Scandalous Conduct in the Avoid-L Webforum


William Jefferys (renowned NASA astrophysicist) and Brent Meeker (unkown US Navy physicist ) direct lies and offenses towards respected parapsychological researcher Dean Radin and... chicken out when the time comes for them to prove their wild assertions.



Pseudoskepticism is always surprising beyond measure. Below is a series of 65 messages (plus a follow up) where I debated these two pseudoskeptics. Prepare to see the feathers all over while they are running from the debate...
(the title of the messages is in bold type, and the author of each message is in green bold type; beware that emails might have changed, just as my email has changed!)

1
Re: The Demon Comes... To discuss some"weaknesses"in"God, The Failed Hypothesis."
From:
"Julio Siqueira" <juliocbdsiqueira@YAHOO.COM>
To:
AVOID-L@HAWAII.EDU
 
Wednesday, March 21, 2007 6:04 AM
 
Lawrence, have you read "Entangled Minds" by Dean
Radin? (2006).

Julio
_________

--- "L. B. Crowell" <lcrowell@SWCP.COM> wrote:

> ----- Original Message -----
> From: Brent Meeker <meekerdb@DSLEXTREME.COM>
> To: <AVOID-L@HAWAII.EDU>
> Sent: Monday, March 19, 2007 6:12 PM
> Subject: Re: The Demon Comes... To discuss
> some"weaknesses"in"God, The Failed Hypothesis."
... ... ...
... ... ...
>
> No no no!!!  To use the language of fluctuations, to
> which is a source of umbrage, quantum fluctuations
> are Markovian.  This means that fluctuation "here"
> and "now" are completely independent of fluctuations
> "there" and "then."  This means that no information
> is communicated by quantum fluctuations.  So if
> there were some God who "buried" His influence in
> such fluctuations they would appear in the
> statistics.  The Princeton "Pear Lab" attempted to
> find such effects to no apparent success, and the
> lab has been closed. 
>

 
2
Re: The Demon Comes... To discuss some"weaknesses"in"God, The Failed Hypothesis."
 
Wednesday, March 21, 2007 6:39 AM
  From:
"Bill Jefferys" <bill@ASTRO.AS.UTEXAS.EDU>
  To:
AVOID-L@HAWAII.EDU
 
At 6:33 AM -0700 3/21/07, John Mazetier wrote:
>On 3/21/07, Julio Siqueira
><<mailto:juliocbdsiqueira@yahoo.com>juliocbdsiqueira@yahoo.com>
>wrote:
>
>>Lawrence, have you read "Entangled Minds" by Dean Radin? (2006).
>>
>>Julio
>>_________
>
>
>Jeeze, Julio...
>
>I wonder how many other old-timers 'round here spewed their morning
>coffee all over their monitors?
>
>Taking Dean seriously?  How about Fred Alan Wolf, or Amit Goswami,
>or perhaps Barney the dinosaur while we're at it?
>
>--J

Good thing I don't drink coffee.

Note to Julio: You have just greatly undermined your own credibility.

Bill

3
Re: The Demon Comes... To discuss some"weaknesses"in"God, The Failed Hypothesis."
 
Saturday, March 24, 2007 5:14 PM
  From:
"Julio Siqueira" <juliocbdsiqueira@YAHOO.COM>
To:
AVOID-L@HAWAII.EDU
 
Hi Bill, two things for now:
 
First, thank you for your feedbacks regarding the US Constitution (Preamble). Indeed I agree with you that the capital letters (in "Blessing") do not have the meaning (and thus the strength!) that I thought they had.
 
Second, my "own credibility" (just like yours) ought to be evaluated by my own works, my own reasonings, my own readings, and my analyses of these readings. Just spot any problem in these and I will either stand corrected (and thank those that corrected me; just like I thanked you) or debate until I can show that my point is correct (*if* I can show...).
 
And I can assure you that I will never give "five stars" (www.amazon.com) to a book that contains as many factual erros as I have showed GTFH to have... Not even "Entangled Minds", "Psi Wars", or "Immortal Remains" (despite being superb writings) reached the five-stars standard for me. Due to problems that I did spot and present in my reviews. (Can you find any review of GTFH by Avoid-L members that did not rate it as five stars? Talking about credibility).
 
But maybe we have oposing views regarding what friends are for...
 
Best Regards,
Julio
________________
 
 
4
Re: The Demon Comes... To discusssome"weaknesses"in"God, The Failed Hypothesis."
 
Saturday, March 24, 2007 6:31 PM
  From:
"Bill Jefferys" <bill@ASTRO.AS.UTEXAS.EDU>
To:
AVOID-L@HAWAII.EDU
 
At 5:37 PM -0700 3/24/07, Julio Siqueira wrote:
Hi Bill,
 
>I think you got it wrong. Indeed what Good said is all correct
>(technically) in his review. *However*, all the problems that he saw
>(i.e. all that he said) simply did not exist in Radin's work.
>Nowadays (and that means: from at least 1980 to now!) it is simply
>stupid to commit the kind of statistic mistakes that Good is
>attributing to Radin. You can find indeed serious statistical
>problems in works from Ian Stevenson's goup. But not from Radin's.

>Why on Earth do you think Nature would agree to publish Radin´s
>correction (with Good's acknowledgement of it!) if the Good's review
>were not shown (and proved!) to be wrong?
 
Common courtesy. It is done all the time.

>What do you think?
 
I know Jack Good personally and would trust his assessment over
yours. I have read Radin's technical papers (though not his popular
books, why should I bother with them after reading what they are
based on?) and find the same errors therein that Jack mentions.
 
This is my last comment on this topic. I haven't the time to waste.
 
Bill
 
PS Did you read my article?
 
5
Re: The Demon Comes... To discusssome"weaknesses"in"God, The Failed Hypothesis."
 
Wednesday, March 28, 2007 6:10 PM
  From:
"Julio Siqueira" <juliocbdsiqueira@YAHOO.COM>
To:
AVOID-L@HAWAII.EDU
 
Hi Bill,

You said:

"I know Jack Good personally and would trust his
assessment over yours."

Well, but that is not what you are doing. Good himself
acknowledged that he was wrong. Good said:

"Radin's claim of my "misunderstanding" relates to
four million guesses of cards made up to the year
1939. In his book (p. 97) he says that the combined
result amounts to "odds against chance" of more than a
billion trillion (10^21). Most readers would take this
as implying that the real odds were not more than a
trillion trillion (10^24) and would be surprised to
hear that by "more than a billion trillion" Radin
meant more than 10^100--increased to 10^2000 after the
book was published (information from Radin; Brian
Josephson, personal communication)"

Please understand, Bill, that I have no way to come to
any "assessment" on that. I know very well who Good
is. But you, who knows him far better than I,
disagrees on him (on Good). My question is a simple
one: why?

But, as it seems, you have no time to spend on that...

Thank you anyway

Best Regards,
Julio
_____________

6
Re: The Demon Comes... To discusssome"weaknesses"in"God, The Failed Hypothesis."
 
Wednesday, March 28, 2007 6:27 PM
  From:
"Brent Meeker" <meekerdb@DSLEXTREME.COM>
To:
AVOID-L@HAWAII.EDU
 
Julio Siqueira wrote:
> Hi Bill,
>
> You said:
>
> "I know Jack Good personally and would trust his
> assessment over yours."
>
> Well, but that is not what you are doing. Good himself
> acknowledged that he was wrong. Good said:
>
> "Radin's claim of my "misunderstanding" relates to
> four million guesses of cards made up to the year
> 1939. In his book (p. 97) he says that the combined
> result amounts to "odds against chance" of more than a
> billion trillion (10^21). Most readers would take this
> as implying that the real odds were not more than a
> trillion trillion (10^24) and would be surprised to
> hear that by "more than a billion trillion" Radin
> meant more than 10^100--increased to 10^2000 after the
> book was published (information from Radin; Brian
> Josephson, personal communication)"

This is not an acknowledgment of error by Good.  He's pointing out that Radin stated one basis for his conclusion in his book and when Good pointed out it was vulnerable to the file drawer effect, Radin changed what he claimed as a basis.

So Good's "error" was that he took something that Radin wrote at face value - an error you seem in danger of making also. 
Brent Meeker
 
7
Re: The Demon Comes... To discusssome"weaknesses"in"God, The Failed Hypothesis."
 
Wednesday, March 28, 2007 7:08 PM
  From:
"Bill Jefferys" <bill@ASTRO.AS.UTEXAS.EDU>
To:
AVOID-L@HAWAII.EDU
 
At 6:10 PM -0700 on 3/28/07, Julio Siqueira wrote:
>Hi Bill,
>
>You said:
>
>"I know Jack Good personally and would trust his
>assessment over yours."
>
>Well, but that is not what you are doing. Good himself
>acknowledged that he was wrong. Good said:
>
>"Radin's claim of my "misunderstanding" relates to
>four million guesses of cards made up to the year
>1939. In his book (p. 97) he says that the combined
>result amounts to "odds against chance" of more than a
>billion trillion (10^21). Most readers would take this
>as implying that the real odds were not more than a
>trillion trillion (10^24) and would be surprised to
>hear that by "more than a billion trillion" Radin
>meant more than 10^100--increased to 10^2000 after the
>book was published (information from Radin; Brian
>Josephson, personal communication)"
>
>Please understand, Bill, that I have no way to come to
>any "assessment" on that. I know very well who Good
>is. But you, who knows him far better than I,
>disagrees on him (on Good). My question is a simple
>one: why?

I don't see Jack saying anywhere in this that he made a mistake... In
fact, it appears from Jack's quote that in his opinion Radin is the
one that made the mistake. Jack puts the word 'misunderstanding' in
scare quotes, after all, and he does so for a reason. By the scare
quotes he is saying, in effect, that Radin *alleges* Jack's
"misunderstanding", but that the misunderstanding is not actual.

Perhaps you are not a native speaker of English and missed this
subtlety, but it is plain as day.

One of the errors that Radin makes constantly in the articles I have
read by him is to confuse p-values with odds. p-values are not odds
and cannot be converted to them.

>But, as it seems, you have no time to spend on that...
>
>Thank you anyway
>
>Best Regards,
>Julio
>_____________
 
8
Re: The Demon Comes... To discusssome"weaknesses"in"God, The Failed Hypothesis."
 
Wednesday, March 28, 2007 7:49 PM
  From:
"Julio Siqueira" <juliocbdsiqueira@YAHOO.COM>
To:
AVOID-L@HAWAII.EDU
 
Hi Brent,

I think that one important thing to do in science is
to seize the oportunity. It is ok that scientists are
not detectives or cops (not even the CSICOPS...). And
I do strongly critize the way things were carried out
in the Benveniste's issue (Nature) back in 1988. But
sometimes things just go on a way that fully justify
further investigation... And this reply from Radin is
one of such examples.

So, drafting a "timeline":

1- Radin writes a book to laymen.
2- Radin includes on it a number (p value; informal)
that perhaps can be used, by an independent
statistician, to recalculate the file-drawer.
3- The "independent" statistician (let's forget for a
moment his "slight" bias on the matter...) uses this
number to recalculate the file drawer, and finds
Radin's calculation to be terribly wrong.
4- Radin acknowledges that, *If The Number Good Used
was the precise one*, then Good's criticism would be
correct. However, so Radin says, he did not put in his
book the actual number (p value).

All this is perfectly ok so far. No one actually
deviated from good standards (I said "good standards",
not "Good's standards"). However... there are two
issues that require a further look from people like us
(like me and you!):

Issue Number one: Nature took very very long to allow
Radin's letter to be published. Odd indeed! Why was it
so?

Issue Number two: the number (p value) that Radin now
comes up with is astronomically smaller than the one
Good was lead to assume to be the true one... Of
course Good is not to blame for that. And actually
Radin is not to blame too. After all, it was a book
for laymen. But the question is: Did Radin really only
put things in a readable way in his book, or did he
actually forged the infinitesimal number (p value)
afterwards?

Now I ask you (and I am sure you understand statistics
astronomically more than I do...): is it really so
difficult for Good to get Radin's raw numbers to check
if his p value (the one in the personal joint
communication with Brian Josephson) was correct?

What do you think about it?

Julio
______________

9
Re: The Demon Comes... To discusssome"weaknesses"in"God, The Failed Hypothesis."
 
Wednesday, March 28, 2007 8:06 PM
  From:
"Julio Siqueira" <juliocbdsiqueira@YAHOO.COM>
To:
AVOID-L@HAWAII.EDU
 
Hi Bill,

You said:

"I don't see Jack saying anywhere in this that he made
a mistake... In fact, it appears from Jack's quote
that in his opinion Radin is the one that made the
mistake. Jack puts the word 'misunderstanding' in
scare quotes, after all, and he does so for a reason.
By the scare quotes he is saying, in effect, that
Radin *alleges* Jack's "misunderstanding", but that
the misunderstanding is not actual."

"Perhaps you are not a native speaker of English and
missed this subtlety, but it is plain as day."

Yes, I am not a native speaker of English and that
does indeed makes me sometimes rather slow to get some
subtleties. In this instance, however, I did see the
"plain as day" stuff. But I understood it more as
"die-hard syndrome" than statistical analysis.

I think that if Good really thought he was right, and
after all those very long months between his negative
review and Radin's answer (published), he could have
scrutinized the sources that Radin used to calculate
his p value, and therefore come and say: "Hey, Guys,
Radin is claiming now this new p value almost as small
as the planck lenght. However, his calculation of this
new p value is just a fake, and I can prove it plain
and simple."

So, what about it? Is Radin p value right or wrong
after all? (that is *the* information we need to bring
him down!).

You also say that Radin confuses p values with odds.
Have you contacted him about it? (He has a blog). Just
to check if he is indeed making a confusion or
something else.

Bye,
Julio
_________
 
10
Re: The Demon Comes... To discusssome"weaknesses"in"God, The Failed Hypothesis."
 
Wednesday, March 28, 2007 8:34 PM
  From:
"Brent Meeker" <meekerdb@DSLEXTREME.COM>
To:
AVOID-L@HAWAII.EDU
 
Julio Siqueira wrote:
> Hi Brent,
>
> I think that one important thing to do in science is
> to seize the oportunity. It is ok that scientists are
> not detectives or cops (not even the CSICOPS...). And
> I do strongly critize the way things were carried out
> in the Benveniste's issue (Nature) back in 1988. But
> sometimes things just go on a way that fully justify
> further investigation... And this reply from Radin is
> one of such examples.
>
> So, drafting a "timeline":
>
> 1- Radin writes a book to laymen.
> 2- Radin includes on it a number (p value; informal)
> that perhaps can be used, by an independent
> statistician, to recalculate the file-drawer.

p-values are not "informal".  Radin either got it wrong the first time or he got it wrong the second time.

> 3- The "independent" statistician (let's forget for a
> moment his "slight" bias on the matter...) uses this
> number to recalculate the file drawer, and finds
> Radin's calculation to be terribly wrong.
> 4- Radin acknowledges that, *If The Number Good Used
> was the precise one*,

Or if it had even been ball-park accurate - like within an order of magnitude.

> then Good's criticism would be
> correct. However, so Radin says, he did not put in his
> book the actual number (p value).
> All this is perfectly ok so far.

Why is it OK?

> No one actually
> deviated from good standards (I said "good standards",
> not "Good's standards"). However... there are two
> issues that require a further look from people like us
> (like me and you!):
>
> Issue Number one: Nature took very very long to allow
> Radin's letter to be published. Odd indeed! Why was it
> so?

Why don't you ask Nature.  Maybe it's because they don't trust Radin since he changed his story.

>
> Issue Number two: the number (p value) that Radin now
> comes up with is astronomically smaller than the one
> Good was lead to assume to be the true one... Of
> course Good is not to blame for that. And actually
> Radin is not to blame too. After all, it was a book
> for laymen.

That's a reason to put in an analogy, not a wrong number.

> But the question is: Did Radin really only
> put things in a readable way in his book, or did he
> actually forged the infinitesimal number (p value)
> afterwards?
>
> Now I ask you (and I am sure you understand statistics
> astronomically more than I do...): is it really so
> difficult for Good to get Radin's raw numbers to check
> if his p value (the one in the personal joint
> communication with Brian Josephson) was correct?

