From: Damien Broderick (d.broderick@english.unimelb.edu.au)
Date: Fri Aug 06 1999 - 06:42:34 MDT
At 04:28 PM 5/08/99 -0700, Hal wrote:
>The latest meta-analysis (where
>multiple studies are combined to get more data) was not able to show
>a statistically significant effect, unlike an earlier meta-analysis.
>The controversy is now over which studies to include, which should not
>be included, why some studies show much higher results than others, etc.
>It is quite a mess.
>
>Online at:
>http://www.sciencenews.org/sn_arc99/7_31_99/fob4.htm
A mess, yes, but an interesting one. Consider this cite from the brief
report above:
========
The metanalysis has generated heated discussion among psychologists. Some
argue that Milton and Wiseman were unjustified in lumping all 30 studies
together because their results were so disparate. Milton contends that a
standard statistical test of variation among the results showed that they
could treat the studies as a uniform set.
Bem says, however, "The reason the effect isn't significant is that there
are three studies that are pulling down the average, and those studies are
very nonstandard." Further, 6 of the 30 studies showed significant psi
effects—more than would be expected by chance, he adds.
Since the metanalysis was completed, nine more ganzfeld studies have been
published. Milton acknowledges that the psi effect would be statistically
significant if the analysis were updated to include these studies. However,
she observes, a single study had an especially strong result, but no
clear-cut effect spans the broad range of investigations. "That will be a
crucial thing to demonstrate in order to make a strong claim that the
studies show a genuine anomaly," she says.
==========
This will strike obdurate skeptics as just so much waffling by the
anomalists - `very nonstandard' indeed, harrumph - but, hey... the bottom
line is, as admitted by the *skeptical* meta-analyst, that including *all*
the extant evidence gives a *positive*, pro-psi result.
This is not the sort of self-evidently bogus,
jeez-I-can-see-through-this-bullshit result that skeptics might expect if
psi were no better than iridology.
Damien Broderick
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 15:04:40 MST