[extropy-chat] Movie: WHAT THE BLEEP DO WE KNOW!?

[Damien Broderick]

At 03:15 PM 11/25/2004 -0800, Adrian wrote:

>The vast majority of attempts to recreate these
>phenomena under conditions of scientific testing have
>failed.

What comprises a `vast majority'? More than half, clearly. More than 90%? 
But again, if the latter, this is simply inconsistent with the results 
found by a number of meta-analyses conducted by competent statisticians 
during the last decade or so. In general, the effect sizes remain stable, 
if small; the number of experiments that fail to repudiate the null is 
about in accordance with the number one must expect (due to noise, etc) in 
any ensemble of tests of a small-effect phenomenon. Read Prof Jessica Utts 
on this: http://anson.ucdavis.edu/~utts/
E.g.: http://anson.ucdavis.edu/%7Eutts/JSE1999.pdf

>Furthermore, given the many many experiments that
>have been conducted, it is expected that a few would
>show anomalous results, even just among the legit
>ones.

The well-known file-drawer hypothesis. This is an ongoing discussion in the 
scientific-paranormal community, but the consensus remains, I gather, that 
the number of competent experiments unpublished due to null results 
required to offset the data summarized in various meta-analyses is so 
absurdly large that the hypothesis is itself found to be extremely improbable.

Damien Broderick