New telepathy test, the sequel.

[Jack by the hedge]

It may be possible to reach a conclusion as to whether Michel H is really telepathic or not in a cumulative way, by doing a large number of simple tests, or by looking at the various kinds of evidence, for example what people say during tests or outside tests.

Also don't forget to consider the cumulative evidence of what people don't say.

For example, people on social media don't say "Can anyone else hear Michel's thoughts in their head?".

Perhaps the clearest evidence that you do not possess a remarkable power is precisely the fact that people do not remark on it.

 

[JayUtah]

Yes, that's right, any (probability) number, provided it is very small.

No, you don't know that. You're trying to pull numbers out of your kiester and call it data. And people are rightly not letting you do that and call it science. Your response is to try to attach a social or emotional penalty to criticizing you, not to fix your science.

...a serious-sounding post, as opposed to a sarcastic one, as some skeptics are trying hard to rebrand it...

No. There have been ample tests here to demonstrate you can't tell a serious post from a sarcastic one. You therefore don't get to be the judge of what someone else's original intent may have been. In any case, the original poster has informed you of his original intent, in case there was ever any question. After that happens, it's just pure arrogance for you to suppose you still know best.

...and the most obvious explanation is that Michel H is really telepathic...

No, that is not the most parsimonious hypothesis. But it's obviously the one you really, really want to believe, so you're blind to its prima facie absurdity. The most parsimonious hypothesis is, unfortunately, the one we can't talk about. There is a straightforward explanation for hearing voices in one's head, and it has nothing to do with telepathy, the CIA, or world peace.

...though not because my test protocols are unscientific, because that's fake news...

No, it isn't. You've been given specific reasons why your protocols are unscientific, and you don't address any of them. You just get very mad and hurl insults at the people who raise them. Then you wave a decades-old diploma as if this certifies that you are the only scientist in the room. When you can address the actual reasons instead of stamping your foot and insisting that you must be a great scientist, then perhaps you'll have a case.

 

[RoboTimbo]

It may be possible to reach a conclusion as to whether Michel H is really telepathic or not in a cumulative way, by doing a large number of simple tests, or by looking at the various kinds of evidence, for example what people say during tests or outside tests.

Or by taking a fair test, which is what an honest person would do. You know, rather than making more laughable "four choice" tests and then applying biased "credibility ratings" after the answers are known, which is dishonest.

 

[JayUtah]

It may be possible to reach a conclusion as to whether Michel H is really telepathic or not in a cumulative way, by doing a large number of simple tests...

No, confidence doesn't accumulate that way. If I ask a large number of people to pick a random number between 1 and 10, I may wrongly arrive at the conclusion that there is some strange phenomenon involved, because in that sort of test the responses you get from humans are not evenly distributed. The most common answers to that question are 3 and 7. It's not telepathy; it's just how people think quantitatively by nature when asked to approximate randomness. You need to be able to discover and control for those kinds of effects if you want to base your conclusion on some number of single-trial tests.

Further, since your hypothesis is that telepathy attaches either to a single person or to pairs of persons, doing large numbers of single-trial tests, where each of many subjects performs only one trial, tells you nothing about single persons. You can't accumulate data from Tom, Dick, and Harry, aggregate it, and then apply it all to Tom. Further, as we go, you change your hypothesis to describe different mechanisms and rules by which telepathy is suggested to operate. All of these are post hoc revisions, intended apparently to explain away disconforming data. You need to abandon the chimeric hypothesis approach, pick one, and enlist the help of practicing scientists to help you design a working protocol. That is, if a fair test is what you really want.

...or by looking at the various kinds of evidence, for example what people say during tests or outside tests.

That requires the experimenter to demonstrate some aptitude in evaluating what people say. You clearly don't have that aptitude. It also requires the experimenter to detect and avoid post hoc biases, such as those that are painfully evident in your approach. You selectively pay attention to what people say outside tests and accept only the data that supports the conclusion you want. Here's Loss Leader falling all over himself telling you outside the test that what he said inside the test was intended as mockery -- and we can all see that it was -- but you give that statement no weight. You simply insist that you know better than the speaker what the speaker really meant.

Finally, a statistical basis for such an approach is incredibly difficult to achieve to any degree of scientifically cognizable rigor. Even when one instigates a coding scheme, employs a panel of evaluators to minimize bias, and applies all the other kinds of blinding controls we've mentioned, the shaky statistical basis for open-ended evaluation generally dooms any attempt at serious research that tries to employ it.

 

[JayUtah]

It may be possible to reach a conclusion as to whether Michel H is really telepathic or not in a cumulative way, by doing a large number of simple tests, or by looking at the various kinds of evidence, for example what people say during tests or outside tests.

