Girl with x-ray eyes - Video

[Richard]

http://video.google.com/videoplay?docid=-3443125822374384435

Enjoy

[KingMerv00]

I say we dissect one of her eyes for the good of humanity.

What? She only needs one.

[Stray Cat]

I still haven't got over spending £2.00 on a pair of x-ray spex when I was a kid :(
Maybe if this girl had a pair of them she'd have got 7 out of 7.
Her missing the metal plate in the guy's head kind of spoke volumes to me.
But at least she put herself up to the test.

[Richard]

See how the other side saw this event.

http://www.skepticalinvestigations.org/Demkina/X-ray.htm

THE GIRL WITH X-RAY EYES

CSICOP CONVERTS VICTORY INTO DEFEAT

by GUY LYON PLAYFAIR

[ObscureReferenceMan]

Just finished watching the video. I hadn't seen it before. It's good to finally see it after reading some of the threads about it. Thanks, Richard!

[Brainache]

Whoever shot that X-ray eyes story did a great job. The camera work was excellent and the editing was none too shabby either.
I felt kind of sorry for the girl, but maybe if she does study medicine she will be able to see how she has been fooling herself.

[Stray Cat]

Whoever shot that X-ray eyes story did a great job. The camera work was excellent and the editing was none too shabby either.
I felt kind of sorry for the girl, but maybe if she does study medicine she will be able to see how she has been fooling herself.

Unless her medicine study only makes it easier for her to fleece people?

Perhaps this is a bit harsh but I can't help thinking that poverty stricken Russia is not the place a qualified doctor can make as much money as a clairvoyant diagnostician would.

As a qualified doctor she would gain more credibility but still using quackery she could do more harm than good... And let's face it, she's good at the quackery already. Imagine if it was backed up with real technical sounding medical jargon, people would fall hook, line and sinker.

[T'ai Chi]

Pretty low probability still, eh?

http://www.tcm.phy.cam.ac.uk/~bdj10/propaganda/

[asthmatic camel]

Pretty low probability still, eh?

http://www.tcm.phy.cam.ac.uk/~bdj10/propaganda/ (http://www.tcm.phy.cam.ac.uk/%7Ebdj10/propaganda/)

T'ai Chi, it seems to me that Natasha was given visual clues in both tests which would have given her the opportunity to make a well-informed guess as to the medical complaints of her "patients".

Do you not think it's somewhat suspicious that she took so much longer to make her diagnoses in the second test, and that she failed to pick up the metal plate in the skull?

She's clearly a very clever young lady, and I wouldn't be at all surprised that she has a "gift" for reading the minutiae of facial and bodily reactions of her fellow beings.

Also of interest is her mother's statement that she is never wrong; obviously, this is not the case.

50 to 1 is impressive, but given the above, hardly definitive proof that Natasha possesses psychic powers. Ultimately, who would you trust with your health care? A teenager from Moldovia or a qualified doctor?

I know which way I'd go.

[jimtron]

OMG, take a look at the guy behind her, about 18-20 seconds into the program. He appears to be moving his scalp with PK power. I don't know about Natasha, but this guy's really got the goods!!


:jaw-dropp

[Mercutio]

Pretty low probability still, eh?

http://www.tcm.phy.cam.ac.uk/~bdj10/propaganda/

So, if we consider this sufficiently low probability to reject the null, what do we conclude?

[Jeff Corey]

Moose and squirrel must die?

[davidsmith73]

So, if we consider this sufficiently low probability to reject the null, what do we conclude?

Not much on the basis of that test because very few control measures were put in place to prevent known means of aquiring information about the patients. The thing that always bugs me about these CSICOP and JREF tests is this "preliminary" testing that, as is the case here, gives the so-motivated sceptic the freedom to cast reasonable doubt on a result that may show something interesting and worthy of further study. Very sneaky.

However, if the proper controls were put in place and we got the same result then I would conclude that she had probably aquired accurate information about the patients by some means other than those that were controlled for in the test. Surely that would arouse some interest into finding out what that mechanism is?

[T'ai Chi]

So, if we consider this sufficiently low probability to reject the null, what do we conclude?

Hard to say. More, better, tests might be needed.

