New telepathy test, the sequel.

[JayUtah]

You already claimed you could drop the 1 from 12 to get 2...

Or more directly, also drop the 2 and get 1, which was also a valid choice. Treating the numeral as a string of digits and considering each one digitwise, one has a choice between the hit and a miss. Choosing the hit with knowledge that it's a hit is pure loading of the dice.

 

[Gord_in_Toronto]

I'm at least four years late to this thread but, by any chance, is the answer "1"? "Or maybe "2"? Or possibly "3"? Or potentially "4"? Or feasibly "5"? Or very likely "6"? Or hopefully "7"? Or God willing "8"? Or just mayhap "9"? Or peradventure "10"? Or perchance "11"?

 

[abaddon]

Or more directly, also drop the 2 and get 1, which was also a valid choice.

Nope, 1 was never a valid choice. 2,3,4,5 were valid, 1 was never in the basket. You owe me a t-shirt.

Treating the numeral as a string of digits and considering each one digitwise, one has a choice between the hit and a miss. Choosing the hit with knowledge that it's a hit is pure loading of the dice.

It's gematria pure and simple. Tedious, I know. But Michel has told us numerous times that he suffers from para-schiz and refuses his meds. It would be offensive if one or any of us made such an accusation, and rightly so. One cannot hurl baseless accusations out of nowhere.

But that is not where we are at. No Michel has flat out stated that he is para-schiz and is off meds intentionally. The rest of us did not simply make that up out of whole cloth, Michel stated so. Now we are stuck with trying to figure out what to do with that. While being cognizant of the MA. Not an easy line to walk.

 

[JayUtah]

Nope, 1 was never a valid choice. 2,3,4,5 were valid, 1 was never in the basket. You owe me a t-shirt.

Oh, okay, I though 1-4 were the valid choices.

It's gematria pure and simple.

Yes.

 

[abaddon]

Double

 

[abaddon]

Oh, okay, I though 1-4 were the valid choices.

That was the previous test. For some reason Michel thinks selecting the range 2-5 this time somehow makes a difference to something, anything. I have no clue why he thinks that. We could as easily pick 3-6, or 113114671 - 113114675 or any other range. It is a daft idea. Yet somehow it is important to Michel.

I have no clue why that might be. And Michel is not willing to identify why.

Once again, WTAF????

 

[JayUtah]

They don't have to be numbers. They could be colors, or animals, or species of tree. Using numbers as the category then improperly invokes all the things we can do with numbers for other purposes, including numerology, all of which are entirely inappropriate when attempting to encode a categorical enumeration for data purposes.

 

[Michel H]

That was the previous test. For some reason Michel thinks selecting the range 2-5 this time somehow makes a difference to something, anything. I have no clue why he thinks that. We could as easily pick 3-6, or 113114671 - 113114675 or any other range. It is a daft idea. Yet somehow it is important to Michel.

I have no clue why that might be. And Michel is not willing to identify why.

Once again, WTAF????

I have been doing a very large number of similar tests in the past, on various websites, and in four different languages (English, French, German and Dutch), and I found that "1" was somewhat problematic, because it is too much associated in excellence (like being first in class at school, and so on).

So I thought it would be a good idea to remove this number, and to do a 2-3-4-5 test instead.

From a purely statistical point of view, just one or two successful tests do not provide a very good p-value, but, when you get similar results from dozens and dozens of tests, it's different ...

It is, however, usually important to carefully distinguish between the answers which sound serious and friendly on the other hand, and those which seem unreliable. In a number guessing experiment, the words which are said are very important, they give insight into people's states of mind.

Regarding the issue of mental disease, I assume the voices I often hear are probably from real people who talk to me telepathically, this is different from psychotic hallucinations.

Trying to convince me that my voices are internally caused by mental illness when people know otherwise is, in my opinion, something which has a criminal dimension. There are unfortunately many people who relish this kind of things.

I wanted to travel far away to see if these voices change, but this has been made (nearly) impossible by the pandemic (I am not yet vaccinated, I don't know if I'll get the jabs or not).

I observe many apparent telepathic phenomena every day, so it seems normal to try to do "something about this", there is also some social pressure for that.

People who participate seriously in my tests are actually (in my opinion) contributing to scientific research, and this something which is interesting, I think.

 

[JayUtah]

From a purely statistical point of view, just one or two successful tests do not provide a very good p-value, but, when you get similar results from dozens and dozens of tests, it's different ...

