Unfortunately, you can't assume that no one will be lying. See for example:
Merged New telepathy test: which number did I write ?
- Thread starter Michel H
- Start date
-
- Tags
- telepathy telepathy test
- Status
Crossbow
Seeking Honesty and Sanity
Then change your OP since it did not ask us to consider the implications of the people being tested providing dishonest answers.
Gord_in_Toronto
Penultimate Amazing
- Joined
- Jul 22, 2006
- Messages
- 26,492
Since the respondent does not actually know the correct number, they can't actually lie about it. They may lie about what they think it is but in such cases the quantum psychic energy field actually places the correct answer into their "lying" response.
Squeegee Beckenheim
Penultimate Amazing
- Joined
- Dec 29, 2010
- Messages
- 32,124
Whether they lie or not is unimportant. Whether they achieve results which are significantly above chance is what's important.
You're looking for a way to subjectively discount some answers while subjectively allowing others. This will invalidate all your results. Your results should be entirely objective.
You're looking for a way to subjectively discount some answers while subjectively allowing others. This will invalidate all your results. Your results should be entirely objective.
turingtest
Mistral, mistral wind...
I think anyone who's tempted to try to answer Michel's question seriously would be well advised to first read this post, which shows how he goes about "assessing credibilities." A taste:
Michel's test was (basically) "what number between 1 and 10 did I write down?" The answer was "10." Scarlettinlondon guessed "1, Right then done it," and Red Baron Farms said "Three" with no comment. Michel accepted SIL's answer as credible because she seemed "rather friendly," and her answer was "related to the target number" ("1" is part of "10"); RBF's answer, OTOH, was rejected as "not-credible" because Michel didn't like his screen name (too "aggressive"), and his answer wasn't anywhere near being right.
Michel, you couldn't shout "confirmation bias" any louder.
Michel's test was (basically) "what number between 1 and 10 did I write down?" The answer was "10." Scarlettinlondon guessed "1, Right then done it," and Red Baron Farms said "Three" with no comment. Michel accepted SIL's answer as credible because she seemed "rather friendly," and her answer was "related to the target number" ("1" is part of "10"); RBF's answer, OTOH, was rejected as "not-credible" because Michel didn't like his screen name (too "aggressive"), and his answer wasn't anywhere near being right.
Michel, you couldn't shout "confirmation bias" any louder.
Loss Leader
I would save the receptionist., Moderator
Since it is very, very, very easy to design an experiment where credibility is not an issue, it is the height of silliness to attempt to answer such a question.
Even trained psychologists require pages of testing before determining whether a particular person is likely to be credible in self-reporting.
Even trained psychologists require pages of testing before determining whether a particular person is likely to be credible in self-reporting.
I think that it all depends on if the recipient's number resembles the number "he" intended to send (the target): if it is either the target number itself, some multiple of that target number, or looks something like that target number (2 and 5 for example) In any of these cases I would certainly count it as a hit, so matter what else the correspondent said. If there is no way to relate the recipient's number to the target number, I would find some reason to question the recipient's credibility and ignore the number that they provided.
Remember that guessing at random any number from among 5 different ones would be 1:5, so any score you obtain using my method of more than 1:5 proves telepathy.
Does this help?
There are other ways to test sending a target number that have been explained to you, but you appear to have ignored these suggestions. I presume that they wouldn't work as well as the method I just described.
Last edited:
One more time: if you wish to avoid possible lying and clueless recipients, just chose a number between 1 and a million. Then the guesses from those lying and the clueless will just disappear in the noise of having 1:1,000,000 chances of guessing a number at random, whereas even 5 people getting your number correctly would probably be statistically significant (given the total number of responders, of course). If you wish to get even better statistics, just repeat the experiment a few times with the people who got the right number the first time. Easy.
Sure, it is possible to design an ESP test with no credibility evaluation (the test you did yourself on this forum is an example of such a test), but, if I did such a test with no credibility assessment, it would probably "fail" (i.e. give results which agree with pure chance), like your test did "fail". Clearly, whether respondents are serious and honest or not is of central importance:
turingtest
Mistral, mistral wind...
I was wrong- you can shout "confirmation bias" louder.
