Author Topic: Podcast #63 - Komsomolskaya Pravda - xray girl  (Read 7694 times)

0 Members and 1 Guest are viewing this topic.

Offline rbairos

  • Off to a Start
  • *
  • Posts: 94
Podcast #63 - Komsomolskaya Pravda - xray girl
« on: October 12, 2006, 10:51:57 AM »
Hi Guys,

I listened to podcast 63, where Rebecca makes reference to a russian girl with x-ray eyes. Apparently this girl scans the body for medical anomalies with a high degree of accuracy.

The first article I came across debunking the power was:

http://www.livescience.com/othernews/reason_demkina_050128.html

In particular:
Quote

We recruited six volunteers, who each had a different medical condition visible on X-rays, plus a "normal" subject who had none of the six target conditions
So the test -- which required her to match at least five of the target medical conditions to the correct subjects -- should have been a breeze. She didn't have to scan their entire bodies for unknown conditions. She was told exactly what to look for and exactly where to look. Yet, it took her more than fours to complete the test and she only matched four of the conditions correctly -- a score that everyone prior to the test had agreed upon would not justify further testing.


Now, Im not implying Natasha really has xray eyeballs, but isn't matching 4 out of 7 statiscally very signifcant ?  
My math may be incorrect, but I give the odds of matching 4 or more out of 7 odds of 50 : 1.  (can anyone confirm?)

This being the case, isn't it a little dishonest to ignore the results outright?


Rob Bairos
Toronto Canada.

Offline Joe Shmoe

  • Chief Rape Apologist
  • Stopped Going Outside
  • *******
  • Posts: 5347
Podcast #63 - Komsomolskaya Pravda - xray girl
« Reply #1 on: October 12, 2006, 11:12:43 AM »
Not if they identify the location of the body and she knows that any illness they send at her would be visible on an xray 4 out of 7 is roughly 50%.  Furthermore an xray technician has a far better record at identifying the problem.

I still don't like the test, it's almost impossible to properly blind this test since most patients know what's wrong with them and the administrators of the test obviously knew the malady affecting each patient (in order to identify the body part).

Offline rbairos

  • Off to a Start
  • *
  • Posts: 94
Podcast #63 - Komsomolskaya Pravda - xray girl
« Reply #2 on: October 12, 2006, 11:24:45 AM »
I dont see getting 4/7 correct,  as about 50% accuracy.

She has 7 subjects, and 7 problems.  She has to match up the
7 problems with the 7 subjects.  

There are 7! (5040) ways of assigning (ordering) the 7 problems.
Getting 7/7 is only a 1 in 5040 chance (statiscally speaking).

I dont have the rest of the numbers with me (wrote up a little
brute-force permutation program) but getting 4 or more right was under 2%.

Offline Freeman74

  • Keeps Priorities Straight
  • ***
  • Posts: 263
    • http://www.myspace.com\freeman74
Podcast #63 - Komsomolskaya Pravda - xray girl
« Reply #3 on: October 12, 2006, 12:11:12 PM »
She could tell other information about the individuals that could give clues to their condition. A 20 year old probably doesn't have a heart problem for example. Women are also less likely to have heart problems makeing an older gentleman the most likely canidate.

I read the study a while ago. I don't have time to refresh myself so what I'm saying may not be 100% accurate.

I beleive that 2 men in the study had their appendix removed (One unknown to the people who setup the test) increasing her chance of guessing one.  But she missed both, Pretty crappy x-ray vision to miss an organ one of the more obvious and should have been easy to diagnose issues. But interestingly since the subjects were clothed there would be no outward signal to help guess that one.

Subtle clues could help eliminate or make better educated guesses.

A better test would have been a group of 10 people.

Some have the issue some do not.

She simply has to identify the ones with the issue.

The experiment is blinded by the subjects being choosen by a non participating third party.

The subjects would have been put in a lineup unaware of what they were being evaluated for.

The issue would be easily apparent on an x-ray and not outwardly obvious.

The girl would have know the issue she had to identify.

This would have taken a lot of the ability to read subtle clues out of the experiment.