Getting the numbers from Radin wouldn't be a check.  If he lied about the p-value, he'd fabricate the data to back it up.  You'd have to take Radin's citations (assuming he has them) and check the sources himself - not so easy.  In any case I find Radin's claim of an enormous number of positive reports dubious

Brent Meeker

11
Re: The Demon Comes... To discusssome"weaknesses"in"God, The Failed Hypothesis."
 
Thursday, March 29, 2007 5:07 AM
  From:
"Bill Jefferys" <bill@ASTRO.AS.UTEXAS.EDU>
To:
AVOID-L@HAWAII.EDU
 
At 8:06 PM -0700 3/28/07, Julio Siqueira wrote:
>Hi Bill,
>
>You said:
>
>"I don't see Jack saying anywhere in this that he made
>a mistake... In fact, it appears from Jack's quote
>that in his opinion Radin is the one that made the
>mistake. Jack puts the word 'misunderstanding' in
>scare quotes, after all, and he does so for a reason.
>By the scare quotes he is saying, in effect, that
>Radin *alleges* Jack's "misunderstanding", but that
>the misunderstanding is not actual."
>
>"Perhaps you are not a native speaker of English and
>missed this subtlety, but it is plain as day."
>
>Yes, I am not a native speaker of English and that
>does indeed makes me sometimes rather slow to get some
>subtleties. In this instance, however, I did see the
>"plain as day" stuff. But I understood it more as
>"die-hard syndrome" than statistical analysis.
>
>I think that if Good really thought he was right, and
>after all those very long months between his negative
>review and Radin's answer (published), he could have
>scrutinized the sources that Radin used to calculate
>his p value, and therefore come and say: "Hey, Guys,
>Radin is claiming now this new p value almost as small
>as the planck lenght. However, his calculation of this
>new p value is just a fake, and I can prove it plain
>and simple."

I don't expect Good to do a huge investigation, based on what Radin
published (and then changed), on this subject. He pointed out a
problem, Radin agreed that Good had done so properly, Radin changed
his numbers, Radin's response was eventually published, and Good's
response to that was published, pointing out Radin's vacillation and
expressing his doubts.
>
>So, what about it? Is Radin p value right or wrong
>after all? (that is *the* information we need to bring
>him down!).

Who knows? He threw out a number, then when a problem was pointed
out, threw out another. This does not sound like responsible science.

>
>You also say that Radin confuses p values with odds.
>Have you contacted him about it? (He has a blog). Just
>to check if he is indeed making a confusion or
>something else.

I am a (skeptical, gadfly[*]) member of the editorial board of the
*Journal of Scientific Exploration*, in which many of Radin's
scientific papers have been published (along with a lot of other
parapsychology). Based on my experience in this position, I have
learned that parapsychologists in general, and Radin in particular,
imagine that p-values are the probabilities of something. They are
not. They don't obey the probability calculus (this is trivial to
show), therefore they cannot be used as probabilities in any
calculation.

I have no interest in pressing Radin on this point. I've already
pointed out in print the common misconception that parapsychologists
have on this (see letters to the editor of *JSE* on my website). I'm
simply making some observations, based on what I know, that may be
relevant to this discussion. Frankly, after all the
*shiHhHhHhnonsense* I have read in Radin's actual papers, anyone who
flogs his work as an example of good science and as good evidence in
favor of paranormal phenomena, automatically loses a significant
amount of credibility in my eyes. I don't base my comments on Radin's
popular works (which I haven't read, because his scientific papers
are so bad), but on his scientific papers.

This is my last comment on this subject. Wish me luck on my cataract
operation today.

Bill

[*] Some people have wondered why I serve on the editorial board of
*JSE*, and I have occasionally been criticized for it. I think that
at at least 90% of the things published in *JSE* are probably wrong,
but I also think that there is a place for a venue where people can
publish on "fringe" topics, since a very small fraction of them might
be worth pursuing and might even be right. Also, the journal does
publish articles critical of other articles that have appeared in it.
Several of the things I published in it were in this category (the
others were book reviews on rather prosaic and uncontroversial
topics). The critical articles I published in *JSE* can be found on
my website http://bayesrules.net/Papers.html/ .

12
Re: The Demon Comes... To discusssome"weaknesses"in"God, The Failed Hypothesis."

Thursday, March 29, 2007 6:46 AM
  From:
"Julio Siqueira" <juliocbdsiqueira@YAHOO.COM>
To:
AVOID-L@HAWAII.EDU
 
Hi Bill,

I will try to read your article as soon as possible,
and then I will try to exchange some viewpoints with
you regarding it and this "Radin Issue".

Your Article is at the link below, right?
http://bayesrules.net/papers/reg.pdf

I have gotten it now.

Bye,
Julio
____________________

"One of the errors that Radin makes constantly in the
articles I have read by him is to confuse p-values
with odds. p-values are not odds and cannot be
converted to them."


--- Bill Jefferys
 
13
Re: The Demon Comes... To discusssome"weaknesses"in"God, The Failed Hypothesis."
 
Friday, March 30, 2007 4:40 PM
  From:
"Julio Siqueira" <juliocbdsiqueira@YAHOO.COM>
To:
AVOID-L@HAWAII.EDU
 
Hi Brent,
 
I will try to get Radin's numbers and citations. I will try contacting him, Brian Josephson (who sent the "personal communication" together with Radin to Good), and one Brazilian psi researcher that I know that has some contact with Radin and with Carlos Alvarado. I think it is a very serious issue, and a very important one too. The least that I can say right now is that if you and Bill Jeffereys are right, then CSICOP is not doing their job correctly (at least right now, I find it very very hard to believe that they cannot, with some time and effort, follow this hot trail of fraud from the part of Radin, if it is truly fraud; the impact that it would have on parapsychology would be devastating).
 
I will report here the feedbacks (or absence of them).
 
Thank you for your comments!
Best Regards,
Julio
__________________
 
 
14
Re: The Demon Comes... To discusssome"weaknesses"in"God, The Failed Hypothesis."
 
Friday, March 30, 2007 4:55 PM
  From:
"Brent Meeker" <meekerdb@DSLEXTREME.COM>
To:
AVOID-L@HAWAII.EDU
 
Julio Siqueira wrote:
> Hi Brent,
>  I will try to get Radin's numbers and citations. I will try contacting him, Brian Josephson (who sent the "personal communication" together with Radin to Good), and one Brazilian psi researcher that I know that has some contact with Radin and with Carlos Alvarado. I think it is a very serious issue, and a very important one too. The least that I can say right now is that if you and Bill Jeffereys are right, then CSICOP is not doing their job correctly (at least right now, I find it very very hard to believe that they cannot, with some time and effort, follow this hot trail of fraud from the part of Radin, if it is truly fraud; the impact that it would have on parapsychology would be devastating).

One can dream.  A long history of fraud after fraud, from Margaret Fox to Florence Cook to Tamara Rand to Uri Geller to Ray Hyman, has had no apparent effect.

Brent Meeker

15
Re: The Demon Comes... To discusssome"weaknesses"in"God, The Failed Hypothesis."
 
Friday, March 30, 2007 5:13 PM
From:
"Julio Siqueira" <juliocbdsiqueira@YAHOO.COM>
To:
AVOID-L@HAWAII.EDU
 
Quite on the contrary! I think it is risky (and sloppy) to speculate quantitatively about the amount of fraud in psi research. However, the frauds that have been uncovered have had a very negative impact on this field, and have been a part of the resistance of science in general to accept this kind of investigation itself.
 
So, only sloppy skepticism (from the part of CSICOP et al) would justify such bizarre negligence on investigating it further and reporting it.
 
But if they are sloppy - and do not live up to their social duty - , I am not.
 
Julio
_______________
 
16
Re: The Demon Comes... To discusssome"weaknesses"in"God, The Failed Hypothesis."
 
Saturday, March 31, 2007 6:46 AM
  From:
"Julio Siqueira" <juliocbdsiqueira@YAHOO.COM>
To:
AVOID-L@HAWAII.EDU
 
Hi Bill,
 
I wish that you will see things better after your operation.
 
“This is my last comment on this subject.”
 
I’ve read it once, and I hope I will keep reading it again for sometime still :-)
 
“I don't expect Good to do a huge investigation, based on what Radin published (and then changed), on this subject. He pointed out a problem, Radin agreed that Good had done so properly, Radin changed his numbers, Radin's response was eventually published, and Good's response to that was published, pointing out Radin's vacillation and
expressing his doubts.”
 
For me, the most important thing in all this is to follow Radin’s mistakes (if they indeed exist) as far as they can lead me. If they lead me to the conclusion that Radin is a sloppy fraud, I will go that far. Nevertheless, I must point out that your paragraph above is far from correct. Since this is more of your concern than of mine, and since it seems that you are not that much interested in correcting what I see as problems (and incorrections) in your own approach to this subject (most certainly due to more important activities), I will not comment further on your deviations from effective standards (unless you, or someone else, ask so, obviously).
 
“Who knows? He threw out a number, then when a problem was pointed
out, threw out another. This does not sound like responsible science.”
 
I think that we all must be engaged in responsible criticism too. Your opinions (i.e. evaluations) regarding Radin may be correct. I will have to delve much further into that to come to a sound and responsible conclusion. Actually your evaluations simply must be correct, given your knowledge and expertise on the subject. At this time, I cannot “take sides.”
 
Let’s see what comes out of this...
 
Best Regards, and I wish you full success in your eye operation.
Julio
_______
 
17
Dean Radin: a Sloppy Fraud? Or is “someone else” to blame?...
 
Saturday, March 31, 2007 4:18 PM
  From:
"Julio Siqueira" <juliocbdsiqueira@YAHOO.COM>
To:
AVOID-L@hawaii.edu
 
Hi Everybody,
 
I am trying to gather as much information as possible (and it is actually very little!) before I try to get any feedback from Dean Radin or Brian Josephson. So I took a look at the pages in Josephson’s site that describe this weird issue between Radin and Nature-Good. The way it is presented, it is very clear that Good-Nature is wrong, and Radin is right. And this is so even in light of all that has been said here by members of the Avoid-L, like Bill Jefferys, Brent Meeker. So please note that I am building upon the feedbacks provided by these very members that I just cited.
 
I will present below a brief series of extracts that portrays the whole issue. Each part will consist of a link (url), followed by the relevant brief extract, followed by my brief comments where I show the apparent inconsistencies of the comments from Brent Meeker and Bill Jefferys.
 
 
Extract # 1:
 
http://members.cruzio.com/~quanta/review.html
 
This below is from I. J. Good’s review (Nature 1996) of Radin’s book:
 
Consider the following typical example. Radin points out that there were 186 publications on ESP card tests worldwide from 1882 to 1939. "The combined results of this four-million trial database [taken at face value]," he says, "translate into tremendous odds against chance--more than a billion trillion to one." (A 'trial' is the guess of one card) He means that the P value is about 10^-21--he is not writing only for the scientific establishment. This P value corresponds to a bulge above 'chance' expectation of 9.5 sigma, where sigma is the standard deviation. (I call that a 'sigmage' of 9.5.)
 
Apart from the possibility of conscious and unconscious fraud and wishful thinking in some fraction of the publications, Radin claims, with no explanation, that, in order to "nullify" the statistical significance, the file drawer would have to contain "more than 3,300 unpublished, unsuccessful reports for each published report". That number 3,300 is a gross overestimate. It should be reduced at least to about 15 (or even to 8).
 
My comment: Jefferys stressed that p values and odds are quite different stuff. However (and I am not doubting Jefferys at all), it seems that it was *Good* that decided to translate these odds into a p value, and not Radin. Note that Radin uses the expression: “taken at face value.” So, IMO, he seems to be open to the possibility of fraud in this database (therefore rendering incorrect and unfair what has been said of him here by some list members...).
 
 
Extract # 2:
 
http://www.tcm.phy.cam.ac.uk/~bdj10/psi/doubtsregood.html
 
This below is a letter from Brian Josephson to Nature, where he explains (or “explains”, if you prefer) the situation (followed by a “clarification” in his webpage).
 
Original letter, sent to Nature Correspondence by Brian Josephson (Nov. 13th., 1997):
 
Again, for the general reader's benefit, he described a P-value which was actually of order 1 in 102000 simply as 'odds of more than a billion trillion to one against chance', so that the latter number does not approximate to the actual P-value as Good assumed.
 
Clarification: the P-value discussed enters into the calculation of the so-called 'file-drawer factor', which indicates how many unpublished studies there would need to be for each one that was published for the significance of the results to fall to chance. With Good's incorrect P-value, the file-drawer factor is around 16 (low enough for it to be not totally unreasonable to argue that this accounts for the P-value calculated on the basis of the published results); with the correct value, derived from the reference that Radin quotes, it is around 3300, which is much harder to argue away, especially in view of the vast total amount of experimentation that would have to be postulated for such an explanation to work. Note that Good chose 1% as the point where a result is significant while Radin uses 5%, which difference accounts for only a small part of the difference between the two quoted file drawer factors.
 
My Comment: it seems that all that is needed is to check for this 1 in 102000 stuff. I have already said that, but now there is more to it (i.e. more strength to this suggestion from me), below...
 
 
Extract # 3:
 
http://www.tcm.phy.cam.ac.uk/~bdj10/psi/doubtsregood.html
 
This below is the letter from Radin to Nature.
 
There follows the letter sent subsequently by Radin (17 Jan 1998):
 
In his review (Oct. 23rd, p.806) of my book The Conscious Universe (1997, HarperEdge), I.J. Good suggests, based on certain misunderstandings on his part, that something "must be" wrong with my statistical arguments in favor of the reality of psychic phenomena. His inability to reproduce my estimate that 3,300 unpublished, unsuccessful experiments would be required for each published ESP card experiment (to nullify the cumulative outcome) follows from his incorrect assumption that my words "more than", used in connection with the cumulative odds against chance for the 186 experiments listed in Pratt et al (1966, "ESP after 60 years", as noted in my book), could safely be replaced by "approximately equal to." The actual p-value is approximately 10-2000, to which application of standard methods (Rosenthal, 1991, "Meta-analytic procedures for social research") gives my reported figure.
 
My Comment: this above is what, as far as I could understand, almost shocked me, in light of what Bill Jefferys said (or was it Brent Meeker?). As it turns out, the source for this number used by Radin is just one book! Can I conclude that it is easier for avoid-l skeptics to call someone a liar than to check this one book for the accuracy of what Radin claims? This book by Pratt and Rhine has (I think) less than 500 pages! (Extra-Sensory Perception after Sixty Years. By J. B. Rhine, J. G. Pratt, C. E. Stuart, B. M. Smith, and J. A. Greenwood, New York, Henry Holt & Co., 1940. Pp. xiv, 463.).
 
 
Extract # 4:
 
http://members.cruzio.com/~quanta/badgood.html
 
Below we see Good’s reply:
 
I.J. Good replies—
 
Radin's claim of my "misunderstanding" relates to four million guesses of cards made up to the year 1939. In his book (p. 97) he says that the combined result amounts to "odds against chance" of more than a billion trillion (10^21). Most readers would take this as implying that the real odds were not more than a trillion trillion (10^24) and would be surprised to hear that by "more than a billion trillion" Radin meant more than 10^100--increased to 10^2000 after the book was published (information from Radin; Brian Josephson, personal communication)
 
My final comment for this message: as far as I have been able to see so far, based on all that I have gathered and based on the feedbacks from Avoid-L members, it seems that there has been some “deviation from good science” in the way some people have commented on this issue here. I will wait for any comments (and preferebly for comments from Bill Jefferys too, which may take some weeks unfortunately) before I forward this info to Radin and Josephson.
 
Best Regards to All,
Julio Siqueira
___________
 
18
Re: Dean Radin: a Sloppy Fraud? Or is � to blame ?...
 
Saturday, March 31, 2007 5:38 PM
  From:
"Brent Meeker" <meekerdb@DSLEXTREME.COM>
To:
AVOID-L@hawaii.edu
 
Julio Siqueira wrote:
...
> *My Comment:* this above is what, as far as I could understand, almost shocked me, in light of what Bill Jefferys said (or was it Brent Meeker?). As it turns out, the source for this number used by Radin is just one book! Can I conclude that it is easier for avoid-l skeptics to call someone a liar than to check this one book for the accuracy of what Radin claims? This book by Pratt and Rhine has (I think) less than 500 pages! (/Extra-Sensory Perception after Sixty Years. By J. B. Rhine, J. G. Pratt, C. E. Stuart, B. M. Smith, and J. A. Greenwood, New York, Henry Holt & Co., 1940. Pp. xiv, 463./).