One can't conclude what Hyman, Skolnik, et al want to.

[T'ai Chi]

Do you not think it's somewhat suspicious that she took so much longer to make her diagnoses in the second test, and that she failed to pick up the metal plate in the skull?


Where did she claim to be able to see metal plates?


Also of interest is her mother's statement that she is never wrong; obviously, this is not the case.


Again, one should test based on performance, not on what they (quite possibly mistaknegly) believe about their performance.

A good basketball player may say 'I never miss a 3 point shot!', when tested he might make 8 out of 10, which is excellent, but by a naive testing of 100% "claim" he'd fail the test.


Ultimately, who would you trust with your health care? A teenager from Moldovia or a qualified doctor?


My current doctor of course. Luckily, my feelings about what I'd choose for healthcare don't have any importance on testing someone's claim, nor should they.

[CFLarsen]

Where did she claim to be able to see metal plates?

She claimed she can see into people, even on a molecular level. If she can see the blood stream and intestinal organs, how on Earth can't she see a metal plate? Why didn't she object when the experiment started?

You need to explain this.

Again, one should test based on performance, not on what they (quite possibly mistaknegly) believe about their performance.

A good basketball player may say 'I never miss a 3 point shot!', when tested he might make 8 out of 10, which is excellent, but by a naive testing of 100% "claim" he'd fail the test.

Sorry, but we have to go with what people claim. We can't just override what they say.

My current doctor of course. Luckily, my feelings about what I'd choose for healthcare don't have any importance on testing someone's claim, nor should they.

No? That's interesting, considering that you have been pointing to the low probability as a reason why there is validity to her performance.

And yet, you won't use her services. You don't practice what you preach.

[T'ai Chi]

What makes you believe that?

[Jeff Corey]

About the same evidence that it would take me to go to a psychic surgeon to remove my hemorhoids.

[CFLarsen]

What makes you believe that?

She claimed she can see into people, even on a molecular level. If she can see the blood stream and intestinal organs, how on Earth can't she see a metal plate? Why didn't she object when the experiment started?

You need to explain this.

[Jeff Corey]

I'll bet he doesn't.

[blutoski]

Pretty low probability still, eh?

http://www.tcm.phy.cam.ac.uk/~bdj10/propaganda/

The key questions are:

1) what does the person claim they can do?
2) what would failure look like?
3) what would an alternative technique look like?

In this case:

1) claims 100% accurate visibility into human body, as if having an x-ray or ct machine in eyes. claims never makes a mistake. makes a living off telling people whether or not they have tumours ranging from microscopic to quite large. charges quite a bit of money for this (one month's salary). Specifically, her literature states: "..can see abnormalities down to the cellular level and that she never ever misses."

2) pass would be 7/7. failure would take the form of missing one or more very obvious missing organs or pieces of skeleton. However, for this demonstration, she was given *extra latitude* and could miss two (5/7) and still 'pass'.

3) an alternative technique would be a) random guesses, b) educated guesses, c) cheating or d) a mix of these. they would score differently, depending on technique, so the 5/7 pass still seems reasonable, as three alternative techniques would be expected to score higher than chance, but only cheating would be expected to get a full score.

She scored 4/7.

It should be noted that Demkina's mother (who is also her public relations manager, is the person who hatched the challenge and contacted the production team via Demkina's agent) originally objected to lowering the standard from 7/7 to 5/7, as it was 'insulting' to imply that her daughter might fail. Off the top of my head, my impression is that she did what a lot of naiive people do when they realize that a 'test' is actually a 'test': she backpedalled and took what she could get, with a backup plan to kvetch after. Demkina is used to doing 'tests' on Russian television that are really 'demonstrations'.


Notice that Scientists' unethical use of media for propaganda purposes does not link to the skeptics' rebuttal of this rumour-mongering: Natasha Demkina, The Girl with Very Normal Eyes (http://web.archive.org/web/20050204152650/http://www.csmmh.org/demkina/answerstocritics.html)




There's a name for making a plan like this. I'm talking about a plan where, if you win, you declare victory, but if you lose, you declare victory by being cheated.

It's called a 'double-cross'.

[CFLarsen]

I'll bet he doesn't.