It was a poorly conducted test but in my opinoin shows that the girl was just using some educated guessing.
quot;Phfft! Facts. You can use them to prove anything." - Homer Simpson

Offline rebecca

  • Seasoned Contributor
  • ****
  • Posts: 595
    • http://www.skepchick.org
Podcast #63 - Komsomolskaya Pravda - xray girl
« Reply #4 on: October 12, 2006, 12:11:38 PM »
You all may be interested to first read this summation of the events, including a point by point rebuttal of criticisms as well as a mention of the one or two criticisms that most true believers have failed to point out.
Skepchick <- Now with updated URL for 2012!

Offline rbairos

  • Off to a Start
  • *
  • Posts: 94
Podcast #63 - Komsomolskaya Pravda - xray girl
« Reply #5 on: October 12, 2006, 12:41:48 PM »
Quote
The investigators raised the bar so that Natasha would fail. The odds of Natasha getting four or more correct matches out of seven was 1 in 50 (or 2 percent).


woohoo I was right.  Now I can go back to what Im supposed to be doing during the day ;)

The article does point out the great lengths the researchers went to in controlling the test, but I have a hard time accepting 1 in 50 is a statisically irrelevant result for a preliminary test, no matter how extraordinary the claim.   I wouldnt chalk it up as a major victory for the skeptical community.

It leads me to think that cold reading is much more effective than many investigators realize, and even greater measures have to be taken to bring results back down to random levels. (ie, about 2/3 chance of getting at most one correct).  If I were the investigators, I would have been sweating each correct answer..

Rob.

Offline Steven Novella

  • SGU Panel Member
  • Well Established
  • *****
  • Posts: 1833
    • http://www.theskepticsguide.org
Podcast #63 - Komsomolskaya Pravda - xray girl
« Reply #6 on: October 12, 2006, 01:03:37 PM »
You actually hit the nail on the head - the positive results were likely due to cold reading (although, 1/50 is not remarkable for pure chance either). The researchers knew that. The bar was set high because they figured she was good at guessing medical ailments - she was a good medical cold reader. Therefore her guessing would not be at chance.

You cannot control for cues in this situation the way you could if you were testing a psychic's ability - because the very physical presence of the subject would provide medical cues to their probable diagnosis.

So, they could have had a much higher number of subjects, and taken more extreme measures to avoid cold reading - or they could just set the bar high enough to account for some non-random guessing.

This is appropriate because of the girl's claim. She was not claiming to get vague impressions, or a psychic sense of their diagnosis. She claimed she had X-ray vision and therefore could see directly into a person's body and see their ailment. If her claim were correct, then she should have gotten 7/7 correct easy. They gave her some wiggle room, I think 5/7 would have still "passed," but not 4/7. And keep in mind, she agreed to the threshold before hand.
Steven Novella
Host, The Skeptics Guide
snovella@theness.com

Offline Joe Shmoe

  • Chief Rape Apologist
  • Stopped Going Outside
  • *******
  • Posts: 5347
Podcast #63 - Komsomolskaya Pravda - xray girl
« Reply #7 on: October 12, 2006, 01:34:13 PM »
Quote from: "rbairos"
I dont see getting 4/7 correct,  as about 50% accuracy.

She has 7 subjects, and 7 problems.  She has to match up the
7 problems with the 7 subjects.  

There are 7! (5040) ways of assigning (ordering) the 7 problems.
Getting 7/7 is only a 1 in 5040 chance (statiscally speaking).

I dont have the rest of the numbers with me (wrote up a little
brute-force permutation program) but getting 4 or more right was under 2%.

I wasn't talking about odds, I was talking about accuracy.

It is 50%, if you were an x-ray tech and you were given xrays for each of those 7 individuals you'd be able to identify with near-100% accuracy the ailment, injury or abnormality.  Someone with the inate ability to see through opaque tissue should have the same level of efficiency (6 of 7 at worst because the one subject with no abnomality might be missed).

Offline Freeman74

  • Keeps Priorities Straight
  • ***
  • Posts: 263
    • http://www.myspace.com\freeman74
Podcast #63 - Komsomolskaya Pravda - xray girl
« Reply #8 on: October 12, 2006, 01:37:36 PM »
Quote
(6 of 7 at worst because the one subject with no abnomality might be missed).