Interesting that data in a sixty year old book which is supposedly at the foundation of a whole field of research is not available in the public domain.  Why don't you just buy the book and tell the list exactly what these results are that Radin says imply odds of 10^21 or 10^2000?

You should also remember that the question is not simply one of doing the statistics correctly.  There were many procedural problems with Rhine's experiments and as he responded to criticisms and tightened controls the effect steadily declined.  One wonders how Radin factored this into his odds.

Brent Meeker

19
Re: Dean Radin: a Sloppy Fraud? Or is � to blame ?...
 
Tuesday, April 3, 2007 10:12 AM
  From:
"Julio Siqueira" <juliocbdsiqueira@YAHOO.COM>
To:
AVOID-L@HAWAII.EDU
 
I will ask Radin and Josephson.

Brent Meeker <meekerdb@DSLEXTREME.COM> wrote:
Julio Siqueira wrote:
...
> *My Comment:* this above is what, as far as I could understand, almost
> shocked me, in light of what Bill Jefferys said (or was it Brent
> Meeker?). As it turns out, the source for this number used by Radin is
> just one book! Can I conclude that it is easier for avoid-l skeptics to
> call someone a liar than to check this one book for the accuracy of what
> Radin claims? This book by Pratt and Rhine has (I think) less than 500
> pages! (/Extra-Sensory Perception after Sixty Years. By J. B. Rhine, J.
> G. Pratt, C. E. Stuart, B. M. Smith, and J. A. Greenwood, New York,
> Henry Holt & Co., 1940. Pp. xiv, 463./).

Interesting that data in a sixty year old book which is supposedly at the foundation of a whole field of research is not available in the public domain. Why don't you just buy the book and tell the list exactly what these results are that Radin says imply odds of 10^21 or 10^2000?

You should also remember that the question is not simply one of doing the statistics correctly. There were many procedural problems with Rhine's experiments and as he responded to criticisms and tightened controls the effect steadily declined. One wonders how Radin factored this into his odds.

Brent Meeker
 
 
20
Re: Fw: Dean Radin: a Sloppy Fraud? Or is "someone else" to blame?...
 
Tuesday, April 3, 2007 1:05 PM
  From:
"Julio Siqueira" <juliocbdsiqueira@YAHOO.COM>
To:
AVOID-L@HAWAII.EDU
 
Now that I have enough feedbacks, I will proceed analysing this issue further (trying to contact Radin/Josephson, etc). As I can see, Jefferys seems to agree that I. J. Good did the very same thing that he (Jefferys) condemned in Radin, that is, converting odds into p-value.
 
Of course we must all bear in mind that if we find out that Radin was a hundred percent correct in this controversy Radin vs Good, this does not mean that Radin's scientific papers (and research) is of good quality. Bill Jefferys has raised strong objections to the quality of Radin's works, on the grounds of statistics mainly. This cannot be downplayed, and asks for a reply from Radin as well.
 
I will be trying to contact, thus, Dean Radin, and I understand it is ok if I pass to him what is written down below in this message (a full message from Jefferys presenting what I see as sound and constructive criticism to psi research) and also what I wrote above. If I understood it wrongly, I kindly ask Professor Jefferys to indicate me so (I will wait some days if need be).
 
Best Regards,
Julio Siqueira
_______________
 
21
Re: Fw: Dean Radin: a Sloppy Fraud? Or is "someone else" to blame?...
 
Tuesday, April 3, 2007 1:42 PM
  From:
"Bill Jefferys" <bill@ASTRO.AS.UTEXAS.EDU>
To:
AVOID-L@HAWAII.EDU
 
At 1:05 PM -0700 4/3/07, Julio Siqueira wrote:
>
>Now that I have enough feedbacks, I will proceed analysing this
>issue further (trying to contact Radin/Josephson, etc). As I can
>see, Jefferys seems to agree that I. J. Good did the very same thing
>that he (Jefferys) condemned in Radin, that is, converting odds into
>p-value.
>
>Of course we must all bear in mind that if we find out that Radin
>was a hundred percent correct in this controversy Radin vs Good,
>this does not mean that Radin's scientific papers (and research) is
>of good quality. Bill Jefferys has raised strong objections to the
>quality of Radin's works, on the grounds of statistics mainly. This
>cannot be downplayed, and asks for a reply from Radin as well.
>
>I will be trying to contact, thus, Dean Radin, and I understand it
>is ok if I pass to him what is written down below in this message (a
>full message from Jefferys presenting what I see as sound and
>constructive criticism to psi research) and also what I wrote above.
>If I understood it wrongly, I kindly ask Professor Jefferys to
>indicate me so (I will wait some days if need be).
>
>Best Regards,
>Julio Siqueira
 
Since you have accused Good of error, don't you think you should get
his opinion before proceeding further? I don't think he has an email
address, but he has a web page with contact information.
 
        http://www.stat.vt.edu/facstaff/ijgood.html
 
Good has proposed techniques for estimating odds from p-values *and
the number of data points in the study*. The second part of this is
crucial. p-values cannot be turned into odds ratios by themselves,
but with this additional piece of information odds ratios can be
estimated, if roughly. I'll bet that this is what Good was talking
about.
 
Bill
 
22
Re: Fw: Dean Radin: a Sloppy Fraud? Or is "someone else" to blame?...
 
Tuesday, April 3, 2007 3:17 PM
  From:
"Julio Siqueira" <juliocbdsiqueira@YAHOO.COM>
To:
AVOID-L@HAWAII.EDU
 
I haven't really accused Good of error. If (If!) Josephson and Radin are not lying, what happened was that Radin used a term that led Good to conclude something wrongly. I still think that both parties acted in good faith (though each one with his own bias). Since Good did not challenge how Radin came up with the 10-2000 stuff (or something tinier), and since we here have challenged this, then I think this is where a problem may be found. I may be a believer. But do not expect me to believe that I will ever find a statistical mistake in Good's works (or in yours...) ;-) . (by the way: as I now understand from this message of yours, Good may have converted odds into p values in an acceptable manner; so I cannot say anymore that there is a conflict between what you said and what he said).
 
Thank you for the feedbacks,
Julio
P.S.: I might still try to contact Good, since right now I am interested in these weaknesses of psi research that you brought up.
______________
 
23
To Bill Jefferys, Regarding Dean Radin
 
Saturday, April 7, 2007 11:33 AM
  From:
"Julio Siqueira" <juliocbdsiqueira@YAHOO.COM>
To:
AVOID-L@HAWAII.EDU
 
Hi Bill,

I am directing this question to you specifically, but
if someone else has any information that might be
useful in this context, I will appreciate additional
replies.

As maybe you have already seen, I sent an emal to Dean
Radin and to Brian Josephson concerning the Good
issue. I sent it to Radin's email and posted it too in
Radin's blog, so I hope he will reply. But as I have
already said, even if he was indeed right and have
acted in good faith (and Good too), that does not
necessarily imply that his works (Radin's) are of good
quality. You said that many of his works have several
problems, especially (I think) in terms of statistics.


I would like, in the very near future (right after he
replies to me), to check for the quality of these
works in light of what you said. To do this, I would
like to ask you if you could indicate to me, say,
three papers from him that have the worst quality or
something like it. It is best if they are recent ones
(i.e. from 1996 on). Also, if you could indicate what
is the problem in any given paper, that would help
enormously.

I'd like just to add that even in my favourable review
of Radin's "Entangled Minds", I pointed out mistakes
from him.
http://paginas.terra.com.br/educacao/criticandokardec/amazon_reviews.htm#radin

Thank you very much, and Best Regards,

Julio Siqueira
____________________
24
Re: To Bill Jefferys, Regarding Dean Radin
 
Saturday, April 7, 2007 12:40 PM
  From:
"Julio Siqueira" <juliocbdsiqueira@YAHOO.COM>
To:
AVOID-L@HAWAII.EDU
 
This is just to inform that Dean Radin has published
my email in his blog:
https://www2.blogger.com/comment.g?blogID=16158865&postID=5786961001437193722

Best Regards,
Julio
___________

25
Re: To Bill Jefferys, Regarding Dean Radin
 
Saturday, April 7, 2007 6:09 PM
  From:
"Julio Siqueira" <juliocbdsiqueira@YAHOO.COM>
To:
AVOID-L@HAWAII.EDU
 
Hi Bill,
 
When I last replied to you I had not seen that Radin had already commented on my message. I did not imagine that he would be so prompt. I post now below his reply that is already there. From it, the key point seems to be this: "An unweighted Stouffer Z score of the studies provided there results in z = 96.5. (A z weighted by the number of trials per study is z = 165.0.)."
 
What do you (and I ask this to Bill Jefferys and also to Brent Meeker and Lawrence Crowel, all who, if my memory serves me well, raised doubts concerning Radin's stand in this issue) think of this number? (for example, does it seem to be an impossible number at first sight?) Right now what I will ask him is what he thinks of the possibility of fraud and methodological errors in this database (from 1880 to 1940). I pointed out to you that it seems that he used the expression "taken at face value", which may indicate a word of caution from him by using it. I think the best thing to ask after it is what results entered as input for the calculation that gave the final z value of 96.5. That may be troubling (i.e. rather painstaking) for him to do. Maybe my contacts here in Brazil have this book from Pratt et al. If need be, I will end up buying it (not so much expensive; but difficult for me to buy it here in Brazil; www.amazon.com does not sell it - of course I know that many list members here most likely do have access to this book in US university libraries...).
 
So, let's how far we can get into this...
 
Best Regards,
Julio
_________________
 
 
The book Extra-sensory perception after sixty years lists all known ESP card tests conducted from about 1880 to the publication of that book in 1940. An unweighted Stouffer Z score of the studies provided there results in z = 96.5. (A z weighted by the number of trials per study is z = 165.0.) I asked a mathematician friend to estimate the actual p pvalue associated with z = 96 and his response was less than 10e-2000. Not knowing the words associated with an odds value associated with that p value, I simply wrote "more than a billion trillion to one," not thinking that anyone would be so rash as to convert that into an exact figure.

I find it disappointing, but not especially surprising, that skeptics find my or my colleagues' studies suspicious. The suspicion arises because they haven't read the actual experimental reports. One of the reasons I wrote The Conscious Universe and Entangled Minds is to help inform scientists that what they think they know about this discipline is, in most cases, demonstrably incorrect.

That people can be highly opinionated without benefit of knowledge is not new, but one hopes that scientists aspire to overcome such unfortunate biases.

I challenge any scientist used to working in conventional areas to seriously investigate an unorthodox topic. If you obtain results that run counter to convention and publish your results, watch what happens to your perceived credibility. It's not a pretty picture.
Saturday, April 07, 2007
 
 
26
update on Dean Radin

Wednesday, April 11, 2007 7:22 AM
  From:
"Julio Siqueira" <juliocbdsiqueira@YAHOO.COM>
To:
AVOID-L@HAWAII.EDU
 
While some here prefer to take the "rational way", and
despise people without proper investigation, I decided
to take the "blind-believer way", and investigate if
Radin's number (10^-2000) was correct (in The
Conscious Universe).

He has answered me fully on his blog, in the thread
"More Lunacy. Not."
http://deanradin.blogspot.com/2007/04/more-lunacy-not.html

Interested readers can check out the current status of
the debate. Radin has been quite supportive of my
re-analysis of his numbers. That is truly the sign of
fraud, isn't it?

Anyway, my next step will be to get the book (Pratt et
all, year 1940) to get the data and put it to the test
(kids job...).

Let's see what the numbers will reveal...

Best Regards,
Julio
_____________

27
Re: update on Dean Radin
 
Wednesday, April 11, 2007 7:39 AM
  From:
"Bill Jefferys" <bill@ASTRO.AS.UTEXAS.EDU>
To:
AVOID-L@HAWAII.EDU
 
I don't see any substantive answer at all on Radin's blog. I do see a
claim that z=96.5, but no supporting evidence. Radin says, "read his
book," but the book is not peer-reviewed, was written for a popular
audience, and is thus useless from my point of view. As I pointed out
earlier, it is true that z=96.5 translates to p<1/2000, it's a
one-line calculation in R [and parenthetically, why did Radin, a
supposed expert on statistics, have to go to someone else to
calculate this simple number?], but as I also stated, it is not clear
that this is of any interest (for many reasons, not the least of
which being that we have no idea how he got z=96.5, what kind of
evidence it is based on, how well the experiments were controlled,
how we can be sure of this, what p-values mean and *don't* mean,
etc., etc.)

Bill

28
Re: update on Dean Radin
 
Wednesday, April 11, 2007 10:19 AM
  From:
"Brent Meeker" <meekerdb@DSLEXTREME.COM>
To:
AVOID-L@HAWAII.EDU
 
Bill Jefferys wrote:
> I don't see any substantive answer at all on Radin's blog. I do see a
> claim that z=96.5, but no supporting evidence. Radin says, "read his
> book," but the book is not peer-reviewed, was written for a popular
> audience, and is thus useless from my point of view. As I pointed out
> earlier, it is true that z=96.5 translates to p>1/2000, it's a
> one-line calculation in R [and parenthetically, why did Radin, a
> supposed expert on statistics, have to go to someone else to
> calculate this simple number?], but as I also stated, it is not clear
> that this is of any interest (for many reasons, not the least of
> which being that we have no idea how he got z=96.5, what kind of
> evidence it is based on, how well the experiments were controlled,
> how we can be sure of this, what p-values mean and *don't* mean,
> etc., etc.)
>
> Bill

He says he calculated the "Stouffer Z" by adding the z-values of the 186 experiments and dividing by the square root of the number.  I've never heard of "Stouffer Z", but it sounds similar to the practice of combining  Gaussians by averaging the variances.  Except in that case you would take the root-sum-square of the z's. 

Brent Meeker

29
Re: update on Dean Radin
 
Wednesday, April 11, 2007 10:55 AM
  From:
"Bill Jefferys" <bill@ASTRO.AS.UTEXAS.EDU>
To:
AVOID-L@HAWAII.EDU
 
At 10:19 AM -0700 on 4/11/07, Brent Meeker wrote:
>Bill Jefferys wrote:
>>  I don't see any substantive answer at all on Radin's blog. I do see a
>>  claim that z=96.5, but no supporting evidence. Radin says, "read his
>>  book," but the book is not peer-reviewed, was written for a popular
>>  audience, and is thus useless from my point of view. As I pointed out
>>  earlier, it is true that z=96.5 translates to p>1/2000, it's a
>>  one-line calculation in R [and parenthetically, why did Radin, a
>>  supposed expert on statistics, have to go to someone else to
>>  calculate this simple number?], but as I also stated, it is not clear
>>  that this is of any interest (for many reasons, not the least of
>>  which being that we have no idea how he got z=96.5, what kind of
>>  evidence it is based on, how well the experiments were controlled,
>>  how we can be sure of this, what p-values mean and *don't* mean,
>>  etc., etc.)
>>  Bill
>
>He says he calculated the "Stouffer Z" by adding the z-values of the
>186 experiments and dividing by the square root of the number.  I've
>never heard of "Stouffer Z", but it sounds similar to the practice
>of combining  Gaussians by averaging the variances.  Except in that
>case you would take the root-sum-square of the z's.
>Brent Meeker

Yeah, that's what he says. Basically he's approximating the z-score
of a combined experiment by assuming they are iid draws from a
population with a constant variance. This isn't quite right, if the
experiments have different variances, but it's probably OK if the
number of trials per experiment doesn't vary much.

My point is that we don't know anything about the experiments. We
don't know who did them, what controls they had, whether cheating
might have been involved, etc. We know that controls in
parapsychology experiments have often been lax, we can point to
specific examples of cheating, etc. Not knowing anything about the
186 experiments that went into this number, it is impossible say
anything reasonable. Radin, no doubt, will assure us that these
experiments have been of the highest quality, that there is no
possibility of spurious results, etc., but should I believe him? Why?
He gives me no reasons.

Furthermore, there is the "filedrawer" effect. Radin mentions standard
discussions (Rosenthal) to the effect that an enormous number of
experiments would have to be in the filedrawer to explain such a huge
effect. But Scargle has shown that the standard discussion is way too
overoptimistic (I posted the reference here a week or so ago).