Are you proposing this because the data-driven approach isn't giving you the answer you wanted? So many of your proposals seem aimed at fixing data you collected in the field, usually in ways that permit subjective or biased influences. Basic scientific prudence suggests that if Michel really wants to believe he's telepathic, he probably shouldn't be in charge of any experiment intended to prove that ability to others. Even if others are involved, a fair test should really have its protocols designed and administered by someone else.

In contrast you set yourself up as the sole judge, happily suggesting that your subjective opinion regarding other people's behavior and intent should attain some sort of scientific rigor. You simply announce that based on academic work from the 1980s you are a great scientist and insinuate that this means any protocol you put forward should be considered to have scientific validity. Even if those weren't enough red flags, when it comes down to doing the actual evaluation you blatantly ignore evidence that disconfirms your evaluation, and try to sweep it under the carpet with feeble categorical claims. The protocol you propose is broken in its design and broken in its execution.

 

[Nay_Sayer]

Or by taking a fair test, which is what an honest person would do. You know, rather than making more laughable "four choice" tests and then applying biased "credibility ratings" after the answers are known, which is dishonest.

Before he ran off his last proposed idea, iirc, was going to be 1-3.

So now the only real question that remains is how much time is everyone else willing to waste talking with Michel?

 

[JayUtah]

Doing a large number of simple (4 answer tests) doesn't take away from the fact that someone still has a 25% chance to guess the answer and you will have done nothing to determine whether correct answers where derived by telepathy or chance/guessing.

I'm not sure, but it seems like he wants to say that if a large number of people all do one-trial tests, and more than a quarter of the "credible" guesses succeed, it can prove he's telepathic. The obvious problem with that proposal is where he said that telepathic abilities vary from person to person and across pairs of persons. Thus if the degree of his telepathic ability is unknown and therefore presumed to be what's under test, and the distribution of telepathic ability across the subject pool is an unknown -- as it would have to be for a crowd-sourced convenience sample -- and the factors affecting the pairwise display of telepathic ability are unknown, then the model has way too many degrees of freedom to provide a solution.

In contrast, if he were a known telepath and each comer had several trials to guess what he's sending, then this could eventually converge to a useful understanding of whether the other subjects were telepathic. Or conversely if all the potential receivers were known telepaths, then his ability to send could be put meaningfully under test.

The insidious problem with the method is that people presented with a list of choices and forced to select from them at random do not do so truly at random. Writers of important multiple-choice tests and questionnaires -- especially psychometric instruments -- have studied this phenomenon at length, and generally develop controls against selection bias. Without such controls, the experimenter might accidentally put the "success" alternative in a selection hot spot and then wrongly attribute the perception of a greater-than-chance response to statistical significance of the hypothesis.

It seems his strategy is to try to hide all the confounding variables in plain sight, wave them away by post hoc fiat (along with data that he doesn't like), and then demand that since he once got a PhD forty years ago there can be no meaningful criticism of his method -- especially from those grotesquely dishonest, unqualified, mean ol' skeptics.

 

[Startz]

Hi Michel,

I've been thinking of this all in a different way. As a practical matter, the way science often proceeds is that the scientist does some preliminary work to demonstrate to themselves that a phenomenon is actually there. After that, the scientist (already convinced they have the right answer) does experiments where the purpose is to convince others.

Here's an example. I recently had an idea about the price responsiveness of the demand for 2-year versus 4-year college enrollment. I had an undergraduate put together some relatively easy to access data and ran some preliminary statistical estimates. I was pleased to get very strong results that supported my suspicions. Now my next step will be to get a graduate student to put together some much more complex data and we'll do more sophisticated statistical work. I am quite confident the results will be the same, but I need to do this second stage to convince others.

You've done some simple tests that convince you about telepathy. All fine. Now the normal scientific next step is to do more standard tests to convince others.

Are you up for that?

 

[JayUtah]

When such an extremely improbable event does occur, this leads you to suspect that some unusual phenomenon or effect is at play...

Yes, but not necessarily the (possibly) unusual effect that you've enshrined in your hypothesis. The problem with practical empiricism is that it's nearly impossible to control for all possible-but-unusual effects that aren't your hypothesis. This is why we rely first on very rigorous protocol design, to reduce them to a manageable set, and second upon data-driven statistical methods that allow us to reason reliably in the face of the inevitable uncertainty that will arise.