No, it isn't. You can't just willy-nilly aggregate p-values from different tests. That's not how samples work.

It is, however, usually important to carefully distinguish between the answers which sound serious and friendly on the other hand, and those which seem unreliable.

It's important to collect good data. It's absolutely cheating to look at each answer and decide whether it's good data after you see whether or not it might help get you to a desired outcome. It's pure dishonestly to concoct elaborate transformations from guesses obviously intended to be facetious to valid answers that are now also "hits."

You were shown how to collect good data. When you did so, your magical abilities mysteriously disappeared entirely, so you wrote off the proper protocols as too onerous for you to repeat. Your bias is loudly evident.

In a number guessing experiment, the words which are said are very important, they give insight into people's states of mind.

No. You have no idea what people's states of mind are, and you have a proven track record of missing obvious sarcasm and humor. You are uniquely unqualified to determine whether other people are serious or kidding.

Regarding the issue of mental disease, I assume the voices I often hear are probably from real people who talk to me telepathically, this is different from psychotic hallucinations.

How do you know that?

Trying to convince me that my voices are internally caused by mental illness when people know otherwise...

Who knows otherwise? Show us their evidence?

There are unfortunately many people who relish this kind of things.

Pure paranoia.

People who participate seriously in my tests are actually (in my opinion) contributing to scientific research, and this something which is interesting, I think.

Nobody is participating seriously in your tests. You need help.

 

[Jack by the hedge]

... It is, however, usually important to carefully distinguish between the answers which sound serious and friendly on the other hand, and those which seem unreliable.

No. It's really much more important *not* to distinguish. Certainly not unless the person making the distinction between friendly and unreliable is blinded to whether the answers are also right or wrong. Isn't that completely obvious?

 

[abaddon]

I have been doing a very large number of similar tests in the past, on various websites, and in four different languages (English, French, German and Dutch), and I found that "1" was somewhat problematic, because it is too much associated in excellence (like being first in class at school, and so on).

Irrelevant.

So I thought it would be a good idea to remove this number, and to do a 2-3-4-5 test instead.

Irrelevant.

From a purely statistical point of view, just one or two successful tests do not provide a very good p-value, but, when you get similar results from dozens and dozens of tests, it's different ...

Irrelevant. We already know that statistics are a far distant country to you.

You already accused me of a partial hit that might be significant, despite the fact that I actively avoided doing so. That is a lie. But we already have established that you have no problem with flat out lying about what people might say.

It is, however, usually important to carefully distinguish between the answers which sound serious and friendly on the other hand, and those which seem unreliable. In a number guessing experiment, the words which are said are very important, they give insight into people's states of mind.

2, 3, 4, 5.

You are going to say that I identified 2, call it a hit, and hurl accusations of dishonesty. That is how you rock. How do I know that? You work it out. You have lied about it before, you are lying about it now. You flat out claimed that you could interpret my responses as support for your nonsense. That is a lie.Stop lying about others.

Regarding the issue of mental disease, I assume the voices I often hear are probably from real people who talk to me telepathically, this is different from psychotic hallucinations.

Make your mind up. Your claim is that the entire planet hears your thought. Obviously false. No you are claiming you hear the thoughts of others. Can you hear my thoughts right now? Bet you don't. What number am I thinking about right now? You have no clue.

Trying to convince me that my voices are internally caused by mental illness when people know otherwise is, in my opinion, something which has a criminal dimension. There are unfortunately many people who relish this kind of things.

Who cares? We as a group have tried to help you, but you want none of it. I could have recommended, various folks have. You reject all of that and want none of it.

I wanted to travel far away to see if these voices change, but this has been made (nearly) impossible by the pandemic (I am not yet vaccinated, I don't know if I'll get the jabs or not).

The only way you are unvaccinated is because you chose it to be so. Rather sad. You are so committed to your nonsense that you are willing to put others at risk. That is not only embarrassing, it is flat out immoral. You should be ashamed of yourself.

I observe many apparent telepathic phenomena every day, so it seems normal to try to do "something about this", there is also some social pressure for that.

Really? Because you seem to do nothing, ever.

People who participate seriously in my tests are actually (in my opinion) contributing to scientific research, and this something which is interesting, I think.

You have no clue how science works.