I believe that.
I and many others suggested a test in which whether respondents were serious or honest or not doesn't matter. Look up just above your post. Do you not want to do this kind of test, and if so, why not.
Loss Leader
I would save the receptionist., Moderator
Your, honor I object because it is devastating to my case.
-Overruled
Good call.
Liar, Liar
https://www.youtube.com/watch?v=Dx32b5igLwA
bruto
Penultimate Amazing
But if you are judging credibility and your subjects know it, then how can you detect when they are faking it?
It seems to me you are being far too complex in a simple task. If telepathy exists and if credible recipients are asked for an answer, they will give it. The result of the test should be better than chance. Discard all facetious answers, and keep those that follow the test procedure. Of those remaining, your mix should be predominantly people who believe in telepathy and have no reason to give false answers, and people who do not believe in telepathy and have no reason to give answers other than their non-telepathic guess.
Any answer that fits the requirements of the test must be considered a serious answer. For the test to be worth anything at all it must include answers from those who believe your theory is bunk, as long as they follow the rules. If you discard them, you test nothing.
abaddon
Penultimate Amazing
Perhaps you missed Michel's definition of "credibility". It works thusly. Post hoc, any answer which can be construed to agree with the "prediction" even by dint of numerological convolution will be admitted. Any answer which does not conform to this is deemed inadmissible and is discarded immediately. In the Michel H idea, all dice are loaded in one direction only. It remains inexplicable to me how Michel cannot see that the dice are being loaded by intentional selection of the results. In Michel's view, and he has stated this out front, if you state any result other than that which he seeks, then you are intentionally lying because you received the correct result and simply do not wish to admit it.
This is what he believes. He has stated so right here in this forum.
Giordano, you are not the first to make such a suggestion (and I wouldn't even be surprised if you had made it before on this forum, under its previous name). But actually real telepathy research is not conducted this way, even Randi did not recommend it (see here), J. B. Rhine for example used five "ESP cards", there are usually four possibilities in ganzfeld telepathy research. Using a (very) large number of possible answers, of "targets", has two drawback. First, my real (?) telepathic specificity, because it is real, is not "unlimited" (I am not attempting to do miracles here). So, if I had written "6743865454" on my modest sheet of paper for example, even if you were ("collectively") very honest, sincere and motivated, I think it might be really very difficult for you to psychically divine it. A second reason has to do with generosity and motivation. It seems that, for many people, even if they know a lot about my modest person through ESP because of a very rare and exceptional worldwide telepathy phenomenon, it is, so to speak, "internally very hard" to admit it. In addition, I am probably perceived as weak and isolated, with little support, and this, in our society can unfortunately sometimes have serious consequences, people might say to themselves "Why should I do it?". If the target pool is small, the psychological effort people have to provide is smaller, the risks of "giving away too much information" are perceived as more limited. If the target pool is large, people will have an impression of giving away a lot of info, and many will be very reluctant to do that.
By the way, I have not yet received an answer to my initial question: which answer is more credible? I also (ask and) invite you to give "credibility ratings" (between -10 and 10) to the two hypothetical answers given in the OP. If you give (in my opinion) good credibility ratings to these two answers (positive or negative), I will try to acknowledge it.
Yes, I indicated that other people also suggested this, so your unwillingness to accept this seems even more telling.
Try this (as I mentioned before) if you think people have trouble choosing from big numbers, have them first think only of and choose only the first digit of six, then only of the second, then only of the third, etc. Pretty soon you have a million or more possible 6 digits, which would have the statistical advantages I mentioned.
This would be more or less equivalent to doing several tests, each with ten different targets, and my latest test was precisely such a test, with ten different targets. But, I think that rather than repeating in an endless way the same kind of test, the issue of credibility should be clarified "in a neutral way". This is why I would like to urge you to answer the question at the end the OP (although the thread has now be merged). I am not Sylvia Browne.
Then how did you do in your latest test? My understanding was that you did very poorly, and that no one guessed the target number. But then, I don't remember the repetition of the tests. How did all the numbers work out?
Why would I give you an answer to your question if you already did the test the our suggested way where reliability or perceived reliability would not play a role?
- Status