And we don't need odds on this one.

[blutoski]

And we don't need odds on this one.

What's sad is that this schmozzle really boils down to conflating two different issues: they failed the challenge, but the challenge was not intended to test for an 'anomaly.' Nobody denies that she did better than chance. Just that this was not the terms of the challenge, so yes, she did fail the challenge. Does she have some kind of ability? Maybe. I think it's worth a test!

If they want to change their claim to something like: "she has some ability to see such-and-such" then hey, apply for the JREF challenge, and the threshold will be negotiated to something everybody will be happy with. Maybe even random chance, who knows? But be prepared for a tighter protocol. (ie: no more peeking at the subjects in advance)

[blutoski]

What's sad is

Actually, what I was starting out to say is that the sad thing is that Wiseman has been ejected from the good graces of the paranormalists. He used to have a much closer relationship with paranormal experimenters, but is now pretty much unable to collaborate.

[Richard]

Let me play the part, if just for a moment, of a real hard skeptic.... You are saying she can see through people's bodies using her x-ray eyes!! Yeah, right. Well, you never know, but come on...

[Richard]

Let me play the part, if just for a moment, of a real hard skeptic.... You are saying she can see through people's bodies using her x-ray eyes!! Yeah, right. Well, you never know, but come on...

Yeah I know, I know... we owe it to science to test the claims.

[The Atheist]

About the same evidence that it would take me to go to a psychic surgeon to remove my hemorhoids.
That's an awful lot of evidence.

[Euromutt]

Please, before posting in this thread, refer to the previous thread (http://forums.randi.org/showthread.php?t=45357) on Natasha Demkina, all 62 pages of it. We've seen most of the "arguments" in her favor repeated there ad nauseam already, and they didn't hold up then either.

[davidsmith73]

What's sad is that this schmozzle really boils down to conflating two different issues: they failed the challenge, but the challenge was not intended to test for an 'anomaly.' Nobody denies that she did better than chance. Just that this was not the terms of the challenge, so yes, she did fail the challenge. Does she have some kind of ability? Maybe. I think it's worth a test!


I agree that there are two issues here, as you said. But here's an interesting question:

What is the motivation of CSICOP of the JREF if they are only interested in seeing whether the challenger succeeds or fails according to some arbitrary set of rules? Why are they not interested in any hint of an ability that falls short of passing the test?

[T'ai Chi]

1) what does the person claim they can do?


Say I claim I can get 100 calls of a coin right out of 100 trials, and my mom says 'Yeah, T'ai chi can!'.

I 'only' get 85. Therefore I fail the test, declared to have 'no powers' by the organized skeptical movement.

However, my actual performance, 85 out of 100 is incredible, and deserves futher study to see if actual skill or a fluke.

In other words, one should test performance, not perception of performance, becuase of the obvious fact that people like to 'talk up' their abilities, or in general, may be mistaken about their true ability.

[T'ai Chi]

We've seen most of the "arguments" in her favor repeated there ad nauseam already, and they didn't hold up then either.

But then, you wonder why many in the organized skeptical movement got so flustered and aggressive... you know, if there was nothing to these critiques of their arguments and evaluations of the outcome?

;)

[CFLarsen]

But then, you wonder why many in the organized skeptical movement got so flustered and aggressive... you know, if there was nothing to these critiques of their arguments and evaluations of the outcome?

;)

Who has said that?

[T'ai Chi]

One may believe that... but then one would be missing the boat, so to speak.

[CFLarsen]

One may believe that... but then one would be missing the boat, so to speak.

Huh? What are you talking about?

Just answer the question, instead of obfuscating.

[T'ai Chi]

Some people are inclined to believe that, apparently.

[CFLarsen]

Believe what??

[Jeff Corey]

But then, you wonder why many in the organized skeptical movement got so flustered and aggressive... you know, if there was nothing to these critiques of their arguments and evaluations of the outcome?

;)

Apparently you missed Claus' question as well as the proverbial boat. Who got flustered and aggressive?

[T'ai Chi]

Say I claim I can get 100 calls of a coin right out of 100 trials, and my mom says 'Yeah, T'ai chi can!'.

