New telepathy test, the sequel.

[steenkh]

I've been pondering why so many posters (including me) try once again to engage with Michel whenever he bumps this thread despite ample evidence that it's pointless, and I've concluded it's because - unlike the majority of the believers who come here with their worthless anecdotes - he clearly has the intellectual capacity to understand why what he is doing will not produce results any more reliable than such anecdotes.

I do not think that Michel H can be convinced of anything as long as he has voices in his head telling him otherwise. His intellectual capacity is irrelevant in this case - except that he is using it to pervert any test to give the only result that he regards as credible.

The reason why we have to continue to debate his posts is that if we don't he will regard it as a victory, and he will link others to it, and point out how the skeptics fell silent because they could not face the fact that ESP is real.

 

[Nay_Sayer]

That is obvious: 10% will be regarded as a significant proof, and four choices will by themselves give 25% success rate. In other words, a solid proof.

Note to Michel H: this post is pure sarcasm.

Not remotely clear enough;

████░██░██░██░▄███▄
░██░░██▄██░██░▀█▄▀▀
░██░░██▀██░██░▄▄▀█▄
░██░░██░██░██░▀███▀

██░▄███▄
██░▀█▄▀▀
██░▄▄▀█▄
██░▀███▀

▄███▄░░▄███▄░░████▄░░▄███▄░░▄███▄░░▄███▄░██▄░▄██
▀█▄▀▀░██▀░▀██░██░██░██▀░▀▀░██▀░▀██░▀█▄▀▀░██▀█▀██
▄▄▀█▄░███████░████▀░██▄░▄▄░███████░▄▄▀█▄░██░░░██
▀███▀░██░░░██░██░██░░▀███▀░██░░░██░▀███▀░██░░░██




Try that

 

[Jack by the hedge]

The thing is, Michel is already convinced he's a telepath. It seems what he would like from us is firstly participation but also affirmation.

In that, he'll forever be frustrated since, however sympathetic we might be for the symptoms he has to cope with, people here are not going to pretend the stuff he's done is valid evidence for the existence of telepathy.

 

[Michel H]

I clicked on the link and have a question. Why only four choices?

I have tried various possibilities before, for example two choices:
https://answers.yahoo.com/question/index?qid=20101213145911AAaUw4r
or seven choices:
https://answers.yahoo.com/question/index?qid=20170330141634AAai019 .

I found that "four choices" is the best compromise: not too easy because people could be somewhat irritated by a too easy test, and not too hard, because some people might not like the effort, and also revealing too much about their abilities. I think four choices usually puts them in the best comfort zone.

Using four possibilities is also what is done in ganzfeld experiments:
In the judging procedure, the receiver is taken out of the Ganzfeld state and given a set of possible targets, from which they select one which most resembled the images they witnessed. Most commonly there are three decoys along with the target, giving an expected rate of 25%, by chance, over several dozens of trials.[9]
(https://en.wikipedia.org/wiki/Ganzfeld_experiment ).

 

[Michel H]

The thing is, Michel is already convinced he's a telepath. It seems what he would like from us is firstly participation but also affirmation.

In that, he'll forever be frustrated since, however sympathetic we might be for the symptoms he has to cope with, people here are not going to pretend the stuff he's done is valid evidence for the existence of telepathy.

"Firstly participation"?

May I remind you that you have already given at least one decent answer in one of my tests? :

... So I change my guess to 3. ...

After a careful examination, your answer was found to be incorrect, and also non-credible.

 

[JayUtah]

I found that "four choices" is the best compromise: not too easy because people could be somewhat irritated by a too easy test...

Subject irritation could be eliminated as a factor by using a better source of subjects. The one you've chosen seems to present you no end of problems, which you then have to solve with subjective post hoc methods that make your experiment irreproducible.

What are the statistical consequences of having so few alternatives?

...and not too hard, because some people might not like the effort

Again, stop using convenience samples and this problem goes away. I've explained the statistical advantage of increasing the number of alternatives. Please comment on that.

...and also revealing too much about their abilities.

This suggests you believe there are profoundly psychic respondents whose manifested abilities will be somehow attenuated by your methodology, for Cthulhu-only-knows what reasons. Based on your prior answers, I guess you mean that Yahoo management and/or the CIA might take note of the people who score highly on your test and that this puts them in danger of assassination. Two immediate problems, if that's the case: The CIA has no way of knowing who scored highly on your test unless you choose to reveal that information, and you admit you have no evidence that profoundly psychic people are in substantial danger. You waved your hands wildly at the possibility, but since you have no evidence to back that up, you're unnecessarily filtering your subjects based on your preconceptions of how they will answer. Finally, please provide an error analysis showing the quantitative boundaries the programmed attenuation will have on the final results.

I think four choices usually puts them in the best comfort zone.