 

[Michel H]

Make your mind up. Your claim is that the entire planet hears your thought. Obviously false. No you are claiming you hear the thoughts of others. Can you hear my thoughts right now? Bet you don't. What number am I thinking about

No, I am not claiming I hear the thoughts of others, I claim I (apparently) hear voices of others in my head, which is completely different.

 

[Michel H]

No, it isn't. You can't just willy-nilly aggregate p-values from different tests. That's not how samples work.

There is no difficulty in fact in combining results from different tests to obtain a new p-value.

 

[JayUtah]

There is no difficulty in fact in combining results from different tests to obtain a new p-value.

Emphatically no.

They are different samples, operating under different protocols and circumstances. They are not directly comparable. There is no one statistical model and dataset to which they all belong.

This is what I do for a living, Michel. You can't bluster your way past me.

 

[Startz]

Emphatically no.

They are different samples, operating under different protocols and circumstances. They are not directly comparable. There is no one statistical model and dataset to which they all belong.

This is what I do for a living, Michel. You can't bluster your way past me.

I have some hesitation about talking about actual statistics, but...

If you know something about the tests, such as that they are from independent data, what's the problem in combining p-values?

(Not that this is likely to ever actually happen in this thread.)

 

[arthwollipot]

My answer was a complete guess, uninfluenced by anyone or anything.

But let's be charitable. I got one right. I had a 25% probability of getting it right just based on random chance alone. We need to repeat the test.

 

[junkshop]

No, I am not claiming I hear the thoughts of others, I claim I (apparently) hear voices of others in my head, which is completely different.

Given that your eardrums are in your head, this would be an entirely expected place to hear the voices of other people. However, as I understand it, these voices that you hear are not those of people in your vicinity speaking out loud, and being heard by everyone within earshot, but rather a psychic phenomenon, only 'audible' to you.

If I'm following your claims correctly, you hear the voices of some people in your head, while everyone else hears your thoughts in theirs. You say that these things are "completely different". How so?

I have more questions, but they can wait.

 

[Michel H]

Emphatically no.

They are different samples, operating under different protocols and circumstances. They are not directly comparable. There is no one statistical model and dataset to which they all belong.

This is what I do for a living, Michel. You can't bluster your way past me.

Let's assume I get 5 credible answers in test 1 (which is a test with 4 possible answers, like 1 or 2 or 3 or 4, or {A or B or C or D}), and that 4 out of these 5 are correct. I then calculate a certain p-value, say p₁, using for example this binomial calculator: https://stattrek.com/online-calculator/binomial.aspx (it is equal to p₁=0.015625).

Next I do a similar test, test 2 (with again four possible answers), 4 are right out of 6 credible, I get p₂.

Now I can also easily combine the results of the two tests. The new p-value is calculated from 4+4=8 hits in 5+6=11 trials, there is no difficulty at all in calculating it.

For an introduction to my method, I suggest you read my first two test analyses on this forum:
http://www.internationalskeptics.com/forums/showthread.php?postid=8607740#post8607740
http://www.internationalskeptics.com/forums/showthread.php?postid=9516155#post9516155.

 

[Gord_in_Toronto]

My answer was a complete guess, uninfluenced by anyone or anything.

But let's be charitable. I got one right. I had a 25% probability of getting it right just based on random chance alone. We need to repeat the test.

Just keep doing it until you get it right. Like 5 times in a row maybe. Then we'll be on to something.

 

[JayUtah]

If you know something about the tests, such as that they are from independent data, what's the problem in combining p-values?

If you know something about the tests, which we do in this case, then knowing that they involved different subjects, different protocols, different uncontrolled circumstances, and different methods of manipulating the data prior to the accounting, we cannot combine them. One never "combine p-values." The p-value is the computed probability that the null hypothesis still governed the outcome, but that the outcome appears to confirm the experiment hypothesis. One therefore combines the data that went into that computation, but only insofar as they come from congruent protocols.

Here, for example, one group of subjects guessed from among numbers 1-10. A different group guessed from among four numbers, variously enumerated. Since the probability of a false positive is different in each trial, you can't aggregate those data.

Customarily you do not aggregate data from different subject pools. For example, telekinesis tests are commonly done with the same test protocols, but using different subject pools at different times and places. We consider them separate experiments and actually compare the p-values to see if they are congruent. We don't aggregate the data, even if the statistical model is the same in each case.

(Not that this is likely to ever actually happen in this thread.)

Not likely, no.