I 'only' get 85. Therefore I fail the test, declared to have 'no powers' by the organized skeptical movement.

However, my actual performance, 85 out of 100 is incredible, and deserves futher study to see if actual skill or a fluke.

In other words, one should test performance, not perception of performance, becuase of the obvious fact that people like to 'talk up' their abilities, or in general, may be mistaken about their true ability.

Wasn't really addressed yet.

[Jeff Corey]

You still haven't answered the question. Specifically, who got "flustered and aggressive"?

[CFLarsen]

Say I claim I can get 100 calls of a coin right out of 100 trials, and my mom says 'Yeah, T'ai chi can!'.

I 'only' get 85. Therefore I fail the test, declared to have 'no powers' by the organized skeptical movement.

However, my actual performance, 85 out of 100 is incredible, and deserves futher study to see if actual skill or a fluke.

In other words, one should test performance, not perception of performance, becuase of the obvious fact that people like to 'talk up' their abilities, or in general, may be mistaken about their true ability.

Wasn't really addressed yet.

OK, we'll go with this line of reasoning:

Since Demkina got 4 right, she must obviously be able to see into people. If she couldn't, she would get far less. Correct?

Which means you accept that she has paranormal powers.

[blutoski]

I agree that there are two issues here, as you said. But here's an interesting question:

What is the motivation of CSICOP of the JREF if they are only interested in seeing whether the challenger succeeds or fails according to some arbitrary set of rules? Why are they not interested in any hint of an ability that falls short of passing the test?

For starters, this wasn't a CSICOP or JREF situation. It was a publicity stunt for the Demkinas. They wanted an on-TV challenge, and the producers found some skeptics with experience testing claims. Wiseman and Hyman were associated with CSICOP, but not representing them. I don't think any of the skeptics involved were associated with JREF.

The rules weren't "arbitrary": they were testing the claim as stated, which was the mandate given them by the television program's producers.

The Demkinas are welcome to go to JREF and take the challenge on their own terms.

[Jekyll]

Say I claim I can get 100 calls of a coin right out of 100 trials, and my mom says 'Yeah, T'ai chi can!'.

I 'only' get 85. Therefore I fail the test, declared to have 'no powers' by the organized skeptical movement.

However, my actual performance, 85 out of 100 is incredible, and deserves futher study to see if actual skill or a fluke.

In other words, one should test performance, not perception of performance, becuase of the obvious fact that people like to 'talk up' their abilities, or in general, may be mistaken about their true ability.

Wasn't really addressed yet.

Well that's because it's not relevant to what's going on.

No one thinks that her performance was a fluke,so let's rewrite your little scenario.

You perform substantially better at cards than would be expected from chance alone. You claim that this is because you're psychic but we all know you're a bit of an amateur magician, so we tighten the controls and surprise, surprise your success rate(which you claimed to be able to maintain because you've got magic powers) drops off.

This doesn't mean that we're catching all the times you cheat, just stopping some of them, so of course we don't expect your success rate to drop all the way to the level chance alone would suggest.

This is why Mercuitio was asking you what it meant to reject the null hypothesis.
It means she did significantly better than chance leads us to expect, but everyone thought that was going to happen anyway, so it's no big deal and actually irrelevant to the important question:

Did she perform the way she should if she had the psychic powers she said she had?

:notm

[blutoski]

Say I claim I can get 100 calls of a coin right out of 100 trials, and my mom says 'Yeah, T'ai chi can!'.

I 'only' get 85. Therefore I fail the test, declared to have 'no powers' by the organized skeptical movement.

Whoah, whoah... this wasn't the conclusion. The conclusion was that she failed the challenge, and merits further investigation. She has been invited to apply for the challenge on her own terms, but declined.



However, my actual performance, 85 out of 100 is incredible, and deserves futher study to see if actual skill or a fluke.

Agreed. Yet... she's not interested. So there ya go.



In other words, one should test performance, not perception of performance, becuase of the obvious fact that people like to 'talk up' their abilities, or in general, may be mistaken about their true ability.

This doesn't make sense. Are you saying that if a person makes a claim, we can ignore it and test something we think, in our opinion, they should be claiming? If they fail such a test, don't you think this gives them the ultimate out? ("They tested something that I don't claim to do, setting me up for failure.")