Does the statistical basis of your anticipated data set enter into your thinking at all?

Using four possibilities is also what is done in ganzfeld experiments:
In the judging procedure, the receiver is taken out of the Ganzfeld state and given a set of possible targets, from which they select one which most resembled the images they witnessed. Most commonly there are three decoys along with the target, giving an expected rate of 25%, by chance, over several dozens of trials.[9]
(https://en.wikipedia.org/wiki/Ganzfeld_experiment ).

The key being, "...over several dozen trials" -- in this case many trials using the same subject. That's not the same as one trial for each of some unknown number of subjects. As Loss Leader correctly pointed out, your study omits the most important protocol one can do in such research. Can you tell us why?

 

[JayUtah]

After a careful examination, your answer was found to be incorrect, and also non-credible.

Why aren't you addressing the posts that illustrate that your real criteria has nothing to do with any measure of credibility, and is entirely correlated to post hoc evaluation of correctness?

 

[RoboTimbo]

I found that "four choices" is the best compromise: not too easy because people could be somewhat irritated by a too easy test, and not too hard, because some people might not like the effort, and also revealing too much about their abilities. I think four choices usually puts them in the best comfort zone.

I find this post to be non-credible. Guessing one out of 10 choices takes no more effort than guessing one out of four choices.

Sorry, your post just seems dishonest and designed to support a delusion rather than seeking a fair test.

 

[Nay_Sayer]

Why aren't you addressing the posts that illustrate that your real criteria has nothing to do with any measure of credibility, and is entirely correlated to post hoc evaluation of correctness?

That correlation being, for those interested; The more questions you ask and more importantly the more you press Michel for an honest answer the less credible you are.

 

[Gamolon]

I found that "four choices" is the best compromise: not too easy because people could be somewhat irritated by a too easy test, and not too hard, because some people might not like the effort,

Not like the effort? Effort doing what exactly. If you did 30 words, all they would need to do is find the word that you telepathically sent out.

and also revealing too much about their abilities. I think four choices usually puts them in the best comfort zone.

How would thirty choices and someone getting your chosen word correct out of those thirty reveal "too much about their abilities" compared to four choices and someone getting it correct out of four?

What does an increased number of choices have to do with the amount of information revealed about their abilities?

 

[Michel H]

Not like the effort? Effort doing what exactly. If you did 30 words, all they would need to do is find the word that you telepathically sent out.



How would thirty choices and someone getting your chosen word correct out of those thirty reveal "too much about their abilities" compared to four choices and someone getting it correct out of four?

What does an increased number of choices have to do with the amount of information revealed about their abilities?

Well, let's assume I have written and circled 29 on my paper, and I only reveal to you in a normal and sensory way (for example, on this forum) that I have written one of the thirty numbers: 1, 2, 3, 4, ..., 28, 29, 30.

Suppose you tell me (correctly): "You wrote and circled '29' ". This would be a rather extraordinary feat, which "proves" telepathy in a much more convincing way than if you just guessed right in a four-possibility test, with a 25% chance of getting it right from pure chance alone.

 

[Startz]

Well, let's assume I have written and circled 29 on my paper, and I only reveal to you in a normal and sensory way (for example, on this forum) that I have written one of the thirty numbers: 1, 2, 3, 4, ..., 28, 29, 30.

Suppose you tell me (correctly): "You wrote and circled '29' ". This would be a rather extraordinary feat, which "proves" telepathy in a much more convincing way than if you just guessed right in a four-possibility test, with a 25% chance of getting it right from pure chance alone.

So if 30 choices is "much more convincing," why do you only want to use 4???

 

[Michel H]

So if 30 choices is "much more convincing," why do you only want to use 4???

Because, as I explained, there are other considerations which play a role: ease of the test (if people have actually only limited information through extra-sensory channels), and a possible desire to not reveal too much, perhaps for mental health protection for everybody.

 

[Loss Leader]

Suppose you tell me (correctly): "You wrote and circled '29' ". This would be a rather extraordinary feat, which "proves" telepathy in a much more convincing way than if you just guessed right in a four-possibility test, with a 25% chance of getting it right from pure chance alone.

Um ... yes. That's why 30 choices is a better design. You just made your own point. The less likely a correct answer can be guessed by chance, the more likely that something other than guesswork is at play.

By your own logic, your tests should have 30 possible answers instead of 4.

 

[JayUtah]

Because, as I explained, there are other considerations which play a role...

And you were told how to mitigate those considerations without doing damage to the confidence of your results. You haven't talked at all about what these compromises do, statistically speaking, to your conclusions. You insist that your protocol is better than any that has been devised here -- most of which eliminate the problems you're trying to work around. But you can't talk meaningfully about how you compensate for the consequences of your workarounds. Your "other considerations" are wagging the dog.

...and a possible desire to not reveal too much, perhaps for mental health protection for everybody.

