast to other fields where
>>>(at least the publication of) replications of existing results are almost
>>>never done (I know that doesn't agree with what you were taught in junior
>>>high science, but it seems to be true).
>
>	I think there are several aspects to this.
>
>	1) *Successful* exact replication of an experiment is
>relatively uninteresting, and if the original results are not
>extraordinary or important enough, it is simply not worth the effort
>to *publish* a replication. This does not mean that replications are
>not being done, only that they are not being published. Many
>experiments get turned into undergraduate class assignments or
>incorporated as part of other experiments. Believe me, if they failed
>to replicate, *that* *would* be published.

>	2) For sufficiently unusual or important results, *published*
>replication is required for acceptance. Look at Particle Physics and
>see how much replication there is when a new particle is announced! Or
>see what happened when the intron/exon structure of Eukaryote DNA was
>discovered!
>				Sarima (Stanley Friesen)
>

It is interesting that many psi experiments, particularly those involving
psychokinesis (sp?) and telepathy,
1: Base their results on statistical evidence (which is acknowledged to
   be less than conclusive).
2: Are expected to yield the same statistical results with a different
   set of subjects, yet providing little, if information about the subjects.

Think of how obsurd this might be.  A researcher in San Francisco, attempting
to duplicate Kinsey's research would very likely get a different result than
a similar research effort in Houston Texas.

In addition, filtering of subjects is considered dubious at best, fraud
more often.

Remember, PSI research is research of the para-Normal.  You might find
to "textbook paranoids" in two different mental hospitals, but what
are your odds of finding two hospitals with the exact same mix of
disorders?

What PSI research needs is more classification and organization of the
factors that appear to effect results.  Unfortunately, attempts to
do this to any significant degree tends to be treated as "bogus science".

Are there any such "classification" efforts under way?

dobro@ulowell.UUCP (Gryphon) (07/23/86)

Phone : (617) 937-0551 
Usenet: ...!{wanginst,masscomp}!ulowell!dobro 
Flame-To: /dev/null/... 
Reply-To: dobro@ulowell.UUCP (Gryphon) 
USMail: P.O.Box 8524, Lowell, Ma. 01853

From article #647 by rb@ccird1.UUCP (Rex Ballard) :
>What PSI research needs is more classification and organization of the
>factors that appear to effect results.  Unfortunately, attempts to
>do this to any significant degree tends to be treated as "bogus science".
>
>Are there any such "classification" efforts under way?

Local, there is one. MOD.PSI
It's purpose: to logically and scientifically come up with
defintions, explanations, limits, criteria, etc.

I wish more people were interested in thinking baout a topic as opposed
to talking about a topic (yes, I'm a fine one to talk, but...)

'Nough said.

						Gryphon


-------------------------------
Disclaimer: I am simply a figment of my immagination,
	and cannot therefore have and opinion.

"Far beyond human ability to classify, or comprehend..."