[davidsmith73]

For starters, this wasn't a CSICOP or JREF situation. It was a publicity stunt for the Demkinas. They wanted an on-TV challenge, and the producers found some skeptics with experience testing claims. Wiseman and Hyman were associated with CSICOP, but not representing them. I don't think any of the skeptics involved were associated with JREF.

My mistake. But I was really making a general comment on the motivation of CSICOP and JREF when it organises tests of psychic claims.


The rules weren't "arbitrary": they were testing the claim as stated, which was the mandate given them by the television program's producers.

What I meant was that the rules of these tests (whether it be independent groups who have a link with CSICOP and JREF, of CSICOP and JREF themselves) change according to the claim of the testee. And thats fine if the interest of the sceptic is simply to see if the testee succeeds or fails in the test.

But as I said, what motivates these kinds of sceptics? Are they not interested in following up an experiment that just falls short of succeeding?

[blutoski]

Whoah, whoah... this wasn't the conclusion. The conclusion was that she failed the challenge, and merits further investigation. She has been invited to apply for the challenge on her own terms, but declined.

I should point out that the statements by Hyman et al stating that there's probably nothing there were their personal opinion and speculation about what would happen if we actually tested her for 'ability' alone, as opposed to 'flawless ability'. That's not the 'official conclusion'. I'm not sure there was any official statement about whether she had any 'ability'. The challenge wasn't designed for that.

I would also like to echo the above poster's (Jekyll) point: your example of coin flipping is not a good analogy because she was almost certainly not 'randomly' guessing. The protocol had significant holes, and there is evidence she was making educated guesses based on visual cues and had cheated, which makes her score even less impressive.

[blutoski]

What I meant was that the rules of these tests (whether it be independent groups who have a link with CSICOP and JREF, of CSICOP and JREF themselves) change according to the claim of the testee. And thats fine if the interest of the sceptic is simply to see if the testee succeeds or fails in the test.

It's more than that: it's a guarantee that we're not testing for the wrong ability or claim. Skeptics are careful to focus their work on actual claims instead of wasting time knocking down strawmen. It's in everybody's interest: the claimant is not being misrepresented, the JREF is not wasting time testing something only to find out later that the claimant announces we've tried to pull a fast one (ie: moving goalposts), and science benefits if it turns out the claim is supported.



But as I said, what motivates these kinds of sceptics? Are they not interested in following up an experiment that just falls short of succeeding?

Yes! That's the purpose of the challenge! But you have to have the cooperation of the claimant. We can't test Natasha without her cooperation - it would be unfair to test her unilaterally, and the only thing worse would be to do it without matching the test to her actual claim. This is the purpose of the challenge: when they want to participate in a bona-fide scientific test, they just have to fill out the form. It's on our dime.

They never do, though.

[T'ai Chi]

Well that's because it's not relevant to what's going on.


Fortunately, it remains relevant.

Because anybody can plainly see that one should evaluate results based on performance, not on pre-result ideas about future performance.

[davidsmith73]

Fortunately, it remains relevant.

Because anybody can plainly see that one should evaluate results based on performance, not on pre-result ideas about future performance.

Exactly. Which makes me wonder what the sceptical motivation is for these pre-arranged success vs failure tests.

[CFLarsen]

How should Demkina be tested, then?

[Moochie]

Say I claim I can get 100 calls of a coin right out of 100 trials, and my mom says 'Yeah, T'ai chi can!'.

I 'only' get 85. Therefore I fail the test, declared to have 'no powers' by the organized skeptical movement.

However, my actual performance, 85 out of 100 is incredible, and deserves futher study to see if actual skill or a fluke.

In other words, one should test performance, not perception of performance, becuase of the obvious fact that people like to 'talk up' their abilities, or in general, may be mistaken about their true ability.

Gamblers sometimes have "incredible" runs of luck -- do they have paranormal ability?

If a doctor wrongly diagnoses 3 out of 7 times I would report her/him to the relevant authorities.

I say shame on this girl's mother and her retinue of bottom feeders for allowing this event to take place. It's more befitting a circus sideshow.

