Merged Telepathy test: which number did I write?

Status
Not open for further replies.
I'm quite happy to participate in a reasonably credible experiment, blinded along the lines Hokulele suggested, although I'd like to know what the point of analysing statements might be: how could suitable criteria be defined and applied? - why not simply require an unadorned number or letter or word guess, without any accompanying statement?

To take it seriously, I'd expect to see the full protocol and details of the analysis to be done before the experiment starts.
 
Last edited:
Michel H said:
Your proposal is interesting, Hokulele, this sounds like an interesting protocol. (Thank you also to Femke; it is possible that I request your assistance at some point, Femke.)
However, I find it a little complicated and I am afraid that answerers (to the telepathy test) who would have to send a private message to the assistant, instead of posting directly on the forum, would be less spontaneous and relaxed, and that their answers would be too "calculated", and that a test with your procedure would end in failure because of that. It is possible that some people don't mind participating once or twice in a telepathy test which looks a little bit like a game
Click to expand...


I disagree. When posting their answers publicly, people are apt to be less spontaneous and natural, as they know their answers will be scrutinized and possibly criticized. This is one reason that scientific psychological studies are done in the strictest anonymity and privacy. In addition, as was mentioned earlier in this thread, people may be influenced by others' reactions, rather than having "pure" reactions of their own.

<snipped the irrelevant hyperbole>

but it is also possible that, once they feel the test's goal would be to rigorously prove that I am "telepathic" (if I really am so), that would "scare" them and their worst instincts would have the upper hand.
Keep also in mind that I am still fairly new here (I have more experience on Yahoo! Answers), it is still not clear what kinds of results I might get here in the long run, with more number-guessing (for example) threads.
Click to expand...


If you are unhappy with the people who respond on this forum, we can combine the more stringent requirements with your preferred venue. Details at the end of this post.

If a more rigorous procedure is needed (more rigorous than what I am currently doing or trying to do), I could ask people on this forum to evaluate credibilities of answers (not credibilities of their own answers, credibilities of other answers), without giving them any information (by sensory means) regarding the numerical correctness (or not) of answers beforehand. If this could produce the right number, telepathy would be proved. I could also ask people to evaluate credibilities themselves after I have provided them with all the information I have (which numbers are correct, and my own "credibility analysis"), this would be much less rigorous, but could nevertheless be interesting (in my opinion).
Click to expand...


Again, I disagree. All this does is add a layer of subjective opinion to a flawed data set. Here is my second proposal:

Have Femke set up a brand new e-mail account, to be used solely for this experiment. She may use any of the free services such as Hotmail or Google, or another service of her choice if this is unacceptable to you or her. You nor any other person will have any access to this account. Only Femke will be able to read the e-mails sent to this address.

On Google Answers, post your request for people to guess your number, and indicate they must send their responses to Femke's e-mail address. Any reponse posted to Google Answers will be discarded. Only answers sent to the address provided will be accepted. Put a deadline for responses so this doesn't drag on too long. Restrict people to only one response.

Femke will then replace the guessed number with the letter X, and post all answers here. You will then publicly perform your analysis and post your results. Femke will then post the numbers guessed by the participants.
 
Hokulele said:
Here is my second proposal:

Have Femke set up a brand new e-mail account, to be used solely for this experiment. She may use any of the free services such as Hotmail or Google, or another service of her choice if this is unacceptable to you or her. You nor any other person will have any access to this account. Only Femke will be able to read the e-mails sent to this address.

On Google Answers, post your request for people to guess your number, and indicate they must send their responses to Femke's e-mail address. Any reponse posted to Google Answers will be discarded. Only answers sent to the address provided will be accepted. Put a deadline for responses so this doesn't drag on too long. Restrict people to only one response.

Femke will then replace the guessed number with the letter X, and post all answers here. You will then publicly perform your analysis and post your results. Femke will then post the numbers guessed by the participants.
Click to expand...
Hokulele, your second proposal seems rather similar to the fist one to me. I think that the method you propose is probably more rigorous than mine, it is possible that I shall use it later. However, keep in mind that, in ESP research, rigor is not the only thing you should worry about. People might be inhibited by a too complicated and serious procedure.

I call the method I'm using here semi-rigorous only, because I may have been a little influenced by my knowledge of the numerical correctness (or not) of the 13 valid answers when I chose the various credibility ratings. The problem of a possible correlation between the various answers during the test is only a minor difficulty because this cannot raise the correct answer rate. I insist, however, that I tried to evaluate all 13 answers credibility-wise (which is different from "quality-wise", those who answered incorrectly may also have made a useful contribution) in a fair way as I usually do. I don't think it would have been possible to obtain a 100% correct answer rate among credible answers if the data had not be of (fairly) good quality, thanks to the 13 answerers. When I look at all answers in post post #177, it seems fairly obvious to me that the (numerically) correct answers are more credible and friendlier than the incorrect ones. I suppose that the people who do not see this do not see it either because they didn't bother to read my analysis carefully, or because they don't want to see it (bad faith). There are of course also some people who just don't have the sense of scientific research (they may, however, have other qualities).
 
Michel H said:
Hokulele, your second proposal seems rather similar to the fist one to me. I think that the method you propose is probably more rigorous than mine, it is possible that I shall use it later. However, keep in mind that, in ESP research, rigor is not the only thing you should worry about.
Click to expand...

Think about the implications of that for a minute.

