PEAR's Global Conciousness hypothesis hit hard

[Delusions_O_Grandeur]

I am a strong believer in a greater mind and many other spiritual beliefs, but I'm NOT prepared to accept bad evidence for it. I'm afraid that I have to resign the GCP to my bin of bad evidence.



I summarized the GCP's hypothesis as: "Events with a worldwide emotional impact will cause random walks generated by Random Event Generators (REG's) to deviate significantly from what could be reasonably expected by chance"

In theory their method is good. They select time windows that they believe overlap with major global events like the 9/11 attacks or a new year's celebration. BEFORE accessing any data they specify a time window and a method of analysis for all the data contained within that time window. This will make any bias impossible. For each window they than calculate the probability that the observed deviation in generated random numbers happened by chance. The results from all windows are all combined using meta analysis techniques.

The chance that the combined deviation occurred by chance is like 10^-8.

Now, the skeptical part of me wants to answer two questions before I'd consider that statistic "extraordinary":

1) Do their analyses and meta analysis somehow cause nice results to emerge no matter which windows are selected?

2) What happens if I were to change the start/end time of each time window a little bit? If a period of abnormal global emotion causes a systematic deviation in series of random numbers this effect shouldn't just disappear into thin air if I change a window containing a global emotion persisting for hours.

Now, I found that the people of the GCP already tried to slide time windows around themselves... And the effect, is DISASTEROUS. If all time windows are sled 5 minutes or more into the past or future, the statistically significant deviation gained by 9 years of ongoing research, is OBLITHERATED. Given the fact that the start/end borders of each window were specified down to the quarter hour and intuitively to some degree, it just does not seem reasonable that the time windows have to be placed so precisely! The explanation they offer hurts my eyes. It’s the experimenter effect, the researcher's anomalous knowledge of exact time window needed to extract significantly deviant data from the database caused the significance. Well that could theoretically make sense, were it not for the 15 minute time resolution. I just don't buy it. For as far as I'm concerned the GCP has concluded absolutely nothing interesting besides the fact that their REG's *do* generate random numbers and that their method works fine in theory. I mean, their offset time window's DID give chance expectation.

So, with a heavy hart I admit that this piece of evidence for supernatural interhuman consciousness is not evidence at all. But I'm still going to believe stubbornly in the supernatural. I simply don't have what it takes to be a skeptic. Hate me. Ha.

 

[Jeff Corey]

I, for one, don't hate any sincere believers. Only the ones who harm others.
But, please use a spell check.

 

[Delusions_O_Grandeur]

Sorry. I was kind of raving there, I'm going over my post now to weed out spelling errors. I don't have an english spelling checker here though.

I'm uncertain where to turn next though. I've got this love/hate relationship with science really. It's kind of like this:


---
Science

It does not care about your emotions
It does not bend to your desires
But these are merciful notions
Because neither does the reality it inquires

Pseudoscience

It will shape reality in your dreams
You shall have the god your soul requires
But it does not prove to be what it seems
Because neither does the reality it inspires

-Delusions_O_Grandeur
---

 

[CFLarsen]

Shapes In The Clouds
A commentary on "Information and Uncertainty in Remote Perception Research", Dunne and Jahn, Princeton Engineering Anomalies Research (PEAR).

PEAR themselves admit - although buried - that there is no effect:

"Once again, there was reasonably good agreement among the six scoring recipes, but the overall results were now completely indistinguishable from chance."(PEAR, p227, Distributive Scoring)

Other articles:

An Evening With Dean Radin
I spent one of the last evenings of September 2002 attending a lecture by Dean Radin, author of "The Conscious Universe", on the Upper East Side, Manhattan, where he told about the Global Consciousness Project.

Book Review: The Conscious Universe, Dean Radin
The subject of the book is psi research, that is research concerning telepathy, clairvoyance and precognition. Radin claims that these phenomena are real and in the book he presents the evidence, which he thinks proves this.

 

[Delusions_O_Grandeur]

Though the article "Shapes in the clouds" is not about the GCP, it does show they have a knack for poor experimental protocols at PEAR. The picture coming into focus is that PEAR simply isn't an institution that I can trust to abstain from poor experimental design and data selection...