Finally, as you & I have discussed offline, observed p-values don't
mean what many people think they mean. They aren't a Type I error
rate probability, they aren't "the probability that the results would
be obtained by chance," they aren't "the probability that the null
hypothesis is true," and they do not have a ready interpretation as
probabilities in terms of the experiment being conducted.

Laplace noted that those who make recitals of miracles "decrease
rather than augment the belief which they wish to inspire; for then
those recitals render very probable the error or falsehood of their
authors. But that which diminishes the belief of educated men often
increases that of the uneducated, always avid for the marvelous."
Given the bad track record of parapsychological experiments, when one
quotes a z-score of 96.5, for a proposition that is highly dubious in
the first place, such a huge z-score only reinforces my opinion that
it is probably due to spurious effects of one sort or another (or
combined). Such a number doesn't increase my confidence that psi is
real.

I would think (a point I have made in print), that if psi can
*really* produce *96-sigma* results, it ought to be powerful enough
that we could construct an experiment that doesn't depend on fancy
statistics and a huge number of trials to produce results. If, for
example, students at Princeton can "influence" the output of a
radioactive random event generator with a tiny p-value, someone
should be able to construct an experiment where a student can move,
say, a single Xenon atom at will around on a silicon substrate. Such
things are routinely done with atomic force microscopes, and the
motion of the atom could easily be detected. Now *that* is something
that would be worthwhile looking at, and would be much more
convincing than any number of statistical experiments of the sort
that parapsychologists tout as "proof". It would also be much more
easily reproducible by others.

So, my comment stands. Yes, I know how Radin calculated the number
(but still I don't know why it is that he couldn't convert a z-score
into a p-value without help!) But I know nothing about the provenance
of those 186 experiments, how they were selected, who did them, how
reliable they are, what experiments were left out or not known about,
etc. So the number is useless, meaningless, and Radin hasn't helped
in his blog entry. His comments were not sufficiently substantive.

Bill

30
Re: update on Dean Radin
 
Wednesday, April 11, 2007 11:58 AM
  From:
"Vic Stenger" <vstenger@MINDSPRING.COM>
To:
AVOID-L@HAWAII.EDU
 
Bill, you and I have been researchers for a long time. Have you ever witnessed a case where a major discovery was made either by metanalysis of several experiments or using sophisticated statistical techniques on a single experiment when the data did not hit you square in the eye?

Vic

31
Re: update on Dean Radin
 
Wednesday, April 11, 2007 12:38 PM
  From:
"Bill Jefferys" <bill@ASTRO.AS.UTEXAS.EDU>
To:
AVOID-L@HAWAII.EDU
 
At 11:20 AM -0700 4/11/07, Brent Meeker wrote:
>Bill Jefferys wrote:
>>  At 10:19 AM -0700 on 4/11/07, Brent Meeker wrote:
>>>  Bill Jefferys wrote:
>>>>   I don't see any substantive answer at all on Radin's blog. I do see a
>>>>   claim that z=96.5, but no supporting evidence. Radin says, "read his
>>>>   book," but the book is not peer-reviewed, was written for a popular
>>>>   audience, and is thus useless from my point of view. As I pointed out
>>>>   earlier, it is true that z=96.5 translates to p>1/2000, it's a
>>>>   one-line calculation in R [and parenthetically, why did Radin, a
>>>>   supposed expert on statistics, have to go to someone else to
>>>>   calculate this simple number?], but as I also stated, it is not clear
>>>>   that this is of any interest (for many reasons, not the least of
>>>>   which being that we have no idea how he got z=96.5, what kind of
>>>>   evidence it is based on, how well the experiments were controlled,
>>>>   how we can be sure of this, what p-values mean and *don't* mean,
>>>>   etc., etc.)
>>>>   Bill
>>>
>>>  He says he calculated the "Stouffer Z" by adding the z-values of the
>>>  186 experiments and dividing by the square root of the number.  I've
>>>  never heard of "Stouffer Z", but it sounds similar to the practice
>>>  of combining  Gaussians by averaging the variances.  Except in that
>>>  case you would take the root-sum-square of the z's.
>>>  Brent Meeker
>>  Yeah, that's what he says. Basically he's approximating the z-score
>>  of a combined experiment by assuming they are iid draws from a
>>  population with a constant variance. This isn't quite right, if the
>>  experiments have different variances, but it's probably OK if the
>>  number of trials per experiment doesn't vary much.
>
>But you do that by taking the RSS of the z's, not SUM_i^N{z_i}/sqrt(N).

Consider a binomial experiment where you flip a coin and register the
excess of heads over tails. The result of one experiment (excess of
heads over tails) will be (h-p*n) where h=number of heads, t=number
of tails, n=h+t and p=probability of getting a head. To determine the
z-score, measure this in standard deviations. For a binomial
experiment, the variance is n*p*(1-p) and the standard deviation is
sqrt(n*p*(1-p)). Assume a fair coin so p=1/2; the standard deviation
is sqrt(n)/2 and the z-score is 2*(h-p*n)/sqrt(n).

Suppose you have done K experiments, all with identical protocols, so
n and p are the same in each experiment[1]. The z-score of a single
experiment is 2*(h_i-p*n)/sqrt(n). If you consider all the K
experiments to be just one huge experiment, the z-score of the
overall experiment is Z=2*(H-p*N)/sqrt(N), where H=\sum{h_i} is the
total number of heads, T=\sum{t_i} is the total number of tails, and
N=H+T=K*n=\sum{n_i}.

Therefore, the overall Z-score

      Z = 2*(H-p*N)/sqrt(K*n)

          = 2*(\sum{h_i}-p*\sum{n_i})/(sqrt(n)*sqrt(K))

          = \sum{2*(h_i-p*n_i)/sqrt(n)}/sqrt(K)

          = \sum{z_i}/sqrt(K)

So Radin's formula is correct.

However, while writing this down, it occurred to me that the z's are
*signed*. That is, if the coin is truly fair, it is as probable to
get an excess of tails over heads (negative z) as heads over tails.
For example, in random event generator experiments, where a Princeton
student attempts to influence the generator to an excess of "heads"
over "tails" (or 1's over 0's), there should be negative z-scores as
well as positive ones. Is it possible that Radin got that huge
z-score by not realizing that the z-scores should be signed? Or is it
possible that a significant number of experiments ended up with
negative z-scores, and they were misreported? Or not reported at all
(filedrawer)? Any of these problems could produce a huge z-score
given such a large number of experiments.

I don't know if any of these suggestions explains Radin's number, but
just toss it out as a possibility.

Bill

[1] This seems to be the approximation Radin is using, even though
the experiments may have different number of trials. If the number of
trials per experiment doesn't vary too greatly, this may not matter
much.
 
32
Re: update on Dean Radin

Wednesday, April 11, 2007 12:43 PM
From:
"Bill Jefferys" <bill@ASTRO.AS.UTEXAS.EDU>
To:
AVOID-L@HAWAII.EDU
At 12:58 PM -0600 4/11/07, Vic Stenger wrote:

>Bill, you and I have been researchers for a long time. Have you ever
>witnessed a case where a major discovery was made either by
>metanalysis of several experiments or using sophisticated
>statistical techniques on a single experiment when the data did not
>hit you square in the eye?

Nope. Parapsychology utterly fails the interocular traumatic test.

Bill
 
33
Re: update on Dean Radin

Wednesday, April 11, 2007 3:36 PM
From:
"Julio Siqueira" <juliocbdsiqueira@YAHOO.COM>
To:
AVOID-L@HAWAII.EDU
Time to be serious (and to be my true self):

Dear Professor William Jefferys,

With all due respect, it seems to me that, IMHO, you
are not following the topic in a fair manner. (I
haven't read the messages that came after this one
below).

Sorry if I am wrong, though.

Best Wishes,
Julio Siqueira
_________



--- Bill Jefferys <bill@ASTRO.AS.UTEXAS.EDU> wrote:

> I don't see any substantive answer at all on Radin's
> blog. I do see a
> claim that z=96.5, but no supporting evidence. Radin
> says, "read his
 

34
Re: update on Dean Radin

Wednesday, April 11, 2007 4:15 PM
From:
"Bill Jefferys" <bill@ASTRO.AS.UTEXAS.EDU>
To:
AVOID-L@HAWAII.EDU

Julio,

You can't be serious! There was absolutely nothing substantive on
Radin's blog, other than his explanation of how he calculated the
z-score and his statement that he had to go find a mathematician to
convert it into a p-value. There was NO information that would allow
me to understand anything about the provenance of the data he used.
If I have missed the supporting evidence, please point to it. I don't
consider his admonition to "read his book" be at all serious; the
book is not peer-reviewed, and is intended for popular consumption.
It is not a serious scientific work.

You ARE wrong. Please rethink this.

If I don't get something more serious, I'll stop responding to your
messages. I don't like wasting my time, and I know how to use a
killfile.

Bill

At 3:36 PM -0700 4/11/07, Julio Siqueira wrote:
>Time to be serious (and to be my true self):
 

35
Re: update on Dean Radin

Wednesday, April 11, 2007 4:25 PM
From:
"Bill Jefferys" <bill@ASTRO.AS.UTEXAS.EDU>
To:
AVOID-L@HAWAII.EDU
At 12:58 PM -0600 4/11/07, Vic Stenger wrote:

>Bill, you and I have been researchers for a long time. Have you ever
>witnessed a case where a major discovery was made either by
>metanalysis of several experiments or using sophisticated
>statistical techniques on a single experiment when the data did not
>hit you square in the eye?

Something else occurred to me, Vic.

As I said in my comments (deleted for brevity), surely a 96-sigma
result would indicate that psi is so powerful that a clean,
statistics-free, hit-you-in-the-eye experiment could be devised that
would demonstrate its existence.

We did a whole bunch of work on our house two years ago. The
carpenters frequently would cut a 10 foot piece of lumber with an
error (one sigma) of 1/32 of an inch. 96 sigmas is 3 inches. If a
10-foot board three inches off, there's no way it be used for the
intended purpose. If carpenters can cut boards with such precision,
so that a 96-sigma error can hit you in the eye so forcefully, how
come the parapsychologists can't come up with a similarly convincing
experiment?

They can't, because what they are looking for isn't there. All they
can do is to conduct these silly experiments, and obfuscate with lots
of statistics.

If Julio doesn't like this, too bad.

Bill
 
36
Re: update on Dean Radin

Thursday, April 12, 2007 7:32 AM
From:
"Julio Siqueira" <juliocbdsiqueira@YAHOO.COM>
To:
AVOID-L@HAWAII.EDU

Dear Professors...

You will not have me ask the two of you if you
understand properly the technical logic behind
meta-analyses...

Vic, you have raised this point almost everywhere. God
may not be omnipresent, but you raising this question
seems to be almost omnipresent. Perhaps I will contact
some science historians on that. But let's stick to
the very words that you two are using now:

"Have you ever witnessed a case where a major
discovery was made either by metanalysis of several
experiments"

Wasn´t that exaclty what Jessica Utts showed in a
paper from 1999 that both of you know far better than
me?
[regarding the use of antiplatelets to reduced
vascular disease]"The Significance of Statistics in
Mind-Matter Research. Jessica Utts. Journal of
Scientific Exploration, Vol. 13, No. 4, pp.615–638,
1999."

And what do you think of this nice "meta-analytic"
device below:
http://www.vla.nrao.edu/

And what do you think of this "humble" (though not
Hubble, as Jefferys might prefer... :-) )list of
"discoveries?"
http://www.nrao.edu/pr/2000/vla20/background/

Isn't that a clear sign of where the logic of the
meta-analysis can take us to? Of its true potential?

But you two are the Authorities to be heard!

Best Wishes,
Julio
__________



--- Bill Jefferys <bill@ASTRO.AS.UTEXAS.EDU> wrote:

> At 12:58 PM -0600 4/11/07, Vic Stenger wrote:
>
> >Bill, you and I have been researchers for a long
> time. Have you ever
 
37
Re: update on Dean Radin

Thursday, April 12, 2007 9:46 AM
From:
"Julio Siqueira" <juliocbdsiqueira@YAHOO.COM>
To:
AVOID-L@HAWAII.EDU
Ok. So, I am sorry.

Maybe it was my naive side playing tricks on me.

I am going to do this: I will buy both Radin's "The
Conscious Universe" and Pratt et al (year 1940). It is
this last book that was the source Radin claims to
have used as input to get the final very big p value.

Some further comments, though,

"There was absolutely nothing substantive on Radin's
blog, other than his explanation of how he calculated
the z-score"

The way I see the discussions here on this topic, this
above was precisely the question we asked. Knowing
that now, we have to see the book (year 1940). I will
do it. And I will bring the data.

"There was NO information that would allow me to
understand anything about the provenance of the data
he used."

I believe it came from the tests performed by Rhine
himself (co-author of the book) and the tests of
others. Yes, that is important information that can
only be acquired by getting the book itself or by
having someone who has access to the book (Radin, for
instance) post it on the web (or send through email,
etc) for us.

"If I have missed the supporting evidence, please
point to it. I don't consider his admonition to 'read
his book' be at all serious;"

Of course it is vital to analyze the data. And that is
why I will get it.

See that you replied to me on 29 Mar 2007 as follows:

[Julio] So, what about it? Is Radin p value right or
wrong after all? (that is *the* information we need to
bring him down!).

[Jefferys] Who knows? He threw out a number, then when
a problem was pointed out, threw out another. This
does not sound like responsible science.

Now: this right above is very serious. If Radin did
indeed "threw out another number", that is fraud pure
and simple. And not only that. He had Brian Josephson,
Nick Herbert, Michael Rossman, and maybe others
pressure Nature with his fraudulent numbers. It is
utterly unacceptable.

The value of this data from 1880 to 1940 per se is,
IMO, minimal (I am sure most parapsychologists, Radin
included, agree with me on this). The seriousness of
Radin's misconduct if he forged a number to pressure
Nature is utterly unexcusable.

That is the point.

Thank you,
Julio
_________





--- Bill Jefferys <bill@ASTRO.AS.UTEXAS.EDU> wrote:

> Julio,
>
> You can't be serious! There was absolutely nothing
> substantive on
 
38
Re: update on Dean Radin

Thursday, April 12, 2007 3:28 PM
From:
"Brent Meeker" <meekerdb@DSLEXTREME.COM>
To:
AVOID-L@HAWAII.EDU
 
> Bill
>
> [1] This seems to be the approximation Radin is using, even though
> the experiments may have different number of trials. If the number of
> trials per experiment doesn't vary too greatly, this may not matter
> much.
>
>

Yes, I see how that would work for benoulli trials, but I think most of the data Radin is referring to is from Rhine's ESP symbol experiments.  Then it's multinomial with four variables.  And since the subject knows how many cards there are of each kind, the quesses are even independent.  So I don't think the equivalent z-value is so simple as you've shown for benoulli trials.  Certainly there will be negative scores, in which the subject gets fewer than the most likely number right.

Brent Meeker

39
Re: update on Dean Radin

Thursday, April 12, 2007 3:43 PM
From:
"Bill Jefferys" <bill@ASTRO.AS.UTEXAS.EDU>
To:
AVOID-L@HAWAII.EDU
 
>are even independent.  So I don't think the equivalent z-value is so
>simple as you've shown for benoulli trials.  Certainly there will be
>negative scores, in which the subject gets fewer than the most
>likely number right.

Rhine amongst others ignored the independence issue, which allows a
clever person to obtain "above average" results due to the fact that
the number of Zener cards of each kind in a deck is known. Persi
Diaconis wrote about this error...common in the parapsychological
world...in an article in *Science* that must have appeared about 30
years ago. I don't know if it's available online but you could look.

In any case, the fact that the z-scores should be signed won't be changed.

I don't know if there is an issue for combining z-scores in such a
situation. That is, I don't know if Rosenthal's meta-analytic
formulae will work in such situations. If the p-values were
calculated as if everything is independent, then as Diaconis shows,
the calculations will bias towards smaller p-values. In other words,
if Rhine and Radin didn't take it into account, it's a matter of
Garbage In, Garbage Out.

Bill


Bill

40
Re: update on Dean Radin

Thursday, April 12, 2007 3:46 PM
From:
"Bill Jefferys" <bill@ASTRO.AS.UTEXAS.EDU>
To:
AVOID-L@HAWAII.EDU

At 7:32 AM -0700 4/12/07, Julio Siqueira wrote:
>Dear Professors...
>
>You will not have me ask the two of you if you
>understand properly the technical logic behind
>meta-analyses...

