From: Damien Broderick (thespike@satx.rr.com)
Date: Sun Jan 01 2006 - 11:32:09 MST
At 09:33 AM 1/1/2006 -0800, Justin wrote:
>If it's "largely" aversive reflex, why is it that attempted
>replications by other parapsychologists even, like of Schmidt's random
>number experiments, tend to fail?
>
>"The main fact that emerges from this data is that 71 experiments gave
>a result supporting Schmidt's findings and 261 experiments failed to
>do so" (Hansel 1989: 185).
I assume this is Hansel's *The Search for Psychic Power: ESP&
Parapsychology Revisited*. I haven't read that, but the earlier editions
were particularly grotesque in their misrepresentations and bizarre
attempts to offer possible fraudulent means for high-scoring (this sort of
thing, as it were: the experimenter might have left his room on the other
side of the University and run across while the subject was guessing and
very quietly with nobody noticing got a chair from an adjacent room and
climbed up on it to look through the transom using a mirror and written
down the subject's calls, then dashed back and copied them into his log as
if they had been the randomised targets). If I had my marked-up copy of the
1980 in edition with me, I'd provide some of the really laughable examples,
of which that invented instance was a close approximation. In general, as I
noted in my own book:
=======================
Although Schmidt has never been found cheating, he's been subjected to some
really quite venomous armchair attacks by such famous critics as Martin
Gardner and Professor C. E. M. Hansel. Certain features of Schmidt's
results, Hansel charged, `suggests that sampling from a common distribution
may have taken place', a phrase which seems to me a direct implication of
fraud. Interestingly, the same features-HI, LO and Baseline distributions
adding up to an overall Gaussian or standard bell-shaped curve-have since
turned up in Jahn's much larger body of evidence, and been incorporated in
the PEAR model of psi.
Hansel did not quite go so far as an accusation of fraud: `Perhaps what has
been written here will indicate reasons for questioning the proof supplied
by Schmidt without implying that he need necessarily have been more than a
careless experimenter...' All these accusations are indirect and
hypothetical; there is absolutely no reason to entertain them other than a
priori and inviolable skepticism itself, proof against every conceivable
counter-example.
============
But in fact 71 significant experiments out of 332 (21.4%) seems pretty
close to what one might expect for a low effect size phenomenon:
http://www.tcm.phy.cam.ac.uk/~bdj10/psi/delanoy/node7.html
<It should be noted that small effects have low statistical power [8, 12,
13]. For example, the aspirin study involved over 22,000 subjects. If there
had only been 3,000 subjects, the investigators would have had less than a
50 percent chance of finding a conventionally significant effect [13].
Given the small effect sizes which are typical in psi experiments, low
replicability is to be expected. Rosenthal [17] notes that ``even though
controversial research areas are characterised by small effects, that does
not mean that the effects are of no practical importance.'' (p. 324).
Indeed, in an article addressing behavioural research in general, Rosenthal
[35] warned: ``Given the levels of statistical power at which we normally
operate, we have no right to expect the proportion of significant results
that we typically do expect, even if in nature there is a very real and
very important effect'' (p. 16).>
Damien Broderick
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:54 MDT