Based on other articles I'm reading here and there I've got the feeling that if you go to any parapsychology research unit and yell "Beware it's the null hypothesis!" people will jump up run around screaming and fly out of windows all over the place.

 

[Kopji]

So, with a heavy hart I admit that this piece of evidence for supernatural interhuman consciousness is not evidence at all. But I'm still going to believe stubbornly in the supernatural. I simply don't have what it takes to be a skeptic. Hate me. Ha.

The supernatural stuff is hard to let go of. A lot of things worth doing are hard. We don't give up on people quite so easily, you have the beginnings of a fine skeptic in you.

 

[T'ai Chi]

Now, I found that the people of the GCP already tried to slide time windows around themselves... And the effect, is DISASTEROUS. If all time windows are sled 5 minutes or more into the past or future, the statistically significant deviation gained by 9 years of ongoing research, is OBLITHERATED.

Your interpretation of what the sliding analysis shows is probably not correct.

Say we're measuring the temperature of a cup of coffee at time t. It is 80 degrees. And you say, well, if you shifted the time it was measured, the temperature would change, see, their claim that the cup of coffee is 80 degrees is flawed.

 

[CFLarsen]

Your interpretation of what the sliding analysis shows is probably not correct.

Say we're measuring the temperature of a cup of coffee at time t. It is 80 degrees. And you say, well, if you shifted the time it was measured, the temperature would change, see, their claim that the cup of coffee is 80 degrees is flawed.

Why change the time windows at all?

 

[blutoski]

Why change the time windows at all?

I think the value of tinkering with the windows is to see how sensitive the protocol is to framing.

To answer T'ai Ch'i's question: the point is that it's suspicious that an event like 9/11 produces "significance" if the window is from September 11th, 9:15am -> October 28th, 8:45pm, but does not produce significance if measured from September 11th, 9:15am -> October 28th, 8:30pm.

The suspicion raised is that the selection is not truly blinded: that it is the result of exploring interval options to find the desired result, then claiming they weren't exploring inverval options, but had selected intervals blindly. ie: it's evidence of deception.

 

[T'ai Chi]

The suspicion raised is that the selection is not truly blinded: that it is the result of exploring interval options to find the desired result, then claiming they weren't exploring inverval options, but had selected intervals blindly. ie: it's evidence of deception.

Well, it is your belief of evidence of deception, anyway.

But think about how stupid that sounds. They post a study on their own deception on the internet?

And think about the impossible heights that self proclaimed skeptics want to raise the bar to; make it so there is significance for all windows of time? After the significance tails off, no matter how long it takes for it to do so, you could just say 'aha, see, there is no longer significance'.

In short, you're the one doing the data mining- but you're just looking for non-signifigance.

 

[CFLarsen]

Well, it is your belief of evidence of deception, anyway.

No, there is evidence of deception. They don't use the same window every time, they vary it so it gives a result.

But think about how stupid that sounds. They post a study on their own deception on the internet?

As long as there are suckers who gobble it up, why not? Surely, you don't want to claim that there are no crooks on the Internet, do you?

And think about the impossible heights that self proclaimed skeptics want to raise the bar to; make it so there is significance for all windows of time?

Nobody says that.

After the significance tails off, no matter how long it takes for it to do so, you could just say 'aha, see, there is no longer significance'.

How long does it take before the significance tails off? Can you show that it is related to the event?

In short, you're the one doing the data mining- but you're just looking for non-signifigance.

No, the data mining is done by the believers, like yourself.

 

[Delusions_O_Grandeur]

Your interpretation of what the sliding analysis shows is probably not correct.

Say we're measuring the temperature of a cup of coffee at time t. It is 80 degrees. And you say, well, if you shifted the time it was measured, the temperature would change, see, their claim that the cup of coffee is 80 degrees is flawed.

I think you've missed the point a bit.