What do you think my calculations in response to Brent's questions were?

>Vic, you have raised this point almost everywhere. God
>may not be omnipresent, but you raising this question
>seems to be almost omnipresent. Perhaps I will contact
>some science historians on that. But let's stick to
>the very words that you two are using now:
>
>"Have you ever witnessed a case where a major
>discovery was made either by metanalysis of several
>experiments"
>
>Wasn´t that exaclty what Jessica Utts showed in a
>paper from 1999 that both of you know far better than
>me?
>[regarding the use of antiplatelets to reduced
>vascular disease]"The Significance of Statistics in
>Mind-Matter Research. Jessica Utts. Journal of
>Scientific Exploration, Vol. 13, No. 4, pp.615–638,
>1999."

Is this significant, and has it held up under time?

>And what do you think of this nice "meta-analytic"
>device below:
>http://www.vla.nrao.edu/

What does this have to do with meta-analysis? I can't find the word
on this webpage.

>And what do you think of this "humble" (though not
>Hubble, as Jefferys might prefer... :-) )list of
>"discoveries?"
>http://www.nrao.edu/pr/2000/vla20/background/

What does this have to do with meta-analysis? I can't find the word
on this webpage.

>Isn't that a clear sign of where the logic of the
>meta-analysis can take us to? Of its true potential?
>
>But you two are the Authorities to be heard!
>
>Best Wishes,
>Julio

Bill
 

41
Re: update on Dean Radin

Friday, April 13, 2007 4:55 AM
From:
"Julio Siqueira" <juliocbdsiqueira@YAHOO.COM>
To:
AVOID-L@HAWAII.EDU

Hello Bill,
 
You said something very interesting below:
 
"Something else occurred to me, Vic. As I said in my comments (deleted for brevity), surely a 96-sigma result would indicate that psi is so powerful that a clean, statistics-free, hit-you-in-the-eye experiment could be devised that would demonstrate its existence. We did a whole bunch of work on our house two years ago. The carpenters frequently would cut a 10 foot piece of lumber with an error (one sigma) of 1/32 of an inch. 96 sigmas is 3 inches. If a 10-foot board three inches off, there's no way it be used for the intended purpose. If carpenters can cut boards with such precision, so that a 96-sigma error can hit you in the eye so forcefully, how come the parapsychologists can't come up with a similarly convincing experiment?"
 
If I understood you correctly, your carpenters had to cut several lumber, and they had to cut them so that they ended up 10-foot long. No one is perfect. Sometimes the lumber would come out 10-foot and 1/32 of an inch bigger than that, or shorter than that. Clumsy carpenters (or kindergarten fellows) would produce lumbers 10-foot and 3 inches long (and sometimes lumbers that would not have 10 feet, but 3 inches less than that actually). Ok?
 
So, each single performing of a piece of lumber cutting is an isolated experiment. If the *effect size is big enough* (and kids do commit big mistakes...), you could have, in just one single experiment (one single lumber-cutting) a 96 sigma (i.e. lumber with a length of 10 feet and 3 inches).
 
However, sometimes the mistakes are not that much evident. Nevertheless, they still exist, and can be detected. Imagine a very very good carpenter. The best one in town (Jesus Christ). Trying to get lumbers 10 feet long (ideal for Roman crosses), he fairly consistently produced an error of 1/1250 of an inch for each piece of lumber (that is about the size of a bacterium). That is the *effect size of His Holy mistake* for each single experiment. I cannot calculate the sigma of that (for each single lumber-cutting). But you can. However, I know that, with only one experiment, we will not get a sigma of 96. But if we pool together all the lumber He has been cutting for the last 2000 years (meta-analysis), then we might come to a 96 sigma.
 
What you claim (you and Vic) is that it is not valid to combine the results Christ's Holy sawing. Is it Right?
 
Julio
________
 
42
Re: update on Dean Radin

Friday, April 13, 2007 10:37 AM
From:
"Bill Jefferys" <bill@ASTRO.AS.UTEXAS.EDU>
To:
AVOID-L@HAWAII.EDU

At 4:55 AM -0700 4/13/07, Julio Siqueira wrote:
>Hello Bill,
>
>You said something very interesting below:
>
>"Something else occurred to me, Vic. As I said in my comments
>(deleted for brevity), surely a 96-sigma result would indicate that
>psi is so powerful that a clean, statistics-free, hit-you-in-the-eye
>experiment could be devised that would demonstrate its existence. We
>did a whole bunch of work on our house two years ago. The carpenters
>frequently would cut a 10 foot piece of lumber with an error (one
>sigma) of 1/32 of an inch. 96 sigmas is 3 inches. If a 10-foot board
>three inches off, there's no way it be used for the intended
>purpose. If carpenters can cut boards with such precision, so that a
>96-sigma error can hit you in the eye so forcefully, how come the
>parapsychologists can't come up with a similarly convincing
>experiment?"
>
>If I understood you correctly, your carpenters had to cut several
>lumber, and they had to cut them so that they ended up 10-foot long.
>No one is perfect. Sometimes the lumber would come out 10-foot and
>1/32 of an inch bigger than that, or shorter than that. Clumsy
>carpenters (or kindergarten fellows) would produce lumbers 10-foot
>and 3 inches long (and sometimes lumbers that would not have 10
>feet, but 3 inches less than that actually). Ok?
>
>So, each single performing of a piece of lumber cutting is an
>isolated experiment. If the *effect size is big enough* (and kids do
>commit big mistakes...), you could have, in just one single
>experiment (one single lumber-cutting) a 96 sigma (i.e. lumber with
>a length of 10 feet and 3 inches).
>
>However, sometimes the mistakes are not that much evident.
>Nevertheless, they still exist, and can be detected. Imagine a very
>very good carpenter. The best one in town (Jesus Christ). Trying to
>get lumbers 10 feet long (ideal for Roman crosses), he fairly
>consistently produced an error of 1/1250 of an inch for each piece
>of lumber (that is about the size of a bacterium). That is the
>*effect size of His Holy mistake* for each single experiment. I
>cannot calculate the sigma of that (for each single lumber-cutting).
>But you can. However, I know that, with only one experiment, we will
>not get a sigma of 96. But if we pool together all the lumber He has
>been cutting for the last 2000 years (meta-analysis), then we might
>come to a 96 sigma.
>
>What you claim (you and Vic) is that it is not valid to combine the
>results Christ's Holy sawing. Is it Right?

That's Radin's claim.

Here's why I don't believe it.

I claim that there are limits to what you can accomplish by taking
more and more data.

You have to remember that I have been dealing with data, real data,
for over 40 years.

I have learned that it is very difficult to pull a signal out of a
large amount of noise.

If you have a signal that is swamped by the noise, there is only a
limited amount that statistics can do for you to pull it out.

Amateur scientists, and I would include most of the parapsychologists
I am aware of, imagine that all you have to do is to take more and
more data and eventually the signal will come out, due to the magic
of "the square root of N". This is an extraordinarily naive belief,
and in real life it will lead to grief if carried too far.

The reason is that there are ALWAYS unknown experimental biases in
any experiment that have nothing to do with the phenomenon you are
trying to investigate. They are always there, they are unavoidable.
It is virtually certain that if you have a per-data-point error of a
certain amount, then at the level of 1/100 of that error, and
sometimes greater, there will be some sort of bias of this sort.

I know of two ways to control such biases: (1) Identify their source
and measure them accurately enough to eliminate them or (2) Do a
better experiment that has a smaller per-data-point error.

Parapsychologists do none of this. Instead, they have a touching, but
naive faith in the power of the square root of N to overcome all
uncertainty.

It is a good rule of thumb that you can't get a signal out of noise
that is much greater than 100 times the signal. Maybe at the outside
1000, but that is very dicey, and would require a very successful and
diligent application of method (1) above. Parapsychologists claim to
be able to do much better than this.

Because of my experience in working with real data, I just don't believe them.

Meta-analysis won't help you. All it does is to make N bigger without
giving any understanding of the underlying problems that are
certainly in the experiments.

If Radin's meta-analysis had any merit, he could do something like
the following, which is often done by real scientists: Divide his
data set into a number of smaller sets of data. He's got 186
experiments, he can construct say 10 groups of 18 or 19 by randomly
choosing them from the pool of 186 (or better, having someone with no
axe to grind do the choosing, like a professional and neutral
statistician). Then do his meta-analysis on each of the 10 groups.
Since the sigma of 96.5 standard deviations is so huge, the sigmas of
the individual experiments will be about 30, still huge. Therefore,
if the different experiments are measuring the same thing, they all
ought to produce the same effect in the same direction to quite high
precision.

I will bet that if he did this, the effects would be all over the map.

This would prove that they aren't measuring anything at all.

I know what Radin would say to this proposal. He will whine that psi
is variable and unpredictable, that will be his excuse.

It's the same excuse that parapsychologists have been giving for 100
years. Look up "sheep and goat effect," "decline effect," etc. It is
all excuses intended to paper over the fact that there is no there
there.

This is why I think the whole
statistical-analysis-of-a-humungus-number-of-trials paradigm that has
dominated parapsychology since Rhine will NEVER be convincing to
those not already convinced.

Parapsychologists need an experiment that passes Jimmie Savage's
interocular traumatic test. I said as much in print over fifteen
years ago, and suggested some ideas about how to go about it. The
ideas that I suggested overcome the objection you suggest: By working
with individual atoms at the microscopic level, it should be possible
to devise clear and convincing experiments, for example telekenesis,
by having people move individual atoms around on a substrate at will.
Since the claims of the random event generator experiments are that
people can control the decay of atomic particles (a much higher
energy process), surely something simple like this would work.

So far, I've seen no interest from the parapsychologists in doing
anything like this. I got a few excuses from people who basically
said, there's no point in doing this because psi might not work that
way. What kind of excuse is that?

Instead, the parapsychologists continue on the same old treadmill,
mindlessly producing large numbers based on huge numbers of trials in
meaningless experiments, with no significant attention to the
problems I mention above.

And they wonder why they can't convince anyone but the true believers!

Bill

43
Re: update on Dean Radin

Saturday, April 14, 2007 2:12 PM
From:
"Julio Siqueira" <juliocbdsiqueira@YAHOO.COM>
To:
AVOID-L@HAWAII.EDU

Hi Bill,

In Utts article, she says:

"For the combined data the
results indicate that the odds of a second occurrence
are reduced by about 25% as a result of taking
antiplatelets. In fact the conclusion from the
abstract of the
original article was clear:"

"Thus antiplatelet treatment can reduce the incidence
of serious vascular events by
about a quarter among a wide range of patients at
particular risk of occlusive vascular
disease (Antiplatelet Trialists’ Collaboration, 1988,
p. 320)."

So, IMO, it seems that this satisfies Stenger's demand
(as far as I know).

Best Wishes,
Julio
P.S.: you can see each radio telescope in VLA as a
single experiment, or data source. Obviously I know
this comparison will look very forceful. But in my
opinion it runs perfectly in line with the logic of
meta-analysis.
______



--- Bill Jefferys <bill@ASTRO.AS.UTEXAS.EDU> wrote:

> At 7:32 AM -0700 4/12/07, Julio Siqueira wrote:
> >Dear Professors...
> >
> >You will not have me ask the two of you if you
> >understand properly the technical logic behind
> >meta-analyses...
 
44
Re: update on Dean Radin

Saturday, April 14, 2007 2:29 PM
From:
"Bill Jefferys" <bill@ASTRO.AS.UTEXAS.EDU>
To:
AVOID-L@HAWAII.EDU

At 2:12 PM -0700 4/14/07, Julio Siqueira wrote:
>Hi Bill,
>
>In Utts article, she says:
>
>"For the combined data the
>results indicate that the odds of a second occurrence
>are reduced by about 25% as a result of taking
>antiplatelets. In fact the conclusion from the
>abstract of the
>original article was clear:"
>
>"Thus antiplatelet treatment can reduce the incidence
>of serious vascular events by
>about a quarter among a wide range of patients at
>particular risk of occlusive vascular
>disease (Antiplatelet Trialists’ Collaboration, 1988,
>p. 320)."
>
>So, IMO, it seems that this satisfies Stenger's demand
>(as far as I know).

Major scientific discovery? I believe that this was one of Vic's condition.

You still haven't indicated whether, ten years later, this research
has withstood the test of time, either.

>Best Wishes,
>Julio
>P.S.: you can see each radio telescope in VLA as a
>single experiment, or data source. Obviously I know
>this comparison will look very forceful. But in my
>opinion it runs perfectly in line with the logic of
>meta-analysis.

It has nothing to do with meta-analysis. If you think it does, then
you know nothing of meta-analysis. The VLA is an interferometer. You
cannot consider the results as being due to taking the data
separately with each telescope, getting a bunch of results, and then
combining them. This isn't how interferometers work. The logic of
interferometers has nothing to do with meta-analysis. The logic of
meta-analysis has nothing to do with how interferometers work.

You are getting rather silly, and you are clearly out of your depth.

Bill

>______
 

45
Re: update on Dean Radin

Saturday, April 14, 2007 5:47 PM
From:
"Julio Siqueira" <juliocbdsiqueira@YAHOO.COM>
To:
AVOID-L@HAWAII.EDU

[Julio]
>> vascular disease]"The Significance of Statistics in
>> Mind-Matter Research. Jessica Utts. Journal of
>> Scientific Exploration, Vol. 13, No. 4, pp.615–638,
>> 1999."
>
 
[Jefferys]
> Is this significant, and has it held up under time?
[Victor]
Absolutely not. Absolutely not.
I hear Utts talk. She tried to say that p = 0.05 was significant. 
What nonsense!
[Julio Now!]
Definetely yes. Just take a look at Psi Wars (Journal of Consciousness Studies, year 2003). The cut value p = 0.05 that you, Vic, point out is indeed very problematic. But Utts does not restrict her reasoning to it. No use making a "straw woman" out of her...
 
Best,
Julio
____________
 
 

Vic Stenger <vstenger@MINDSPRING.COM> wrote:
On Apr 12, 2007, at 4:46 PM, Bill Jefferys wrote:

> At 7:32 AM -0700 4/12/07, Julio Siqueira wro
 
46
Re: update on Dean Radin

Monday, April 16, 2007 5:28 PM
From:
"Julio Siqueira" <juliocbdsiqueira@YAHOO.COM>
To:
AVOID-L@HAWAII.EDU

Hi Bill,

Some few items for you to ponder below...

--- Bill Jefferys <bill@ASTRO.AS.UTEXAS.EDU> wrote:

> At 2:12 PM -0700 4/14/07, Julio Siqueira wrote:
> >Hi Bill,
> >
> >In Utts article, she says:
> >
> >"For the combined data the
> >results indicate that the odds of a second
> occurrence
> >are reduced by about 25% as a result of taking
> >antiplatelets. In fact the conclusion from the
> >abstract of the
> >original article was clear:"
> >
> >"Thus antiplatelet treatment can reduce the
> incidence
> >of serious vascular events by
> >about a quarter among a wide range of patients at
> >particular risk of occlusive vascular
> >disease (Antiplatelet Trialists’ Collaboration,
> 1988,
> >p. 320)."
> >
> >So, IMO, it seems that this satisfies Stenger's
> demand
> >(as far as I know).
>
> Major scientific discovery? I believe that this was
> one of Vic's condition.

Oh yes. Discoveries that help save lives are
definetely not major discoveries. And surely "major
discovery" is a very well defined term.

>
> You still haven't indicated whether, ten years
> later, this research
> has withstood the test of time, either.

And you have not indicated whether it has not.

>
> >Best Wishes,
> >Julio
> >P.S.: you can see each radio telescope in VLA as a
> >single experiment, or data source. Obviously I know
> >this comparison will look very forceful. But in my
> >opinion it runs perfectly in line with the logic of
> >meta-analysis.
>
> It has nothing to do with meta-analysis. If you
> think it does, then
> you know nothing of meta-analysis. The VLA is an
> interferometer. You
> cannot consider the results as being due to taking
> the data
> separately with each telescope, getting a bunch of
> results, and then
> combining them. This isn't how interferometers work.
> The logic of
> interferometers has nothing to do with
> meta-analysis. The logic of
> meta-analysis has nothing to do with how
> interferometers work.
>
> You are getting rather silly, and you are clearly
> out of your depth.