You aren't interested in either. You deliberately avoid any sort of control that would eliminate unwanted effects, even those that would be trivial to eliminate. You propose to deal with them using subjective post hoc analysis -- a major no-no even when the experimenter doesn't have a stake in the outcome. And when the statistics don't work out the way you want, you revise your subjective post-experiment data-culling strategy ad hoc. You try to describe your personal opinions as if they could be quantified as a probability that overrides collected data and "fixes" the statistics. What you're doing is no more scientifically valid than if you had simply said, "It's my opinion that the p-value of my experiment should be taken as < 0.05."

I struggle with how you expect an intelligent person to come away with anything but the idea that you're trying to dress up a bunch of speculation and wishful thinking to make it look like science, and hope you can browbeat, gaslight, and bluster your way past any criticism. It really is not that hard to be a good scientist, and your evident presumption that no one can confidently see the obvious flaws in your methods really doesn't save you.

 

[Nay_Sayer]

Hi Michel,

I've been thinking of this all in a different way. As a practical matter, the way science often proceeds is that the scientist does some preliminary work to demonstrate to themselves that a phenomenon is actually there. After that, the scientist (already convinced they have the right answer) does experiments where the purpose is to convince others.

Here's an example. I recently had an idea about the price responsiveness of the demand for 2-year versus 4-year college enrollment. I had an undergraduate put together some relatively easy to access data and ran some preliminary statistical estimates. I was pleased to get very strong results that supported my suspicions. Now my next step will be to get a graduate student to put together some much more complex data and we'll do more sophisticated statistical work. I am quite confident the results will be the same, but I need to do this second stage to convince others.

You've done some simple tests that convince you about telepathy. All fine. Now the normal scientific next step is to do more standard tests to convince others.

Are you up for that?

No. He believes all his tests are more than sufficient for everyone and is confounded as to why his esp abilities are not as readily obvious to everyone as they are to him.

Hilited is said in context, Michel does not have ESP abilities and I do not want this taken out of context by him.

 

[JayUtah]

So now the only real question that remains is how much time is everyone else willing to waste talking with Michel?

As with all such threads, for as long as it remains entertaining to do so. We slow down to see the aftermath of a road accident, but eventually there's no more to see and we speed up again to pursue more highbrow interests. We watch Gaslight to see Charles Boyer try to berate poor Ingrid Bergman into submission, sometimes because we like the ending and other times because it's a cautionary tale. (And no matter how hard they tried, they couldn't make the 1940 version go away.)

When watching Michel make a laughingstock of himself isn't the order of the day, these threads often provide for others interested in experiment design to discuss how it's done, with the OP's specific scenario in mind. It ends up being entertaining and useful to some.

 

[JayUtah]

I do not want this taken out of context by him.

Aw, when has he ever taken anything out of context?

 

[Darat]

Just as a little bit of background.

This forum even when it had the name JREF at the top was not part of the official JREF activities.

None of the moderators here were ever appointed by anyone who worked for the JREF.

The JREF never had anything to do with Loss Leaders appointment as a moderator.

No moderator on the forum has ever spoken for the JREF.

There was no "prestige" associated with being a moderator on the JREF forum.

For the record I was the last person ever to be appointed to the moderating team by the JREF.

 

[Loss Leader]

It is true that Loss Leader seems sometimes to have something of a double personality, with a "Loser" side and a "Leader" side, as he probably understands it himself.

A "Loss Leader" is the part of a magnetic computer tape that has no magnetic coating. It is meant to be threaded through the reader and into the receiving tape reel. Coming first, it is the tape's "leader." And, since it has no magnetic coating, any information one tried to record on it would be "lost." Thus, early computer programmers were advised to be sure they started recording after the "loss leader."

I chose the handle some twenty years ago after reading Count Zero by William Gibson. In that novel, the main character chose his handle "Count Zero" by reading a book on old computer code that said, "When receiving an interrupt, decrement the counter to zero." As I didn't want to just steal Gibson's character's name, I got a book of old computer code and found the passage that reminded programmers to wind the computer tape past the loss leader.

Here's a man threading one:

Here's a woman surrounded by them:

As to the job and qualifications to become a moderator:

My only job as a moderator is to enforce the Membership Agreement. You will note that the Agreement says nothing about "promoting critical thinking and providing the public with the tools needed to reliably examine paranormal, supernatural, and pseudoscientific claims."

In fact, much of my job has been to protect people who espouse ridiculous claims from personal and uncivil attacks. In my time as a moderator, I have protected anti-vaxers, 9/11 conspiracy theorists, JFK conspiracy theorists, electric universe proponents, bigfoot hunters, Amanda Knox guilters, pretend psychics, straight-up Nazis, and more. None of these people shared my personal beliefs. None of them were critical thinkers. Many of them actually held positions offensive to me personally.