M.

[Jekyll]

Fortunately, it remains relevant.

Because anybody can plainly see that one should evaluate results based on performance, not on pre-result ideas about future performance.

Ok then. We throw out your "pre-result ideas about future performance", namely any comparison with the null hypothesis.

What we're left with is a subject who's magical power miraculously decrease when controls to make it harder to cheat are present.

What do you think Occam's razor suggests in this case?

[Mojo]

She wasn't really trying, or she was feeling stressed, or she never claimed to be able to do what she has been reported as claiming, even on her own website, or appendices can grow back in Russia, or appendices can look like worms, or appendices can roam freely about the abdomen, or the people doing the test should have been looking for x-rays, or psychic abilities are actively evasive, or...

;)

[blutoski]

Fortunately, it remains relevant.

Because anybody can plainly see that one should evaluate results based on performance, not on pre-result ideas about future performance.

Mm. But performance must be compared to a null hypothesis, right? In the example you gave with coin-flipping, you can recognize ability by comparing to random chance, because that's a reasonable alternate explanation. The experiment is intended to resolve between two possibilities.

In Demkina's case, then, sure, we could re-test. But the comparison would have to be other people making educated guesses, not random chance. If ordinary people do 4/7 in the same situation, it would be reasonable to conclude that she's making educated guesses.

There are entire magic tricks that rely on this type of thing. Some things are not random when humans are involved. We have prejudices.

I'll give you an example: which patient has the metal plate in his head? I'd pick the one patient with a hat on. That's exactly what she did (and she got it wrong). Which patient has no medical conditions? I'd pick the young athlete, considering most of the conditions listed were typical of old people. (she got this one right).

I think most people would have 'guessed' the same. I predict clustering the way you get with random numbers, favourite colours, &c.

To be frank, with all the protocol violations, we should just toss the results. But there was a show to broadcast, so there ya go.

[Mercutio]

If a doctor wrongly diagnoses 3 out of 7 times I would report her/him to the relevant authorities.
This is a crucial point, and deserves repeating.

The obtained probability in this demonstration may well have been "sufficiently low probability to reject the null" for some people; others, though, will hold to a stricter criterion. Alpha level is not automatically .05, after all; it must depend on the costs of type I and type II error.

[Jeff Corey]

That is also why the previous example of getting 85 heads out of 100 tosses was disingenuous. The Binomial Test yields a p<.000000001.

[T'ai Chi]

Gamblers sometimes have "incredible" runs of luck -- do they have paranormal ability?


If they do consistently, significantly beyond what chance would dictate, they they should be studied more.

(read that sentence very carefully)

[T'ai Chi]

What do you think Occam's razor suggests in this case?

I'm interested in actual performance, not ideas about performance, and not OR philosophy.

[Mercutio]

I'm interested in actual performance, not ideas about performance, and not OR philosophy.

Have you contacted her yet about doing more testing?

[Jeff Corey]

If they do consistently, significantly beyond what chance would dictate, they they should be studied more.

(read that sentence very carefully)

You should write more carefully. Check out the "edit" button.

[blutoski]

If they do consistently, significantly beyond what chance would dictate, they they should be studied more.

(read that sentence very carefully)

I disagree. The alternative explanation is not always 'chance'.

I was walking on the road the other day, and all the cars were driving on the right. The odds of all these cars driving on the right by pure chance is incredibly low. However, this isn't a supernatural event: people are in control of the cars, and there's a perfectly good explanation for my observation. No need to pick up the phone and dial Randi.


What we would want to compare with is a baseline that's appropriate for the type of phenomenon. In the case of Natasha, a baseline would be educated gueses, since that's completely within her ability as a human being with a brain and years of practice.

[CFLarsen]

If they do consistently, significantly beyond what chance would dictate, they they should be studied more.

(read that sentence very carefully)

And you have made it clear that Demkina should be studied more.

Please point to the results where Demkina has performed "consistently, significantly beyond what chance would dictate".

[Flange Desire]

Say I claim I can get 100 calls of a coin right out of 100 trials, and my mom says 'Yeah, T'ai chi can!'.

I 'only' get 85. Therefore I fail the test, declared to have 'no powers' by the organized skeptical movement.

