Anyone heard this from PEAR?

[davidhorman]

Every couple of weeks I get a copy of ComputerWeekly sent to me, even though I've never subscribed and we tell them we don't want to renew our non-existent subscription.

Anyway, on the back page this week was this:

A question of mind over machine

Princeton University researches claim to have detected small but "statistically significant" signs that the human mind can interact with machines to affect their output. The Princeton Engineering Anomalies Research program studies the interaction of human consciousness with sensitive physical devices, systems and processes. Don't be surprised if Windows crashes more often now it knows what you think of it.

I don't remember PEAR finding anything statistically significant lately - is this just a bad journalist reporting old, wrong news?

David

[Zep]

They had a few column-inches to spare, and no inclination to do any checking of facts. Really, it sounds like a humour piece, not anything substantial.

And yes, PEAR claim to have found something but dodge and weave the issue completely when pressed for actual data. It's been discussed here ad nauseum - do a search if you are really interested: PEAR, EGGS.

[CFLarsen]

Or, naturally, read all about it here. (http://www.skepticreport.com/psychics/shapesintheclouds.htm) ;)

[Alkatran]

I don't even want to get into how badly computers would perform if we could change stored/displayed values while they were running.

[Bronze Dog]

Originally posted by Alkatran
I don't even want to get into how badly computers would perform if we could change stored/displayed values while they were running.
You mean there are computers that don't run badly? ;)

[Ersby]

from "On the false hypothesis of psi meditated shift of statistical average in tests with random number generators", Pallikari
"The micro-psychokinesis (micro-PK) hypothesis states that intention can mentally shift the statistical average of the outcomes of an inherently random process in the desired, pre-stated direction. The most recent meta-analysis based on all available to date micro-PK data, testing the direct influence of human intention on the outcomes oftrue random number generators (RNG), does not support this hypothesis. Furthermore, a very large-scaleexperiment set out by a consortium of independent research groups to replicate the micro-PK hypothesis, has failed to show the effect."

[Robin]

I made reference to PEAR here http://forums.randi.org/showthread.php?s=&threadid=58690 (in the science section, I was giving the benefit of the doubt), and quoted some of the excuses they give about replication, including this from http://www.princeton.edu/~pear/sos.pdf :
To begin with, it will be necessary to broaden quite generously the definition of experimental replicability to accommodate the hyper-statistical character of the interactive processes and their participants, the intrinsically elusive nature of many of the phenomena, and the subjective dimensions to which they relate.
A related site is the Global Consciousness Project which actually does make it's data available. They quote impressive statistical significance but this is based on varying the statistical procedure for each event they test.

[Zep]

Originally posted by Robin
A related site is the Global Consciousness Project which actually does make it\'s data available. Well, highly selective, pre-processed data, really.
Originally posted by Robin
They quote impressive statistical significance but this is based on varying the statistical procedure for each event they test. And by being highly selective and arbitrary about the range of data they examine.

[Robin]

Originally posted by Zep
Well, highly selective, pre-processed data, really.
Actually I think the database has raw data in it. For example a lot of the earlier stuff that claims to have data from 3 'eggs' actually has a most of the data missing (but nevertheless extravagant results claimed).
And by being highly selective and arbitrary about the range of data they examine.
Absolutely. I experimented on my test data and found I only had to tweak the range in this way for less than 10% of the events in order to create an impressive statistical significance.

[Ducky]

Originally posted by BronzeDog
You mean there are computers that don't run badly? ;)

I believe they're called "Macintosh."

Never had a problem with my g5 running protools.

[SmooveK]

My Mac ran like crap so I bought a new laptop a year ago and double installed Windows XP/SuSe. Never an error since except for when I dropped it.

Different strokes for different folks telling different jokes.