No, there are no super-psychics out there afraid to take your test. You need to stop letting your fantasies attempt to dictate your science.

 

[RoboTimbo]

Well, let's assume I have written and circled 29 on my paper, and I only reveal to you in a normal and sensory way (for example, on this forum) that I have written one of the thirty numbers: 1, 2, 3, 4, ..., 28, 29, 30.

Suppose you tell me (correctly): "You wrote and circled '29' ". This would be a rather extraordinary feat, which "proves" telepathy in a much more convincing way than if you just guessed right in a four-possibility test, with a 25% chance of getting it right from pure chance alone.

LOL. No wonder you don't want to do a test with 30 choices. Do you even read what you type?

 

[JayUtah]

The less likely a correct answer can be guessed by chance, the more likely that something other than guesswork is at play.

Right, that's the whole concept of the p-value. To refer again to what he said about you, he claims your one correct answer proves you have psychic ability. But let's scrutinize the nuts and bolts of that. There were four alternatives to choose from, one of them having been predesignated as the "success" outcome. If one choice is allowed, then the null hypothesis is that you have a 25% chance of choosing randomly and getting a hit. The hypothesis under test is that you are telepathic, and thus able to know which is the right answer and choose it deliberately. What the p-value asks is what is the probability that a particular set of data arose by the null hypothesis that just happened also to confirm the hypothesis. Typically science wants a probability less that 0.05 that the null hypothesis accidentally produced confirmatory results for the operative hypothesis. In your case the p-value is trivial to obtain: the probability that one trial selecting at random from four alternatives obtained a hit is 0.25, far higher than what we generally consider statistically significant.

 

[Michel H]

Right, that's the whole concept of the p-value. To refer again to what he said about you, he claims your one correct answer proves you have psychic ability. But let's scrutinize the nuts and bolts of that. There were four alternatives to choose from, one of them having been predesignated as the "success" outcome. If one choice is allowed, then the null hypothesis is that you have a 25% chance of choosing randomly and getting a hit. The hypothesis under test is that you are telepathic, and thus able to know which is the right answer and choose it deliberately. What the p-value asks is what is the probability that a particular set of data arose by the null hypothesis that just happened also to confirm the hypothesis. Typically science wants a probability less that 0.05 that the null hypothesis accidentally produced confirmatory results for the operative hypothesis. In your case the p-value is trivial to obtain: the probability that one trial selecting at random from four alternatives obtained a hit is 0.25, far higher than what we generally consider statistically significant.

I see nothing wrong in your statistical analysis, JayUtah. But don't forget Loss Leader did not just write a number, he said:

I am seeing a 4 very clearly. It's almost as though I had written it myself.

and confirmed later:

... Early on, I used my telepathic powers to see into your ... mind ... your thoughts were very easy to read ...

and, by doing so, may have revealed a great deal about his mental processes (in spite of what he said later).

Keeping a purely statistical approach, what is the probability that a moderator on the prestigious James Randi Educational Foundation would say:

I am seeing a 4 very clearly. It's almost as though I had written it myself.

Probably very small, and this is one argument for rejecting the null hypothesis.

 

[Startz]

Loss Leader wrote

Please, please, please, please, please stop quoting my six and a half year-old post as evidence of anything whatsoever.

I did not receive your thoughts. I did not see any number. I was in no way influenced by you.

I made my post as a joke. I wrote it sarcastically. My intention was to be cruel. My intention was to mock you and make your pretend test look as absolutely foolish as possible.

When you ignore him, you greatly reduce your credibility.

 

[JayUtah]

I see nothing wrong in your statistical analysis, JayUtah.

Then you should concede that the statistical analysis doesn't support your conclusion.

But don't forget Loss Leader did not just write a number, he said...

All statements which he has copiously disavowed and which the rest of us handily recognize as having been facetious from the start. In any case, your subjective impression of what Loss Leader was saying does absolutely nothing to fix your statistics.

...may have revealed a great deal about his mental processes (in spite of what he said later).

No. You're simply trying to substitute your home-grown psychoanalysis for the clear statements of the actual person in question. That's arrogant, rude, and entirely unscientific. If I don't get to speculate about your mental processes, you don't get to do it either to someone else. You have demonstrated many times your inability to recognize sarcasm. Therefore your personal judgment regarding what is or is not sarcastic has no evidentiary value -- statistical or otherwise. In any case, accepting the subject's statement when it appears to support your belief and rejecting it when it does not is completely dispositive of your "method." You simply ignore anything you don't want to hear.

Keeping a purely statistical approach, what is the probability that a moderator on the prestigious James Randi Educational Foundation would say:

Probably very small...

How is your speculation "keeping a purely statistical approach?" You simply mentioned the concept of statistics and then offered your speculation. Tell us again what a great scientist you are.