Sheldrake tests telephone telepathy

[Prester John]

yes they are, should have paid more attention.

Heres my take on the stats:

Experiments Calls Right % right p
Preliminary 30 13 43 0.02
Series 1 201 75 37 0.00007
Series 2 322 134 42 5 x 10-11
S. Bloomfield 18 9 50 0.02
Totals 571 231 40 4 x 10-16

these are the results shown in the paper. They look pretty good and are statistically significant (very) using the tests they have used. The tests used seem reasonable enough.

My biggest problem with the results is they seem to be skewed, nearly everyone scored above average. When you draw a histogram of % correct guesses it looks like a normal distribution, but with one side (the non significant results) chopped away. we are left with just the significant results. This alone is strongly suggestive of incomplete data.

you would expect the results to be reasonably normally distributed even if they showed greater than expected correct caller guessing.

edit- the use of the word average is wrong, expected would be a better word. (hey i was rushing to go home)

 

[Paul C. Anagnostopoulos]

I'm guessing this idea will die from lack of volunteers, but anyway.

I'm leaning toward an email experiment, just because it's less hassle. I've asked Pam Smart for their protocol. The problem is that we all know each other, so there is nothing to prevent the selected sender from contacting the recipient outside the protocol. We could set things up so that the senders do not know who the recipient is, using anonymous email addresses, but the recipient will know the senders and can contact them to set up a cheat.

And will the telepathy work if the sender is addressing the recipient through an anonymous email address and so doesn't know who the recipient is? How would those telepathic vibes get from the sender to the recipient?

~~ Paul

 

[Paul C. Anagnostopoulos]

John said:

My biggest problem with the results is they seem to be skewed, nearly everyone scored above average. When you draw a histogram of % correct guesses it looks like a normal distribution, but with one side (the non significant results) chopped away. we are left with just the significant results. This alone is strongly suggestive of incomplete data.

Sorry, I don't understand what you're saying. Do you think there ought to be some series with lower % right numbers?

~~ Paul

 

[JamesM]

I'm guessing this idea will die from lack of volunteers, but anyway.

If nothing comes of it, this may be interesting as a purely academic exercise.

The problem is that we all know each other, so there is nothing to prevent the selected sender from contacting the recipient outside the protocol.

That's why I would suggest that the experimenter contact the sender under a very strict time limit, to reduce the chance that they will have time to both send the sneaky cheating message (by whatever means) and send the real message.

We could set things up so that the senders do not know who the recipient is, using anonymous email addresses, but the recipient will know the senders and can contact them to set up a cheat.

What if everyone has to use the same email address and the content of the email is determined solely by the experimenter and the mailers have no way of knowing what it is until they're told to send it. If the email is cc'ed to a third party, it can be compared to the instructions sent by the experimenter to see if the emailer changed it in any way as a way of signalling to the recipient.

And will the telepathy work if the sender is addressing the recipient through an anonymous email address and so doesn't know who the recipient is? How would those telepathic vibes get from the sender to the recipient?

It seems to me that the mailer should know the identity of the recipient to make it similar to the telepathy by phone experiment.

 

[Moz]

A question...

... did his experiment record only incorrect precitions of the caller's identity, or did it also include the event that a "telepathically" anticipated call *did not* take place?

Surely this is a factor. Focussing on the identity of the caller might have lead the experimenters away from including the *fact* (or otherwise) of a call as an event worth recording in the "experiment".

 

[Prester John]

yes basically nearly every one did better than expected, this is not what you would expect

if you propose that the sample would do better than expected, say they are psychic and can identify 40% rather than the expected 25%. You would still get some, who by chance do less well than expected and only get 10 or 20 % correct. So you should get a distribution with a mean of 40% and something close to a normal distribution.

This data is not like that. Large numbers of people got 30-50% correct with the numbers getting 50+ % tailing off in the "normal" way. but at 0-29% there are practically no people. This is indicative of incomplete data - those results should be there.

There are a number of ways to proceed:

To compensate statistically for it is possible, simply fill in the missing data by mirroring the data present.This would allow a statistical analysis of what is more likely the full data set.
However as the data is not complete it is acceptable to discard the entire study as there are probably other flaws if there is missing data. Of course the authors should be given their opportunity to defend the apparent missing data

 

[JamesM]

Re: A question...

... did his experiment record only incorrect precitions of the caller's identity, or did it also include the event that a "telepathically" anticipated call *did not* take place?

Do you mean, were there "blank" runs, where no phone call came, but the person made a guess, anyway? If so, no, that wasn't part of the protocol: no guess was made until the phone rang. So, no ringing, no guess. Perhaps this sort of thing could be included if you could get some sort of automated call up service as a possible caller.

 

[CFLarsen]

I'm guessing this idea will die from lack of volunteers, but anyway.

Give it time. It's only been a few hours.

 

[Paul C. Anagnostopoulos]

Well, let's take a look at the videotaped trials.