I greatly doubt that emotional bursts like the one
above can be the path to... "good science"... (or to
faithful evaluation of... Radin's work)

If the VLA is not a good metaphore to meta-analysis,
then it is not. It is that simple.

Best Wishes,
Julio
___________
 
47
Re: update on Dean Radin

Tuesday, April 17, 2007 6:43 AM
From:
"Bill Jefferys" <bill@ASTRO.AS.UTEXAS.EDU>
To:
AVOID-L@HAWAII.EDU


At 5:28 PM -0700 4/16/07, Julio Siqueira wrote:
>Hi Bill,
>
>Some few items for you to ponder below...
>
>--- Bill Jefferys <bill@ASTRO.AS.UTEXAS.EDU> wrote:
>
>> At 2:12 PM -0700 4/14/07, Julio Siqueira wrote:
>> >
>> >"Thus antiplatelet treatment can reduce the
>> >incidence of serious vascular events by
>> >about a quarter among a wide range of patients at
>> >particular risk of occlusive vascular
>> >disease (Antiplatelet Trialists’ Collaboration,
>> >1988,
>> >p. 320)."
>> >
>> >So, IMO, it seems that this satisfies Stenger's demand
>> >(as far as I know).
>>
>> Major scientific discovery? I believe that this was
>> one of Vic's condition.
>
>Oh yes. Discoveries that help save lives are
>definetely not major discoveries. And surely "major
>discovery" is a very well defined term.

Here are some examples of major scientific discoveries: Newton's
laws, the wave nature of light, special and general relativity,
quantum mechanics, evolution by natural selection, Mendelian
inheritance, the structure of DNA, plate tectonics, the expansion of
the universe, etc. A major scientific discovery has huge
ramifications down the line, in many different fields unrelated to
the putative field of the discovery itself. It is the sort of thing
for which honors follow: prizes, election to prestigious national
academies, etc.

That aspirin is useful in some people to reduce mortality is NOT a
major scientific discovery. It is a technical improvement in
treatment of a particular narrow condition. It is routine, everyday
science, not Nobel-prize material.
I note that the actual reduction looks to be about 20%, not a quarter
(25%), as asserted by Utts (see figure below).
>> You still haven't indicated whether, ten years
>> later, this research
>> has withstood the test of time, either.
>
>And you have not indicated whether it has not.

It has been discovered since Utts' paper that aspirin prophylaxis is
not useful in some populations. I, for example, would be put at
greater risk from other conditions should I to use it than the
benefit I might get. Women seem not to get much benefit from it.
People who have not had heart attacks get less if any benefit. So,
while is a certain particular population the prophylaxis is still
useful, it is not universally useful. In that limited sense it
appears to have stood the test of time.

Now, the other claim you made is that this discovery was a result of
meta-analysis. I will now humor you briefly and PRETEND that it is a
major discovery. I will show that the discovery was NOT because of
meta-analysis. My exhibit is this figure (Figure 1 from the Utts
paper):
 


When you examine this figure, you see immediately that the results of
the individual studies are all consistent with each other. All but
three show an improvement in mortality, and the three that do not are
all of low accuracy and are still consistent with the remaining
studies that do show this. One of the studies shows a very large
improvement, greater than the 95% error bar, but that's consistent
with the fact that there are 25 studies shown, and one can expect one
in twenty to be outside its 95% confidence interval. Anyone who
looked at this chart would conclude that it is highly likely that
aspirin is a useful treatment in these cases. Just by eyeball alone,
one can see that the line that is drawn (dashed vertical line) is a
pretty good guess at what the studies would show when combined. You
don't need meta-analysis to see this, and the people that conducted
the meta-analysis didn't need it either! All the meta-analysis did
was to tighten the error bars and give a more accurate overall
picture of where the odds ratio is (about 0.8, a modest improvement).
Basically, all the meta-analysis amounts to is taking a weighted
average of the results along with improving the error bar by an
amount roughly proportional to 1/sqrt(N). I can get the basic result
by eyeball without doing any calculation, just from the figure.

And I have already pointed out to you that you can't go very far with
1/sqrt(N). In this case, the signal is larger than most of the error
bars, so it is an example of data that easily passes Jimmie Savage's
interocular traumatic test. You don't need fancy statistics, the
result is obvious in the data themselves.
 
So I say that even before meta-analysis, people were pretty sure of
what they would find out when meta-analysis was done. So to attribute
this result to meta-analysis, instead of just common sense, is a huge
stretch.

To continue: I would bet that if a similar chart of Radin's 183
studies were drawn, it would look nothing like this figure. I would
predict that the points and their error bars will be all over the
map, and inconsistent with each other [all the data plotted above are
consistent]. In other words, what Radin is doing, in contrast to the
above study, is pulling noise out of noise, not signal out of noise.
To summarize: This is not an example of a major scientific discovery
made possible by meta-analysis. It isn't a major scientific discovery
in the first place, and even if it were, it is one that would have
been made had formal meta-analysis not or not been performed.
 
>> >Best Wishes,
>> >Julio
>> >P.S.: you can see each radio telescope in VLA as a
>> >single experiment, or data source. Obviously I know
>> >this comparison will look very forceful. But in my
>> >opinion it runs perfectly in line with the logic of
>> >meta-analysis.
>>
>>It has nothing to do with meta-analysis. If you think it does, then
>>you know nothing of meta-analysis. The VLA is an interferometer.
>>You cannot consider the results as being due to taking the data
>>separately with each telescope, getting a bunch of results, and
>>then combining them. This isn't how interferometers work. The logic
>>of interferometers has nothing to do with meta-analysis. The logic
>>of meta-analysis has nothing to do with how interferometers work.
>>
>>You are getting rather silly, and you are clearly out of your depth.
>
>I greatly doubt that emotional bursts like the one
>above can be the path to... "good science"... (or to
>faithful evaluation of... Radin's work)

Just stating the facts. No emotion!

90% of everything is crap -- Sturgeon's Law. One of the major
responsibilities of scientists is to recognize crap and point it out
when they see it.

>If the VLA is not a good metaphore to meta-analysis,
>then it is not. It is that simple.

If you had any notion about what meta-analysis is, you wouldn't have
suggested this example.

Bill

48
Update on Dean Radin - New DATA !

Wednesday, April 18, 2007 5:46 AM
From:
"Julio Siqueira" <juliocbdsiqueira@YAHOO.COM>
To:
AVOID-L@HAWAII.EDU

This is specially for William Jefferys and to Brent Meeker. [this message is also being sent to Dean Radin, to Brazilians psi researchers of the Interpsi group, and to Brazilians psi survivalist researchers of the ECAE group].
 
I have placed on my site the book "ESP After 60 Years" (Pratt et all, 1940). This is the source Radin claims to have used to reach the p value of about 10-2000, so now I believe it will be possible to tell fact from fraud (wherever it lies...). I have split the book in five parts. And I have made it available both in doc format and in pdf format. The links are below:
 
Doc format:
 
http://paginas.terra.com.br/educacao/criticandokardec/rhine_1940_pg_0_100.doc
http://paginas.terra.com.br/educacao/criticandokardec/rhine_1940_pg_101_200.doc
http://paginas.terra.com.br/educacao/criticandokardec/rhine_1940_pg_201_300.doc
http://paginas.terra.com.br/educacao/criticandokardec/rhine_1940_pg_301_400.doc
http://paginas.terra.com.br/educacao/criticandokardec/rhine_1940_pg_401_466.doc
 
Pdf format:
 
http://paginas.terra.com.br/educacao/criticandokardec/rhine_1940_pg_0_100.pdf
http://paginas.terra.com.br/educacao/criticandokardec/rhine_1940_pg_101_200.pdf
http://paginas.terra.com.br/educacao/criticandokardec/rhine_1940_pg_201_300.pdf
http://paginas.terra.com.br/educacao/criticandokardec/rhine_1940_pg_301_400.pdf
http://paginas.terra.com.br/educacao/criticandokardec/rhine_1940_pg_401_466.pdf
 
Any problems with downloading, please contact me (juliocbsiqueira@terra.com.br).
 
Also, my copy of "The Conscious Universe" (Dean Radin, 1996) has already shipped (from www.amazon.com), and I should get it in one week or two. The exact (and full) wordings of Radin in this book, regarding this whole issue, will be available then.
 
Further, I sent to Radin one piece of technical criticism that appeared in more recent exchange of messages on the Avoid-L skeptic forum. I reproduce below the criticism (or questioning) and Radin's answer:
 
Criticism/Questioning:
 
Consider a binomial experiment where you flip a coin... ... ...
So Radin's formula is correct.
However, while writing this down, it occurred to me that the z's are
*signed*. That is, if the coin is truly fair, it is as probable to
get an excess of tails over heads (negative z) as heads over tails.
For example, in random event generator experiments, where a Princeton
student attempts to influence the generator to an excess of "heads"
over "tails" (or 1's over 0's), there should be negative z-scores as
well as positive ones. Is it possible that Radin got that huge
z-score by not realizing that the z-scores should be signed? Or is it
possible that a significant number of experiments ended up with
negative z-scores, and they were misreported? Or not reported at all
(filedrawer)? Any of these problems could produce a huge z-score
given such a large number of experiments.
I don't know if any of these suggestions explains Radin's number, but
just toss it out as a possibility.
 
 
Radin's Answer, exactly the way he presented it in his email to me:
 
"Is it possible that Radin got that huge z-score by not realizing that the z-scores should be signed? "
 
No, it is not possible. Z scores are distributed normally around a mean of 0.
 
"Or is it possible that a significant number of experiments ended up with negative z-scores, and they were misreported? "
 
Mistakes are always possible, but the table listing these results also provides the number of hits and trials in each experiment, from which it is easy to determine the hit rate and thus the z score. The results as reported look correct to me and were very carefully vetted by the original authors, and by many readers, over the last 60 years.
 
"Or not reported at all (filedrawer)? "
 
This is point of Good's comment in Nature, and my response. There are simple ways of calculating the number of supposedly missing studies required to neutralize reported results. In this case the estimated filedrawer would have to be extremely large. Did selective reporting happen? Probably. Did it happen often enough to nullify the overall effect? I doubt it.
 
- Dean
xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
 
My Conclusion so far:
 
It seems that the main critics of Radin on Avoid-L have not found any technical problem with his statistic calculations regarding the Radin vs Good (Nature review) issue.
 
Best Regards,
Julio
______________
 
49
Re: Update on Dean Radin - New DATA !

Wednesday, April 18, 2007 6:32 AM
From:
"Bill Jefferys" <bill@ASTRO.AS.UTEXAS.EDU>
To:
AVOID-L@HAWAII.EDU

At 5:46 AM -0700 4/18/07, Julio Siqueira wrote:

>Further, I sent to Radin one piece of technical criticism that
>appeared in more recent exchange of messages on the Avoid-L skeptic
>forum. I reproduce below the criticism (or questioning) and Radin's
>answer:
>
>Criticism/Questioning:
>
>Consider a binomial experiment where you flip a coin... ... ...
>So Radin's formula is correct.
>However, while writing this down, it occurred to me that the z's are
>*signed*. That is, if the coin is truly fair, it is as probable to
>get an excess of tails over heads (negative z) as heads over tails.
>For example, in random event generator experiments, where a Princeton
>student attempts to influence the generator to an excess of "heads"
>over "tails" (or 1's over 0's), there should be negative z-scores as
>well as positive ones. Is it possible that Radin got that huge
>z-score by not realizing that the z-scores should be signed? Or is it
>possible that a significant number of experiments ended up with
>negative z-scores, and they were misreported? Or not reported at all
>(filedrawer)? Any of these problems could produce a huge z-score
>given such a large number of experiments.
>I don't know if any of these suggestions explains Radin's number, but
>just toss it out as a possibility.
>
>
>Radin's Answer, exactly the way he presented it in his email to me:
>
>"Is it possible that Radin got that huge z-score by not realizing
>that the z-scores should be signed? "
>
>No, it is not possible. Z scores are distributed normally around a mean of 0.
>
>"Or is it possible that a significant number of experiments ended up
>with negative z-scores, and they were misreported? "
>
>Mistakes are always possible, but the table listing these results
>also provides the number of hits and trials in each experiment, from
>which it is easy to determine the hit rate and thus the z score. The
>results as reported look correct to me and were very carefully
>vetted by the original authors, and by many readers, over the last
>60 years.
>
>"Or not reported at all (filedrawer)? "
>
>This is point of Good's comment in Nature, and my response. There
>are simple ways of calculating the number of supposedly missing
>studies required to neutralize reported results. In this case the
>estimated filedrawer would have to be extremely large. Did selective
>reporting happen? Probably. Did it happen often enough to nullify
>the overall effect? I doubt it.
>
>- Dean
>xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
>
>My Conclusion so far:
>
>It seems that the main critics of Radin on Avoid-L have not found
>any technical problem with his statistic calculations regarding the
>Radin vs Good (Nature review) issue.

I thank you for posting the book. BUT...

You apparently did not transmit to Radin the information about the
filedrawer problem discussed in Jeff Scargle's paper. Scargle shows
that the size of the filedrawer required to neutralize results is
much smaller than the usual calculation gives, and in fact can be
quite small indeed. Thus, this part of my question has not been dealt
with.

Nor does Radin deal with the fact that observed p-values cannot be
interpreted as probabilities in any meaningful sense, as Berger and
Delampady, and Berger and Sellke have pointed out. A p-value is not
the Type I error rate for the investigation, for example, though it
is often claimed to be such (the Type I error rate, conditioned on
the p-value, is generally larger than the p-value, in a conditional
frequentist calculation); it is not probability of the null
hypothesis being false (that requires a Bayesian approach); nor is it
the probability that the "results were obtained by chance" (that is
always 100%, by definition, since in any stochastic process the
results, whatever they are, are always obtained by chance). Indeed,
p-values don't have a reasonable probabilistic interpretation in
terms of the actual experiment that has been performed.

Did Radin perform the calculation I suggested, that is, randomly
dividing up the experiments into, say 10 groups of 18 or 19,
calculating the z-scores, and seeing if they are giving the same
basic result, or are (as I suspect will turn out to be the case) all
over the map? This is the acid test to see if what he is doing is
actually pulling a real signal out of noise, or (as I suspect)
pulling noise out of noise.

Did Radin deal with the problem of quality control? That is, we know
that Rhine's early experiments were poorly controlled, and that the
size of the effect declined with improving controls. How do we know
that all of the data going into his calculation were so well
controlled that biases were so small as to be ignorable? The point
is, unless the biases can be demonstrated to be much smaller than any
effect being reported, his calculation amounts to GIGO.

Are we absolutely certain that none of the data in his calculation
involved cheating? (I am not accusing Radin of cheating, but am
deeply concerned about the provenance of the historical data he
used.) Even a small amount of cheating by the original investigators
(now all dead, I presume) would render his calculation meaningless.
We know that some highly touted experiments in parapsychology
involved cheating (e.g., Soal and Bateman, as exposed by Betty
Markwick). How do we verify, to everyone's satisfaction, that no
cheating was involved in these data?

As I pointed out, I do not believe anyone who claims to pull results
out of data at levels much smaller than 0.01 times the standard
deviation of the elemental data, or possibly 0.001 if extremely
carefully done (which I doubt in general for parapsychology
experiments). In the case of binary choices, whether at equal or
unequal probabilities, the standard deviation of a single experiment
is of order 1/2, so I would be very suspicious of any reported effect
size smaller than about 0.005 and would not believe at all any effect
size smaller than 0.0005 (on a scale of 0 to 1). I would attribute
such results to experimental defects, and not to psychic effects.

Why don't parapsychologists devise experiments that pass the
interocular traumatic test? Let's face it, no one who is not already
convinced will be convinced by experiments of the kind that Radin
uses in his calculation.

I've pointed all of these things out to you before.