Michel H said:
When I look at all answers in post post #177, it seems fairly obvious to me that the (numerically) correct answers are more credible and friendlier than the incorrect ones. I suppose that the people who do not see this do not see it either because they didn't bother to read my analysis carefully, or because they don't want to see it (bad faith).
Click to expand...

How we evaluate their credibility isn't really relevant. The relevant point is that having the person who already knows the "correct" number evaluate the responses makes your study worthless. Even if there is no impropriety on your part there is the opportunity for impropriety, making your results unreliable.
 
Michel H said:
Hokulele, your second proposal seems rather similar to the fist one to me. I think that the method you propose is probably more rigorous than mine, it is possible that I shall use it later. However, keep in mind that, in ESP research, rigor is not the only thing you should worry about. People might be inhibited by a too complicated and serious procedure.
Click to expand...


Responding via e-mail is no more complicated that responding via an Internet forum. And I have to ask, why is rigor less important in ESP research than in proper scientific research?

I call the method I'm using here semi-rigorous only, because I may have been a little influenced by my knowledge of the numerical correctness (or not) of the 13 valid answers when I chose the various credibility ratings.
Click to expand...


I would say it is likely to be more than just a little. You could easily prove me wrong by taking a more rigorous test. Are you willing?

The problem of a possible correlation between the various answers during the test is only a minor difficulty because this cannot raise the correct answer rate.
Click to expand...


Sure it can. Even if it is by accident, people are more likely to change their answers based on others. There tends to be two ways this can happen, either everyone chooses the most popular answer, or, in a test like this, people deliberately try to choose an answer different from everyone else. This results in the almost equal distribution we saw here. The whole, "I can't pick that because it was already taken" phenomenon. You can easily avoid many different types of bias by having the responses be private.

I insist, however, that I tried to evaluate all 13 answers credibility-wise (which is different from "quality-wise", those who answered incorrectly may also have made a useful contribution) in a fair way as I usually do.
Click to expand...


Again, I disagree. Calling improper punctuation (Im) worse than improper capitalization (wrong) is purely an opinion, one that is likely to be biased based on the outcome you desired. If the answer with "Im" had been correct, would you have been more likely to look kindly on the mistake?

I don't think it would have been possible to obtain a 100% correct answer rate among credible answers if the data had not be of (fairly) good quality, thanks to the 13 answerers.
Click to expand...


And here again, I disagree. I believe the data was of very poor quality, given that people could, and respond to, each other's answers. Try this little test sometime: Give a group of people a public poll where they have to choose a number from 1-4. Most likely, there will be an even distribution amongst all four numbers, much as we saw here. Give a similar group of people another poll, but one where they cannot see how anyone else answered until they have made their choice. Typically, people will choose either 2 or 3, and ignore 1 and 4. Don't just trust me, actually try this experiment for yourself.

This is fairly typical testing for a psychological claim.

When I look at all answers in post post #177, it seems fairly obvious to me that the (numerically) correct answers are more credible and friendlier than the incorrect ones. I suppose that the people who do not see this do not see it either because they didn't bother to read my analysis carefully, or because they don't want to see it (bad faith). There are of course also some people who just don't have the sense of scientific research (they may, however, have other qualities).
Click to expand...


Then it should be even more obvious when the answers are masked prior to the evaluation, no? We would know for sure your analysis was based on the inherent credibility of the response. Why would you think such a clear difference would be masked if you didn't know the answers? Isn't telepathy based on knowing without seeing?
 
Spindrift said:
My second guess is that Michel won't return.
Click to expand...

Second guessing? Sorry, your answer is not serious. This is disqualified. Oh, you were wrong? Well, then it's qualified because uh, let's see... uh... you had no spelling or grammatical errors. So your answer is accepted and is clear proof that skepticism is a failure.
 
devnull said:
I guess at the end of the day Michel is free to run his experiments any way he wishes, as long as he doesnt expect anyone to take him seriously.
Click to expand...


With the protocols he's proposing, no one will.
 
New telepathy test: which number did I write?

Hi, I would like to invite you to participate in a (new) simple telepathy test.

It is not the first time I propose a telepathy test on this forum, the previous one is here: http://www.internationalskeptics.com/forums/showthread.php?t=241593 .

At about 22:47 on this Thursday August 22 (Brussels, Belgium time), I wrote carefully one of the four numbers: "1", "2", "3", "4" on my sheet of paper, and I surrounded it with a circle. Then, I wrote it again twice.

I shall repeat this number from time to time during this test.

It was selected using this random number generator: http://www.random.org/integers/ .

I ask you to write it here (if you think you know it, even with a doubt). You may also answer "I don't know".

So, your answer should be one of the four numbers: "1", "2", "3", "4", or "I don't know".

A comment might be useful, but is not indispensable.

Please note that the number I wrote has no meaning, it was just produced by the generator.

A MD5 hash code for a complicated sentence containing the number I ask you to telepathically guess (like, for example: "the number to guess is 5 f4315d 3b1fcd81") is:
e5ca98da86a6e4c582700847e587c3ac

It was obtained on this website:
http://www.md5hashgenerator.com/ .

I shall reveal the actual sentence I used to produce this MD5 hash at the end of the test, after I have revealed the number I ask you to guess. This way, you'll be able to verify my number.

Thank you for participating.

Note: I do these tests because I believe I have a particular tendency to (telepathically) communicate my thoughts to others, and I am seeking to prove this through online telepathy experiments.
 
Status
Not open for further replies.

Back
Top Bottom