Hawksley: 3 callers significant, 3 not. Overall p=.000001
Hawksley: 2 callers significant, 3 not. Overall p=.0008
Hawksley: 2 callers significant, 2 not. Overall p=.04

Reeves: 1 caller significant, 3 not. Overall p=.05

Morsman: 1 caller significant, 3 not. Overall p=.20

Marcovici: 2 callers significant, 2 not. Overall p=.0003

These experiments show a wider range of results.

~~ Paul

 

[Paul C. Anagnostopoulos]

Here is Pam Smart's response to my request for the email protocol:

It is important to have people who are connected emotionally. They send their guess one minute before the trial time and the person who gets picked sends their email with a Cc copy to the co-ordinator at the exact trial time.

Are email time stamps that accurate?

~~ Paul

 

[Interesting Ian]

Here is Pam Smart's response to my request for the email protocol:

quote:
--------------------------------------------------------------------------------
It is important to have people who are connected emotionally. They send their guess one minute before the trial time and the person who gets picked sends their email with a Cc copy to the co-ordinator at the exact trial time.
--------------------------------------------------------------------------------


Are email time stamps that accurate?
~~ Paul [/B]

I absolutely 100% agree about them having to be connected emotionally. But for some reason it's just something that skeptics can't get their heads around

 

[CFLarsen]

Here is Pam Smart's response to my request for the email protocol:
Are email time stamps that accurate?

~~ Paul

Far from it. It depends entirely on the clock settings on the mail servers.

If they hadn't checked this beforehand, they are worse experimenters than both Steve Grenard and Gary Schwartz!!

 

[CFLarsen]

I absolutely 100% agree about them having to be connected emotionally. But for some reason it's just something that skeptics can't get their heads around

Please point to the experiment that proves that people have to be "connected emotionally".

How is "emotion" measured and defined?

 

[JamesM]

Far from it. It depends entirely on the clock settings on the mail servers.

If they hadn't checked this beforehand, they are worse experimenters than both Steve Grenard and Gary Schwartz!!

So if I send an email from my Yahoo account to my Hotmail account, what determines the time that appears on the email I receive?

 

[Darat]

I absolutely 100% agree about them having to be connected emotionally. But for some reason it's just something that skeptics can't get their heads around

What level of "emotional connectedness" is needed and how would it be established?

 

[Paul C. Anagnostopoulos]

Ian said:

I absolutely 100% agree about them having to be connected emotionally. But for some reason it's just something that skeptics can't get their heads around.

Does this go for contacting dead people, too? Does it go for influencing random number generators (I feel a real affinity for RNGs, don't you know)?

~~ Paul

 

[Interesting Ian]

Please point to the experiment that proves that people have to be "connected emotionally".

I can't point to any experimental data. It's kind of from personnal experience, from anecdotes, and what I instinctively feel.

How is "emotion" measured and defined?

Oh it can't be That's because it's real. Science deals with measurements. That's because it deals in a world of unreality

 

[CFLarsen]

So if I send an email from my Yahoo account to my Hotmail account, what determines the time that appears on the email I receive?

That's a very, very good question! And it needs to be answered, if this test is going anywhere (which I hope it is - if we can't pull it off, who else can? )

I can't point to any experimental data. It's kind of from personnal experience, from anecdotes, and what I instinctively feel.

Funny, but I never saw you as a "touchy-feely" guy. Guess I was wrong...

You can't back up your claim, then. OK. Your claim is therefore invalid.

Oh it can't be That's because it's real. Science deals with measurements. That's because it deals in a world of unreality

I hope you are joking. You certainly have been unable to back up your posts!

 

[patnray]

quote:
--------------------------------------------------------------------------------
There was a striking difference between Sue’s performance with familiar and unfamiliar callers. With the two familiar callers, she was right 25 times out of 35 (71%; p = .00000001). With the unfamiliar callers, she was right only 5 times out of 35 (14%), not significantly different from the chance level (see Figure 1). The difference between success rates with familiar and unfamiliar callers was very significant statistically ( p = .000001).
--------------------------------------------------------------------------------

What percentage of Sue's guesses were the two callers she was emotionally connected to? If she guessed them 80% of the time it would skew the results for both classes.

Also, when selecting a subclass of the data, such as the 2 callers she was familiar with, what is the correct way to calculate the expected chance results?

If I always guessed A, I would be right 100 % of the time when A calls and 50% of the time in any subgroup of 2 that includes A.

It would seem that any analysis of the data should include examination of any patterns in the receiver's guesses.

 

[Paul C. Anagnostopoulos]

Claus, an email message is time-stamped by the SMTP server receiving the message from your mail app.

Patnray, both of the cited papers have tables that break down guesses by caller and guess. For example, in one series, Hawksley never even guessed one of the callers; she divided her guesses between the other three. She had insignificant results for two callers; significant results for the other two; signficant results overall.

I'm telling you, folks, without a statistician we're going nowhere. In the case of email telepathy, for example, I cannot imagine what the sensory leak is. Seems to me the results are due either to something amazing, cheating, or bad statistics.

~~ Paul