Bill
 
50
Re: Update on Dean Radin - New DATA !

Wednesday, April 18, 2007 9:49 AM
From:
"Brent Meeker" <meekerdb@DSLEXTREME.COM>
To:
AVOID-L@HAWAII.EDU

Julio Siqueira wrote:
> This is specially for William Jefferys and to Brent Meeker. [this message is also being sent to Dean Radin, to Brazilians psi researchers of the Interpsi group, and to Brazilians psi survivalist researchers of the ECAE group].
>  I have placed on my site the book "/ESP After 60 Years/" (Pratt et all, 1940). This is the source Radin claims to have used to reach the p value of about 10^-2000 , so now I believe it will be possible to tell fact from fraud

The usual false dichotomy: fact or fraud.  As pointed out in the paper on ganzefeld experiments the problem can be unintended bias.  In any case it is unlikely that fraud can be discovered from published reports.  The fraud and bias take place in the testing, not in the reporting or the analysis.

> (wherever it lies...). I have split the book in five parts. And I have made it available both in doc format and in pdf format. The links are below:
>  Doc format:
>  http://paginas.terra.com.br/educacao/criticandokardec/rhine_1940_pg_0_100.doc
> http://paginas.terra.com.br/educacao/criticandokardec/rhine_1940_pg_101_200.doc
> http://paginas.terra.com.br/educacao/criticandokardec/rhine_1940_pg_201_300.doc
> http://paginas.terra.com.br/educacao/criticandokardec/rhine_1940_pg_301_400.doc
> http://paginas.terra.com.br/educacao/criticandokardec/rhine_1940_pg_401_466.doc
>  Pdf format:
>  http://paginas.terra.com.br/educacao/criticandokardec/rhine_1940_pg_0_100.pdf
> http://paginas.terra.com.br/educacao/criticandokardec/rhine_1940_pg_101_200.pdf
> http://paginas.terra.com.br/educacao/criticandokardec/rhine_1940_pg_201_300.pdf
> http://paginas.terra.com.br/educacao/criticandokardec/rhine_1940_pg_301_400.pdf
> http://paginas.terra.com.br/educacao/criticandokardec/rhine_1940_pg_401_466.pdf
>  Any problems with downloading, please contact me (juliocbsiqueira@terra.com.br).
>  Also, my copy of "The Conscious Universe" (Dean Radin, 1996) has already shipped (from www.amazon.com), and I should get it in one week or two. The exact (and full) wordings of Radin in this book, regarding this whole issue, will be available then.
>  Further, I sent to Radin one piece of technical criticism that appeared in more recent exchange of messages on the Avoid-L skeptic forum. I reproduce below the criticism (or questioning) and Radin's answer:
>  *Criticism/Questioning:*
>  Consider a binomial experiment where you flip a coin... ... ...
> So Radin's formula is correct.
> However, while writing this down, it occurred to me that the z's are
> *signed*. That is, if the coin is truly fair, it is as probable to
> get an excess of tails over heads (negative z) as heads over tails.
> For example, in random event generator experiments, where a Princeton
> student attempts to influence the generator to an excess of "heads"
> over "tails" (or 1's over 0's), there should be negative z-scores as
> well as positive ones. Is it possible that Radin got that huge
> z-score by not realizing that the z-scores should be signed? Or is it
> possible that a significant number of experiments ended up with
> negative z-scores, and they were misreported? Or not reported at all
> (filedrawer)? Any of these problems could produce a huge z-score
> given such a large number of experiments.
> I don't know if any of these suggestions explains Radin's number, but
> just toss it out as a possibility.
>   *Radin's Answer, exactly the way he presented it in his email to me:*
>  "Is it possible that Radin got that huge z-score by not realizing that the z-scores should be signed? "
>  No, it is not possible. Z scores are distributed normally around a mean of 0.

This response is self-contradictory!  A variable that is normally distributed around a mean of zero is as likely to be negative as positive.  If that is the case the answer would be "Yes, it is possible."  Averaging a large number of such variable values would produce zero.  I wonder what Julio thinks he's doing if he can't even understand the answers he gets?
>  "Or is it possible that a significant number of experiments ended up with negative z-scores, and they were misreported? "
>  Mistakes are always possible, but the table listing these results also provides the number of hits and trials in each experiment, from which it is easy to determine the hit rate and thus the z score. The results as reported look correct to me and were very carefully vetted by the original authors, and by many readers, over the last 60 years.
>  "Or not reported at all (filedrawer)? "
>  This is point of Good's comment in /Nature/, and my response. There are simple ways of calculating the number of supposedly missing studies required to neutralize reported results.

The simple ways require assuming that the studies give null results, but what if the file-drawer bias is to suppress negative, not just null, results?

Brent Meeker
 

51
Re: Update on Dean Radin - New DATA !

Wednesday, April 18, 2007 2:14 PM
From:
"Bill Jefferys" <bill@ASTRO.AS.UTEXAS.EDU>
To:
AVOID-L@HAWAII.EDU

At 9:49 AM -0700 4/18/07, Brent Meeker wrote:
>Julio Siqueira wrote:

>>  > "Is it possible that Radin got that huge z-score by not realizing
>>>that the z-scores should be signed? "
>>
>>   No, it is not possible. Z scores are distributed normally around a
>>mean of 0.
>
>This response is self-contradictory!  A variable that is normally
>distributed around a mean of zero is as likely to be negative as
>positive.  If that is the case the answer would be "Yes, it is
>possible."  Averaging a large number of such variable values would
>produce zero.  I wonder what Julio thinks he's doing if he can't even
>understand the answers he gets?

I read Radin's response differently, to wit: It is not possible that
I (Radin) did not realize that z-scores should be signed, because I
(Radin) know that they [should be] distributed normally around a mean
of 0.

To be sure, it is a bit difficult to parse Radin's reply, and I may
not have done so correctly.

I don't know what Julio thinks, though.

Bill
 
52
Re: Update on Dean Radin - New DATA !

Wednesday, April 18, 2007 4:45 PM
From:
"Julio Siqueira" <juliocbdsiqueira@YAHOO.COM>
To:
AVOID-L@HAWAII.EDU

Dear Professor Bill Jefferys,

I will follow all the "roads" as soon as I can (and as
far as my intellectual capability will let me). I do
need some time, though. Getting the book (Pratt et al,
1940) has been very tiresome in a time when I am still
ill (Strong flu or weak dengue fever? Who knows,
and... who cares :-), for I am getting better now).

Radin has been cooperative and, IMO, quite candid. I
have to proceed very carefully, especially because
each step may entail the need for careful and rather
tense (i.e. difficult) reading. (Just one example: I
have tried hard to shoehorn the VLA into a
meta-analytic framework, so far with no success. :-(
The interferometer logic seems utterly different from
meta-analytic logic. But despite that, I continue
boldly to seek any link... :-) ).

For now, unfortunatelly, all I can do is plagiarize
Swartznegger: "I will be Back!"

Best Wishes,
Julio
_____________
 
53
Re: Update on Dean Radin - New DATA !

Sunday, April 22, 2007 7:46 AM
From:
"Julio Siqueira" <juliocbdsiqueira@YAHOO.COM>
To:
AVOID-L@HAWAII.EDU

Hello Bill,
 
My comments below, preceded by [Julio] (they will appear in blue, in case you can see colors in your system.
 
>What you claim (you and Vic) is that it is not valid to combine the
>results Christ's Holy sawing. Is it Right?
 
That's Radin's claim.
 
Here's why I don't believe it.
 
I claim that there are limits to what you can accomplish by taking more and more data.
 
[Julio] Interesting point.
 
You have to remember that I have been dealing with data, real data, for over 40 years.
 
[Julio] "Real Data"... ?
 
I have learned that it is very difficult to pull a signal out of a large amount of noise.
 
[Julio] And this "signal" would be the "real data", I suppose.
 
If you have a signal that is swamped by the noise, there is only a limited amount that statistics can do for you to pull it out.
 
[Julio] Ok.
 
Amateur scientists, and I would include most of the parapsychologists I am aware of, imagine that all you have to do is to take more and more data and eventually the signal will come out, due to the magic of "the square root of N". This is an extraordinarily naive belief, and in real life it will lead to grief if carried too far.
 
[Julio] I think Jessica Utts explained this very clearly in her paper (1999), by saying that meta-analysis can be well performed or poorly performed.
 
The reason is that there are ALWAYS
 
[Julio] "ALWAYS and NEVER are two words that you should always remember never to say"...
 
unknown experimental biases in any experiment that have nothing to do with the phenomenon you are trying to investigate. They are always there, they are unavoidable.
 
[Julio] You are exaggerating too much in this piece above.
 
It is virtually certain that if you have a per-data-point error of a certain amount, then at the level of 1/100 of that error, and sometimes greater, there will be some sort of bias of this sort.
 
[Julio] Ok, let's suppose so. I just cannot figure out how this "bias" would influence the result in, say, Ganzfeld experiments. This "bias" that you talk about (in light of the highly exquisite design of the Ganzfeld protocol) seems, to me, far more amazing than ESP itself. And I am not saying that this "bias" is impossible (for if I were, I would actually be using the word "never", that I never use...). But it would be highly indicative (perhaps proof itself) of something highly interesting and relevant. That is precisely what skeptics Brugger and Taylor pointed out in their paper to the Journal of Consciousness Studies in 2003 (in the special volume "Psi Wars").
 
I know of two ways to control such biases: (1) Identify their source and measure them accurately enough to eliminate them or (2) Do a better experiment that has a smaller per-data-point error.
 
Parapsychologists do none of this. Instead, they have a touching, but naive faith in the power of the square root of N to overcome all uncertainty.
 
[Julio] I can't help saying that this that you are saying above is contrary to everything that I know about the present-day state-of-the-art in parapsychology. Maybe many parapsychologists (and their research) fit this tragic scenario that you picture. Perhaps most of them fit this scenario. But certainly not all of them. Many carry out very good experiments. Of course when I say this, it is from the point of view of someone who is outside of the parapsychological community. Maybe those who know it really well and from inside know that it is all fraud and pretence. But that is not what the papers of psi researchers lead us to believe. And more: when it comes to parapsychology, we have the help (or *supposedly* have the help...) of the organized skeptic movement (CSICOP et al). Checking the arguments from the skeptics (or better, from those who believe psi does not exist) against the arguments from the believers (i.e. psi researchers), the arguments from skeptics are very poor and weak. And I am talking about people like Hyman, Alcock and Blackmore.
 
[Julio] So, you said just above that parapsychologists (1) do not try to identify the source of biases and do not try to control them. I believe Honorton is shaking in his tomb right now (even if there is no afterlife!). His agreement with Hyman, it seems, meant nothing to some... "skeptics". As to (2) the "better experiment" with a smaller per-data-point error, I don't understand this concept. I would appreciate if you could explain it.
 
It is a good rule of thumb that you can't get a signal out of noise that is much greater than 100 times the signal.
 
[Julio] Supposing that telepathy exists, what would be, in the Ganzfeld protocol, the signal and the noise? Would the signal be the effect size of telepathy and the hit rate? Would the noise be the methodological flaws? (by the way... which ones? It has been more than ten years since any skeptic forwarded any putative methodological flaw in Ganzfeld...). Why do you believe (or has concluded) that the ratio signal/noise in this would be smaller than 1/100?
 
Maybe at the outside 1000, but that is very dicey, and would require a very successful and diligent application of method (1) above. Parapsychologists claim to be able to do much better than this.
 
[Julio] At this point, I would ask the same that I just asked above, but this time regarding the PEAR micro-PK experiments...
 
Because of my experience in working with real data, I just don't  believe them.
 
[Julio] Bill, the score is this: it may very well be that you are right. You and the CSICOPers. But what really seems is that you are dead wrong. You report the status of the psi research unfaithfully. Just like CSICOPers like Hyman and Alcock. You say they (psi researchers) do not care to improving their methodology. Psi researchers claim just the opposite. And they prove it with articles and etc. Counter arguments from you (and Hyman-Alcock-Blackmore, etc) are just nowhere to be found (regarding the improving in methodology). Maybe they will come out now... I am not forgetting the things that you said, regarding statistical flaws from the psi researchers, Radin include. Yes, that can be the link to the rotten core of psi research. But so far, honestly, it does not seem to be a real link. It just seems to be... noise.
 
Meta-analysis won't help you. All it does is to make N bigger without giving any understanding of the underlying problems that are certainly in the experiments.
 
[Julio] Again, Utts put it very clear in her paper (1999). Used correctly, meta-analyses may be of great help. Perhaps meta-analysis addresses the item 2 that you referred to (per-data-point error).
 
If Radin's meta-analysis had any merit, he could do something like the following, which is often done by real scientists: Divide his data set into a number of smaller sets of data. He's got 186 experiments, he can construct say 10 groups of 18 or 19 by randomly choosing them from the pool of 186 (or better, having someone with no axe to grind do the choosing, like a professional and neutral statistician). Then do his meta-analysis on each of the 10 groups. Since the sigma of 96.5 standard deviations is so huge, the sigmas of the individual experiments will be about 30, still huge. Therefore, if the different experiments are measuring the same thing, they all ought to produce the same effect in the same direction to quite high precision.
 
[Julio] He himself said that neither he nor his colleagues care that much about these experiments from 1940. Yes, this that you propose above is good indeed. I will forward it to Radin. If he wants to write books talking about this old data, then I think he must be ready to look at this data in the proper way. You may be surprised by what I will say, but I do not like meta-analyses. And more, I do think (feel) Radin exaggerates and misuses this tool often. Just as I criticized him in the way he used the term "consciousness" in his recent book (Entangled Minds; 2006), I feel that there is a high risk of "garbage in, garbage out" when one (Radin) uses meta-analysis in a rather lax way or ventures into meta-meta-analysis and etc. So, perhaps, I already agree with the core of your criticism to Radin (if this is the core). But I strongly disagree with you regarding the extent to which this flaw from Radin (if really a flaw) devastates the bulk of his works and the essence of his "findings."
 
I will bet that if he did this, the effects would be all over the map.
 
[Julio] Now both you and Brent Meeker can do it quite easily on your own... The book is online, and free. Anyway, I will pass to Radin your suggestion.
 
This would prove that they aren't measuring anything at all.
 
[Julio] Type II error now?
 
I know what Radin would say to this proposal. He will whine that psi is variable and unpredictable, that will be his excuse.
 
[Julio] Perhaps it is worth asking first and shooting later...
 
It's the same excuse that parapsychologists have been giving for 100 years. Look up "sheep and goat effect," "decline effect," etc. It is all excuses intended to paper over the fact that there is no there there.
 
[Julio] Come on, Bill, let's not fool ourselves. They do not claim that psi is THAT elusive. They claim that it is stable enough to be detected, and so far it seems that they are right. By the way, why did Hyman say, in 1999, that Radin and Utts were the two who made the most to establish the reality of ESP? Just courtesey? Perhaps CSICOPers should think over before saying certain things...
 
This is why I think the whole statistical-analysis-of-a-humungus-number-of-trials paradigm that has dominated parapsychology since Rhine will NEVER be convincing to those not already convinced.
 
[Julio] Science is much more than that. For example, Michael Shermer's insistant complaint is the absence of a theory for psi. It is a fact that psi is highly rejected by the academia. The reason for this is, to me, a deep mystery. I think it is greatly (but not solely) due to CSICOP's "bad fingers."
 
Parapsychologists need an experiment that passes Jimmie Savage's interocular traumatic test. I said as much in print over fifteen years ago, and suggested some ideas about how to go about it. The ideas that I suggested overcome the objection you suggest:
 
[Julio] Let me comment on your idea below:
 
By working with individual atoms at the microscopic level, it should be possible to devise clear and convincing experiments, for example telekenesis, by having people move individual atoms around on a substrate at will. Since the claims of the random event generator experiments are that people can control the decay of atomic particles (a much higher energy process), surely something simple like this would work.
 
[Julio] A good experiment. I myself, if I had the required expertise and apparatus, would put it to the test! By the way, why haven't you or any other skeptic ever attempted it? (Stanley Jeffreys tried similar protocols to PEAR ones, and ended up, some years ago, finding introductory evidence for micro-PK - He must have been severely pressured afterwards, for he even wrote an article recently in the Skeptical Inquiry where he seems to have completely wiped out of his mind any memory of his positive finding...). However... I do not know very well the nature of this quantum mechanic event that you refer to (single atoms on the move). Using my high school physics terminology (and not quantum mechanics derived, but from classical mechanics), I would point out that if the system that you are talking about is either at static equilibrium or at dynamic equilibrium, it would not be a good comparison to what is tested in micro-PK. What they test is systems that are not in equilibrium (you may here point out a misunderstanding of mine in this regard...). So, first, why would one expect that a phenomenon that operates in systems not in equilibrium would also operate in systems that are in equilibrium? Second, if the "result" would be... random, then we would go back to the same situation that they have in PEAR. Merely a shift in the probability that the event would occur. There are also further problems. But for now, these will suffice.
 