Not to nitpick, but I don't think 6 of 7 is possible. due to the setup she had 7 ailments and 7 subjects. A miss would mean improperly assigning 2 making the best case with an error 5/7.

Sorry, its minor, but I'm anal that way I guess.
quot;Phfft! Facts. You can use them to prove anything." - Homer Simpson

Offline Joe Shmoe

  • Chief Rape Apologist
  • Stopped Going Outside
  • *******
  • Posts: 5347
Podcast #63 - Komsomolskaya Pravda - xray girl
« Reply #9 on: October 12, 2006, 01:41:46 PM »
There were 6 conditons and one "normal" subject.  The normal subject can be mis-identified as having an abnormality not in the list of things that could be wrong.  I don't think they actually told her the target conditions, if they did then 4 of 7 is much less impressive.

Offline Freeman74

  • Keeps Priorities Straight
  • ***
  • Posts: 263
    • http://www.myspace.com\freeman74
Podcast #63 - Komsomolskaya Pravda - xray girl
« Reply #10 on: October 12, 2006, 01:43:11 PM »
She was aware of the conditions. Not knowing the conditions at all and getting 4 right by chance would have been astronomical.
quot;Phfft! Facts. You can use them to prove anything." - Homer Simpson

Offline Dirty J. Martini

  • Well Established
  • *****
  • Posts: 1818
  • I'm Pro-Accordian and I Vote!
Podcast #63 - Komsomolskaya Pravda - xray girl
« Reply #11 on: October 12, 2006, 03:13:58 PM »
I don't understand why she was able to see the subjects. Wouldn't it have been a valid test to put the patients behind, say, a paper screen. Something that would have no effect on XRays, but would still allow her to see the ailments (if she actually had the ability)? Heck, in that case, they could even have been naked to make it less likely that any outer clothing was "blocking" her abilities.

Offline rbairos

  • Off to a Start
  • *
  • Posts: 94
Podcast #63 - Komsomolskaya Pravda - xray girl
« Reply #12 on: October 12, 2006, 04:32:29 PM »
From the link Rebecca sent, the girl claimed her ability was not
hampered by clothing, but obstructed by a cloth hanging in front
of the participant.

I know this sounds silly, but I think both sides can agree that whatever
the girl was doing, it did not involve x-rays per se.
So disproving it was actual traditional x-ray technology, or that her accuracy was less than 100% wouldn't be very informative.  Believers could claim she has some type of empathic ability, etc.

For the record, here's the statistical probability of each score,
all external factors removed:

Score 0 ( 36.79 % )
Score 1 ( 36.80 % )
Score 2 ( 18.33 % )
Score 3 (  6.25 % )
Score 4 (  1.39 % )
Score 5 (  0.42 % )
Score 6 (  0 % )
Score 7 (  0.02 % )

Rob.

Offline Freeman74

  • Keeps Priorities Straight
  • ***
  • Posts: 263
    • http://www.myspace.com\freeman74
Podcast #63 - Komsomolskaya Pravda - xray girl
« Reply #13 on: October 12, 2006, 04:40:09 PM »
I believe that it was proposed to have some sort of screen. She rejected it and they accepted the rejection.

Its really pretty silly, Its just too bad that this case was treated this way. A poorly designed and executed test gives them amunition no matter how weak it really is.

On the surface to the uneducated observer this test could be taken as positive results.  This acutally damages our cause in the public eye, It shows a scientifically conducted experirment with seemingly higher than expected results.

I hope that there has been a lesson learned here and the test will be more stringent from now on.  We have to rely on facts, they can rely on an advertising campaign, and this didn't help our cause.
quot;Phfft! Facts. You can use them to prove anything." - Homer Simpson

Offline JD

  • Seasoned Contributor
  • ****
  • Posts: 563
Podcast #63 - Komsomolskaya Pravda - xray girl
« Reply #14 on: October 12, 2006, 06:48:31 PM »
This seems like a fairly silly test from the start.  If she's generating X-Rays from her eyes, what exactly do you need any patients(or patience) for?  Just have her look at a detector for a second, and then send her on her way.
roblem with the forum?  email me: leykial *at* yahoo *dot* com
Turn on, tune in, drop out, and shut the hell up, hippie.
"I'm a pretty good researcher ." - Neal Adams

 

personate-rain