However, my actual performance, 85 out of 100 is incredible, and deserves futher study to see if actual skill or a fluke.

In other words, one should test performance, not perception of performance, becuase of the obvious fact that people like to 'talk up' their abilities, or in general, may be mistaken about their true ability.
No.

[Jekyll]

I'm interested in actual performance, not ideas about performance, and not OR philosophy.
You should know that you are dismissing the whole of statistics when you say that.

Your null hypothesis is just an "idea about performance" and nothing more.

As I said in my last post, the facts of the actual performance are that her magical abilities decrease when tighter controls are imposed.

That's it. That is all we can know without using these "ideas about performance", we can't know that she performs "significantly beyond what chance would dictate". You are consistently shooting yourself in the foot with this ill-thought argument.

[CFLarsen]

You should know that you are dismissing the whole of statistics when you say that.

Which is interesting, given the professional background of T'ai's: Statistics.

Now....why would someone dismiss his own field of knowledge (which he prides himself with) when searching for a paranormal phenomenon.....?

[T'ai Chi]

Have you contacted her yet about doing more testing?

Have you?

[T'ai Chi]

You should know that you are dismissing the whole of statistics when you say that.


In your (wrong) opinion only.

I'm saying I'm more interested in data than theory. What don't you understand about that?


, we can't know that she performs "significantly beyond what chance would dictate".


Um, of course one can, since we can calculate the probabilities of getting r out of n 'hits'. That was the whole idea of the experiment by Hyman et. al. Else she just gets r out of n and no one can say if getting r is 'good' or 'bad'.

[Jekyll]

In your (wrong) opinion only.
Please tell me how I'm wrong. Explain just how the null hypothesis you were so anxious to test against is not an "idea about performance".

I'm saying I'm more interested in data than theory. What don't you understand about that?
This makes it all the more confusing that you're not responding to my point that the data shows that her abilities diminish with tighter controls.

Um, of course one can, since we can calculate the probabilities of getting r out of n 'hits'. That was the whole idea of the experiment by Hyman et. al.
Of course you can and should do this but it is based entirely upon your "ideas about performance".
Else she just gets r out of n and no one can say if getting r is 'good' or 'bad'. Which is exactly why your claim that that "one should [not] evaluate results based on... pre-result ideas about future performance." is so stupid.

[T'ai Chi]

Please tell me how I'm wrong.


This was already done. I humbly suggest you try understanding what I wrote a little harder.


This makes it all the more confusing that you're not responding to my point that the data shows that her abilities diminish with tighter controls.


A simplistic "diminish" isn't detailed enough to make a judgement. One has to mathematically state what your terms mean. Will you?


Which is exactly why your claim that that "one should [not] evaluate results based on... pre-result ideas about future performance." is so stupid.

Not at all, you simply persist in your misunderstanding.

[Jekyll]

This was already done. I humbly suggest you try understanding what I wrote a little harder.
Which post number?


A simplistic "diminish" isn't detailed enough to make a judgement. One has to mathematically state what your terms mean. Will you?
Sure, "Her supporters vouch for the accuracy in her typical diagnoses (her mother even claimed that she never errs). Natasha informed the producer that our proposed test would be much less than demanding than her typical reading."
From http://www.csicop.org/specialarticles/natasha2.html
4/7 is at best 57% accurate, it is considerably less impressive when you consider the inter-related nature of the claims. I'll be generous and ignore that because I can't be bothered with the working.

Elimination of unfalsifiable claims and many cold reading techniques reduced her apparent absolute accuracy by something in the region of 40%.

Unfortunately, I can't analyse the data any further without using "ideas about performance".:rolleyes:

[T'ai Chi]

No rational reason given yet for why one should test judgements about ability rather than actual performance.

[CFLarsen]

The arrogance is quite staggering, isn't it?

[blutoski]

The arrogance is quite staggering, isn't it?

It's not arrogance... that last post was gibberish.

[monoman]

If they do consistently, significantly beyond what chance would dictate, they they should be studied more.

(read that sentence very carefully)

Firstly, i think Natasha and her representatives would refuse to have her tested again since she obviously struggled under the given conditions.