So far, I've seen no interest from the parapsychologists in doing anything like this. I got a few excuses from people who basically said, there's no point in doing this because psi might not work that way. What kind of excuse is that?
 
[Julio] The actual replies would be of great interest... But I must say that I consider it very bad that, out of the whole of the parapsychology community, no one has stepped out to try this test.
 
Instead, the parapsychologists continue on the same old treadmill, mindlessly producing large numbers based on huge numbers of trials in meaningless experiments, with no significant attention to the problems I mention above.
 
And they wonder why they can't convince anyone but the true believers!
 
[Julio] Now it is you that are playing the true believer. You believe that they can only convince the true believers. That can hardly be true. I myself am someone that just couldn't care less for telepathy or PK (micro or macro) or clairvoyance and etc. Also God is very low in my priority list (and in my agenda...). Afterlife is topmost to me, but it receives almost no support from any of these oddities of psi research. And, after carefully weighing the evidence pro and con, and the arguments pro and con, I came to consider that, scientifically, ESP is already proved. (And so what...). If the next ten years bring to us a clear disproof of ESP (which, in the last 150 years, has never happened... despite Vic' failed propaganda), I would not care at all. It is like the micro spacial dimensions. 11 or 10 or 9, to me it is just the same. And the macro spacial dimensions? Are they three or four? (and some even claim that macro spacial dimensions are not three, but two. Ok with me, as long as I do not feel "flat" :-) ). It seems really true that skeptics, too, have turned into non-skeptics. Sometimes it is said that Daryl Bem is one such example. Also, it is clear that the guys from the Journal of Consciousness Studies are pro psi. So, Bill, be a psi foe and psi fighter if you will. But do not play ostrich thinking the impact of psi is too little in academia. Well, but actually you know that far better than me...
 
Best Wishes,
Julio
______________
 

 
54
Re: Update on Dean Radin - New DATA !

Sunday, April 22, 2007 7:48 AM
From:
"Julio Siqueira" <juliocbdsiqueira@YAHOO.COM>
To:
AVOID-L@HAWAII.EDU

Hi Bill,
 
Let's go again. [Julio]
 
Date: Tue, 17 Apr 2007 09:43:31 -0400
From:"Bill Jefferys" <bill@ASTRO.AS.UTEXAS.EDU> 
Subject: Re: update on Dean Radin To:AVOID-L@HAWAII.EDU
 
At 5:28 PM -0700 4/16/07, Julio Siqueira wrote:
>Hi Bill,
>
>Some few items for you to ponder below...
>
>--- Bill Jefferys <bill@ASTRO.AS.UTEXAS.EDU> wrote:
>
>> At 2:12 PM -0700 4/14/07, Julio Siqueira wrote:
>> >
>> >"Thus antiplatelet treatment can reduce the
>> >incidence of serious vascular events by
>> >about a quarter among a wide range of patients at
>> >particular risk of occlusive vascular
>> >disease (Antiplatelet Trialists’ Collaboration,
>> >1988,
>> >p. 320)."
>> >
>> >So, IMO, it seems that this satisfies Stenger's demand
>> >(as far as I know).
>>
>> Major scientific discovery? I believe that this was
>> one of Vic's condition.
>
>Oh yes. Discoveries that help save lives are
>definetely not major discoveries. And surely "major
>discovery" is a very well defined term.

Here are some examples of major scientific discoveries:
 
[Julio] Ok, now we can get on agreement, or we can "agree to disagree," but at least we know what we are talking about. Let's see.
 
Newton's
laws, the wave nature of light, special and general relativity,
quantum mechanics, evolution by natural selection, Mendelian
inheritance, the structure of DNA, plate tectonics, the expansion of
the universe, etc. A major scientific discovery has huge
ramifications down the line, in many different fields unrelated to
the putative field of the discovery itself. It is the sort of thing
for which honors follow: prizes, election to prestigious national
academies, etc.
 
[Julio] That is *your* definition of a major discovery. I may disagree (and indeed I do). But I know what you are talking about. However, IMO, there will come the day when meta-analysis will give to us a true Jefferys/Stenger-Major-Scientific-Breakthrough. Perhaps psi is already the very first example.
 
[Julio] Anyway, talking about "major discoveries" and "prizes", below is a list from Nobel Prize Org:
 
Years 2006 to 1988.
 
"for their discovery of RNA interference - gene silencing by double-stranded RNA"
 
"for their discovery of the bacterium Helicobacter pylori and its role in gastritis and peptic ulcer disease"
 
"for their discoveries of odorant receptors and the organization of the olfactory system"
 
"for their discoveries concerning magnetic resonance imaging"
 
"for their discoveries concerning 'genetic regulation of organ development and programmed cell death'"
 
"for their discoveries of key regulators of the cell cycle"
 
"for their discoveries concerning signal transduction in the nervous system"
 
"for the discovery that proteins have intrinsic signals that govern their transport and localization in the cell"
 
"for their discoveries concerning nitric oxide as a signalling molecule in the cardiovascular system"
 
"for his discovery of Prions - a new biological principle of infection"
 
"for their discoveries concerning the specificity of the cell mediated immune defence"
 
"for their discoveries concerning the genetic control of early embryonic development"
 
"for their discovery of G-proteins and the role of these proteins in signal transduction in cells"
 
"for their discoveries of split genes"
 
"for their discoveries concerning reversible protein phosphorylation as a biological regulatory mechanism"
 
"for their discoveries concerning the function of single ion channels in cells"
 
"for their discoveries concerning organ and cell transplantation in the treatment of human disease"
 
"for their discovery of the cellular origin of retroviral oncogenes"
 
"for their discoveries of important principles for drug treatment"
 
1901 on:
 
"for his work on serum therapy, especially its application against diphtheria, by which he has opened a new road in the domain of medical science and thereby placed in the hands of the physician a victorious weapon against illness and deaths"
 
"for his work on malaria, by which he has shown how it enters the organism and thereby has laid the foundation for successful research on this disease and methods of combating it"
 
"in recognition of his contribution to the treatment of diseases, especially lupus vulgaris, with concentrated light radiation, whereby he has opened a new avenue for medical science"

That aspirin is useful in some people to reduce mortality is NOT a
major scientific discovery. It is a technical improvement in
treatment of a particular narrow condition. It is routine, everyday
science, not Nobel-prize material.
 
[Julio] "not Nobel-prize material." ... ... ...
 
I note that the actual reduction looks to be about 20%, not a quarter
(25%), as asserted by Utts (see figure below).
>> You still haven't indicated whether, ten years
>> later, this research
>> has withstood the test of time, either.
>
>And you have not indicated whether it has not.

It has been discovered since Utts' paper that aspirin prophylaxis is
not useful in some populations.
 
[Julio] Quite expectable.
 
I, for example, would be put at
greater risk from other conditions should I to use it than the
benefit I might get.
 
[Julio] Me Too (Hemorrhagic Dengue Fever).
 
Women seem not to get much benefit from it.
People who have not had heart attacks get less if any benefit. So,
while is a certain particular population the prophylaxis is still
useful, it is not universally useful.
 
[Julio] Yes.
 
In that limited sense it
appears to have stood the test of time.
 
[Julio] Hummmmm...

Now, the other claim you made is that this discovery was a result of
meta-analysis. I will now humor you briefly and PRETEND that it is a
major discovery.
 
[Julio] Yes. Time for some humor. Let's pretend this is the kind of thing that gives Nobel Prizes in physiology and medicine... :-) .
 
 I will show that the discovery was NOT because of
meta-analysis.
 
[Julio] And I will show you that you are resorting to post hoc strategies.
 
My exhibit is this figure (Figure 1 from the Utts
paper):

When you examine this figure, you see immediately that the results of
the individual studies are all consistent with each other.
 
[Julio] Bill, don't we get the same impression from the figures that I now placed on my site? (also from this paper of Utts). See the link below:
http://paginas.terra.com.br/educacao/criticandokardec/ESP_interocular_traumatic_test.htm
 
 All but
three show an improvement in mortality, and the three that do not are
all of low accuracy and are still consistent with the remaining
studies that do show this. One of the studies shows a very large
improvement, greater than the 95% error bar, but that's consistent
with the fact that there are 25 studies shown, and one can expect one
in twenty to be outside its 95% confidence interval.
 
[Julio] This, to me, looks so very similar to the graphs about Ganzfeld, both pre-PRL and PRL...
 
Anyone who
looked at this chart would conclude that it is highly likely that
aspirin is a useful treatment in these cases.
 
[Julio] The same for esp in Ganzfeld.
 
Just by eyeball alone,
 
[Julio] The same for esp in Ganzfeld.

one can see that the line that is drawn (dashed vertical line) is a
pretty good guess at what the studies would show when combined.
 
[Julio] The same for esp in Ganzfeld.
 
You
don't need meta-analysis to see this,
 
[Julio] The same for esp in Ganzfeld.!!!
 
and the people that conducted
the meta-analysis didn't need it either!
 
[Julio] The same for esp in Ganzfeld.
 
All the meta-analysis did
was to tighten the error bars and give a more accurate overall
picture of where the odds ratio is (about 0.8, a modest improvement).
Basically, all the meta-analysis amounts to is taking a weighted
average of the results along with improving the error bar by an
amount roughly proportional to 1/sqrt(N). I can get the basic result
by eyeball without doing any calculation, just from the figure.
 
[Julio] The same for esp in Ganzfeld?

And I have already pointed out to you that you can't go very far with
1/sqrt(N).
 
[Julio] (actually, I can't go anywhere with it... due to my lack of sophistication in statistics!)
 
In this case, the signal is larger than most of the error
bars, so it is an example of data that easily passes Jimmie Savage's
interocular traumatic test.
 
[Julio] And what about the graphs regarding Ganzfeld? Isn't the signal larger than most of the error bars too? If not, why not? (I can't tell the difference).
 
You don't need fancy statistics, the
result is obvious in the data themselves.
 
So I say that even before meta-analysis, people were pretty sure of
what they would find out when meta-analysis was done.
 
[Julio] The same for esp in Ganzfeld. Anyway, in both cases, they decided to do the meta-analysis. If you claim that they did it for no reason, I will find it very hard to believe...
 
 So to attribute
this result to meta-analysis, instead of just common sense, is a huge
stretch.
 
[Julio] Now "common sense" became the guiding light of science... Common sense told me, about ten years ago, that life after life was a proved fact. After reading the parapsychological literature, I got convinced that I was wrong (thanks to Ian Stevenson and his honest presentation of data and his thoughtful reasoning). Common sense assured us that black people were inferior, that women were not very much smart, and so on. Common sense told me homeopathy works (meta-analyses have proved that it does not).
 
[Julio] Let's quote Utts, to see how *she* reports this very same "common sense" that you talk about...: "Notice that very few of the individual studies were 'statistically significant' as illustrated by the fact that their confidence intervals cover the chance odds ratio of 1.0. A naive 'vote-count' of the number of studies for which the null hypothesis (that antiplatelets have no effect) could be rejected, would appear as if there was little or no use for these drugs in preventing vascular disease reoccurrence."
 
[Julio] It seems to me that at least 15 of these 25 studies were not statistically significant. Common sense...

To continue: I would bet that if a similar chart of Radin's 183
studies were drawn, it would look nothing like this figure.
 
[Julio] I do not know if I am willing to bet either way... ;-) . Anyway, I think it would be so very interesting to get this info (remember, the book is now online).
 
I would
predict that the points and their error bars will be all over the
map,
 
[Julio] So very different from Aspirin and Ganzfeld...
 
and inconsistent with each other [all the data plotted above are
consistent]. In other words, what Radin is doing, in contrast to the
above study, is pulling noise out of noise, not signal out of noise.
 
[Julio] No, dear professor. You cannot say this *until* you make your own calculations. Even in regard to this 1940 data that no one cares about. But, anyway, let me state my own position regarding this: I think these experiments from 1880 to 1940  most likely were indeed detecting news (i.e. true findings) out of noise. I think that the potential methodological flaws render this database almost useless, though still highly respectable. I think Radin should not be capitalizing, in any way, in these old experiments. At least not in this "meta-analytic" way that he engaged in. And I feel very suspicious about his meta-meta-analyses... (not that he is dishonest. Just that he, I think, may get too much signal out of the noise).
 
To summarize: This is not an example of a major scientific discovery
 
[Julio] Yes, no Nobel Prizes ever... :-) as we clearly saw from the list of (i.e. reasons for) Nobel Prize winners.

made possible by meta-analysis.
 
[Julio] Good point. It was not made possible by meta-analysis. Therefore, you just placed yourself at a conundrum: either both the antiplatelets' value and esp in Ganzfeld were not legitimized/discovered scientifically through meta-analyses, or both antiplatelets' value and esp in Ganzfeld were legitimized/discovered scientifically through meta-analyses. I let you free to choose. Either way, you "lose" (and, IMO, society wins).
 
It isn't a major scientific discovery
in the first place,
 
[Julio] Sure. We've seen that "ad nauseum"... thanks to Nobelprize.org...
http://nobelprize.org/nobel_prizes/lists/all/all_laureates_y.html
 
and even if it were, it is one that would have
been made had formal meta-analysis not or not been performed.
 
[Julio] "If", in the History of Human Societies and in the History of Science is but a sloppy gamble (that is what I meant when I said you would be engaging in post hoc excuses...). Especially in the hands of those who do not belong to the field (like you - the field of medicine and biomedical sciences). You say it would have been discovered anyway. I say, I do not know.
 
>> >Best Wishes,
>> >Julio
>> >P.S.: you can see each radio telescope in VLA as a
>> >single experiment, or data source. Obviously I know
>> >this comparison will look very forceful. But in my
>> >opinion it runs perfectly in line with the logic of
>> >meta-analysis.
>>
>>It has nothing to do with meta-analysis. If you think it does, then
>>you know nothing of meta-analysis. The VLA is an interferometer.
>>You cannot consider the results as being due to taking the data
>>separately with each telescope, getting a bunch of results, and
>>then combining them. This isn't how interferometers work. The logic
>>of interferometers has nothing to do with meta-analysis. The logic
>>of meta-analysis has nothing to do with how interferometers work.
>>
>>You are getting rather silly, and you are clearly out of your depth.
>
>I greatly doubt that emotional bursts like the one
>above can be the path to... "good science"... (or to
>faithful evaluation of... Radin's work)

Just stating the facts. No emotion!
 
[Julio] Huummmmm...

90% of everything is crap -- Sturgeon's Law.
 
[Julio] And 90% of crap is "gold" -- Clinical Bacteriologists' Law.
 
One of the major
responsibilities of scientists is to recognize crap and point it out
when they see it.
 
[Julio] That is: to be Type-II-Error Prone citizens. Very revolutionary...

>If the VLA is not a good metaphore to meta-analysis,
>then it is not. It is that simple.

If you had any notion about what meta-analysis is, you wouldn't have
suggested this example.
 
[Julio] Again, type-II-error proneness... You miss too much of the signal that is in the noise, and end up not seeing as many alternatives as it would be desireable. In the case above, you only saw one alternative (that is, my not knowing what meta-analyses are) when actually there were two... (the other one would be: Julio does not know what the VLA really is - interferometer). It so happens that the right alternative is (was) the second one. But anyway, the logic of the VLA is that you can achieve with a lot what you will not achieve with just a few. A nice metaphor. Metaphors, just like meta-analyses have their strengths and potentials, and also have their limits. Knowing our limits is the key to avoid type II errors, and to avoid going, as you put it, well beyond our depth. This time, it was you who went far beyond your depth, or, since you are an astronomer, it might be more suitable to say: far beyond your height! It just happens that it is usually safer to "surface" (in apnea) than to... "land!"
 
Best Wishes,
Julio
P.S.: according to Utts (1999), Ganzfeld meet well the bayesian challenge. According to Brent Meeker (or better, according to the thesis on Ganzfeld and bayesian approach that he indicated to me - year 2004), the impression is rather the same...
______________  
 
 
55
Re: Update on Dean Radin - New DATA !

Sunday, April 22, 2007 7:50 AM