To use the coffee cup hypothesis: Say you're testing if the coffee cups temperature rises over time. If you pick a certain t1 and t2 10 hours apart, and between those two times the coffee cups temperature went up say, 20 degrees, than you'd say that would support the hypothesis. But if you take t1+5 minutes and t2+5 minutes and suddenly, the coffee cup cooled 3 degrees, then I'd say you can't possibly reliably say the coffee cup heats up over time but that the coffee temperature just behaves eratically. To tell if in certain time frames the coffee temperature goes up on average, you need to take a couple of thousand of windows *blindly*. I could get massive significane, thanks to the erratic behavior, by selecting my start time in a valley and my end time at a peak. But by selecting blindly, I prevent myself from doing this. Now, if in all these thousands of windows there is a very small average temperature rise, and I selected my start/end times using unclear criteria and don't care if I give or take 15 minutes, this effect should not go away if I shift my windows by 5 minutes. If the effect DOES go away, it suggests I picked some of those goodlooking and *very narrow* peaks and valleys now and then.

 

[Paul C. Anagnostopoulos]

Have they predicted anything yet, or is it all still retrodiction?

~~ Paul

 

[Squishua]

"Events with a worldwide emotional impact will cause random walks generated by Random Event Generators (REG's) to deviate significantly from what could be reasonably expected by chance"

A simple enough test could be devised to determine this.

Have multiple sets of random numbers generated by multiple random number generators during the time of a "worldwide emotional impant," as well as sets generated during other (randomly selected) time intervals. Of course there must be a double blind protocol for the actual identity of each set until the testing is concluded (post hoc predictions aren't very impressive).


Let the GCP perform their statistical analysis on each set of numbers to determine the deviation of their random walks and make their conclusions. When the identity of each set is revealed, do the target sets tend to have a greater deviation that the other sets? Are the results consistent?

That's all they'd have to do to convince me.

 

[T'ai Chi]

Have they predicted anything yet, or is it all still retrodiction?

I think they've predicted more significant deviations from chance than average for global events.

 

[T'ai Chi]

If the effect DOES go away, it suggests I picked some of those goodlooking and *very narrow* peaks and valleys now and then.

You've misunderstood the purpose of looking at various windows.

It is to see if the significance is robust to the choice of windows. It is not.

This is not the same as saying the windows were intentionally pre-selected to cause the significance as you attempt to imply.

 

[CFLarsen]

You've misunderstood the purpose of looking at various windows.

It is to see if the significance is robust to the choice of windows. It is not.

This is not the same as saying the windows were intentionally pre-selected to cause the significance as you attempt to imply.

They were not pre-selected at all, T'ai Chi. They were selected post-hoc. After they knew what they should look for.

 

[Delusions_O_Grandeur]

Indeed the experiment was done to find out if the effect was persistant when sliding the time frame a bit. However, I don't think they expected to get a peak of significane so dependant on the offset. They were at pains to explain it, and in case you didn't read it, they tried to explain it in terms of the experimenter effect, silently abandoning their original hypothesis. Now if the peak were say, an hour or 2 wide I'd find their original hypothesis still plausible.

I'm *implying* they selected their time frams. I'm not, however, saying thay actually did it. But the nice thing about science is that in the end it does not have to boil down to trust. Replication of the experiment with the given protocols will determine the ultimate fate of the hypothesis.

Oh, and please don't think I'm hell-bent on finding insignificance. I believe a global consciousnes exists, but this experiment just does not support it.

 

[Gord_in_Toronto]

Well, it is your belief of evidence of deception, anyway.

But think about how stupid that sounds. They post a study on their own deception on the internet?

And think about the impossible heights that self proclaimed skeptics want to raise the bar to; make it so there is significance for all windows of time? After the significance tails off, no matter how long it takes for it to do so, you could just say 'aha, see, there is no longer significance'.

In short, you're the one doing the data mining- but you're just looking for non-signifigance.

If you actually read the article An Evening with Dean Radin as identified by CFLarsen in a previous post, you will find this reported:

...during the days before the attacks, there were several instances of the eggs picking up data that showed the same fluctuation as on September 11th. When I asked Radin what had happened on those days, the answer was: "I don't know."

If there is any data mining and deception going on, I don't think it's on the skeptic's side.

 

[Paul C. Anagnostopoulos]

I think they've predicted more significant deviations from chance than average for global events.

I know that's what they say. Have they predicted specific times when this should occur and then analyzed the data to determine that it did occur?

~~ Paul