Secondly, if i was the tester and Natasha got 4 out of 7 i'd like to test her again......except she missed a bleeding metal plate in someones head!

[Mercutio]

Have you?

I asked you first. Besides, you seem much more passionate about the subject. If you like, though, I would be happy to assist you in designing a better experiment. I think you will agree that blutoski and Jekyll each have relevant comments about the adequacy of a simple a priori probability model as comparison. There would be no reason for you to repeat that mistake.

[blutoski]

Firstly, i think Natasha and her representatives would refuse to have her tested again since she obviously struggled under the given conditions.

As mentioned above, she has been invited and has refused.



Secondly, if i was the tester and Natasha got 4 out of 7 i'd like to test her again......except she missed a bleeding metal plate in someones head!

I think that the post-mortem that resulted has been so confusing that we have to revisit this protocol.

T'ai Chi is not unusual in that many of my peers also fail to understand that we're not dealing with random chance as a baseline for these events. Humans can do better than chance all the time just by making educated guesses. If you line ten people up and ask "Which are boys and which are girls," it's not too hard to sort them out. If you ask "Which one has a metal plate in his head," it's a good shot it's the guy with the hat.

This is subtle, and is the primary reason that scientists are not especially qualified to examine these claims: we need magicians, because the fourth legitemate hypotheses is cheating. James Randi's psi mole demonstration was proof of this.

A good protocol should not only have the resolving power necessary to be decisive from a professional point of view, but should also be decisive from a layperson's point of view.

For example, an alternative protocol would be to select patients with the same condition, instead of giving her a chance to slot them based on their visible differences (eg: assigning the missing leg to the woman who limped - duh)

The condition in question should have no external cues.

Ideally, the patients should not know about their condition.

There should also be more than seven patients.

A scenario like this would be to have patients who are going in for CT testing for, say, pancreatic cancer, and seeing if she can detect tumours in those who have them, and not detect tumours in those who do not.

This would satisfy the "compared to chance" baseline T'ai Chi is seeking.

[blutoski]

T'ai Chi is not unusual in that many of my peers also fail to understand that we're not dealing with random chance as a baseline for these events.

I should point out that this is a problem that extends beyond psi: the misidentification of baselines is also the key argument behind ID. ("It's impossible for all those to have come together by chance" - well no scientist claims it was chance... they consider evolution to be a mechanism that operates differently from chance.)

[Jekyll]

No rational reason given yet for why one should test judgements about ability rather than actual performance.
I'll tell you what; I'll frame the whole thing in frequential stats and hypothesis tests just to make it easier for you.

H0: Neither Natasha nor her mother are liars or cheaters. The results of the test were consistent with her claimed everyday accuracy.

H1: H0 is false.

Obviously, if we do actually take Natasha's mother at her word we find that there is a probability of 0 that the null hypothesis is true.

Let's be generous, we'll assume that when N.'s mother said 100% accuracy she only meant 97%.

So we calculate the probability of getting that 4/7 or worse based on 97% accuracy:

p=Sum(i=0...4) .97^i+(1-.97)^(7-i)*nCp(7,i)
=8.63*10^-4

Oh, that's no good, we're bound to be rejecting H0 for pretty much any alpha. Let's try again with 95% accuracy.

P=3.76*10^-3.

92% accuracy?

P=1.40*10^-2.

At 91% accuracy (9/10 is not something that can ever be described as 100% accurate) you get p=1.93*10^-2 which is a result comparable to the p-value Brian Josephson was talking about in your first link.

And you have to drop Natasha's "never wrong" accuracy all the way down to 87% in order to find a p-value that lies within your 95% confidence interval. At this point, you are already invalidating H0 with your assumptions that neither she nor her mother can be trusted to tell the truth about her abilities.

In which case, why would you test her?

[Spektator]

It's not arrogance... that last post was gibberish.

Aaaahhhhhhhhhh yes! But as gibberish goes, it was quite arrogant, wasn't it?

[BillC]

Have you?I wrote to her in both Russian and English last year, at a time when she had a personal website, but a very much more limited one than at present. I asked her if she still carried out readings. I received no reply.

[T'ai Chi]

I asked you first.


Have you?