You, for example, have been targeted multiple times by people questioning or diagnosing your mental health. I helped remove those posts to keep your thread about testing your pretend paranormal powers.

My only tools as Moderator are editing posts, moving offending posts, and infracting people who break the Membership Agreement. Were I to edit posts to suit my own views, move or delete posts that I disagreed with, or take any other action based on my personal preferences, I would be in violation of the Membership and Moderator Agreements and would be swiftly kicked off the team and likely banned. I know because it has actually happened.

The requirements for being selected as a Moderator are only these: that the person volunteers; that he/she has shown a calm demeanor; that he/she has shown an overall ability to abide by the Membership Agreement; and that he/she is active on the forum.

I am not allowed and have not used my position as moderator to favor or disfavor any position, whether it is skeptical or not.

Your beliefs that a moderator would be more likely to give a credible answer are wrong. We're allowed to lie. We're allowed to be sarcastic. We're allowed to be outright scornful.

Just as a little bit of background. This forum even when it had the name JREF at the top was not part of the official JREF activities. None of the moderators here were ever appointed by anyone who worked for the JREF. The JREF never had anything to do with Loss Leaders appointment as a moderator. No moderator on the forum has ever spoken for the JREF. There was no "prestige" associated with being a moderator on the JREF forum. For the record I was the last person ever to be appointed to the moderating team by the JREF.

And there you have it.

 

[Michel H]

What's the difference between "proving" something and "convincingly proving" something?

Doing a large number of simple (4 answer tests) doesn't take away from the fact that someone still has a 25% chance to guess the answer and you will have done nothing to determine whether correct answers where derived by telepathy or chance/guessing.

There are some well known statistical tools to deal with these issues (binomial distribution, calculation of p-value and so on). Ganzfeld experiments are also successions of little tests, called trial or sessions, with a 25% probability of success from random chance alone for each test: https://en.wikipedia.org/wiki/Ganzfeld_experiment.

If your hit rate is for example 50% (instead of the 25% expected from random chance alone), and your sample is large enough, then one generally says that an ESP effect has been found.

 

[RoboTimbo]

There are some well known statistical tools to deal with these issues (binomial distribution, calculation of p-value and so on). Ganzfeld experiments are also successions of little tests, called trial or sessions, with a 25% probability of success from random chance alone for each test: https://en.wikipedia.org/wiki/Ganzfeld_experiment.

If your hit rate is for example 50% (instead of the 25% expected from random chance alone), and your sample is large enough, then one generally says that an ESP effect has been found.

None of your tests has returned results better than chance.

 

[JayUtah]

There are some well known statistical tools to deal with these issues...

Yes. Assume your readers are quite familiar with them.

If your hit rate...

It was explained to you at length how your proposed trials are structurally very different from a ganzfield trial. You didn't address it, which means you either know your experiment is bogus for that reason, or you don't know that -- and therefore don't know how to construct a suitable statistical model for this type of research.

 

[p0lka]

Why do you doubt a Faraday cage would work?

because the tinfoil hat on their head doesn't stop the voices?

 

[Startz]

A "Loss Leader" is the part of a magnetic computer tape that has no magnetic coating. It is meant to be threaded through the reader and into the receiving tape reel. Coming first, it is the tape's "leader." And, since it has no magnetic coating, any information one tried to record on it would be "lost." Thus, early computer programmers were advised to be sure they started recording after the "loss leader."

I chose the handle some twenty years ago after reading Count Zero by William Gibson. In that novel, the main character chose his handle "Count Zero" by reading a book on old computer code that said, "When receiving an interrupt, decrement the counter to zero." As I didn't want to just steal Gibson's character's name, I got a book of old computer code and found the passage that reminded programmers to wind the computer tape past the loss leader.

This is totally fascinating. I had no idea--I've definitely learned something from this thread.

Also, with apologies for any excess thread drift, does anyone know of a source for the plastic canisters that used to be used ship mag tapes?

 

[Michel H]

Since you consider that no poster is ever sarcastic, why do you apply your "credibility ratings" to answers after you know whether they are correct or not? Why not just take all answers as given?

When I assign credibility ratings (see for example the analysis of my second test on this forum:
http://www.internationalskeptics.com/forums/showthread.php?p=9516155#post9516155), I am careful to do it regardless of the correctness of the post (perhaps like a mod who would examine whether a post complies with the Member Agreement regardless of whether the poster is a Democrat or a Republican).

I believe, and continue to believe that this is not only very doable, but also easily verifiable by readers who follow the work.

In a telepathy experiment, if friendly and serious-sounding answers tend to be more accurate than those which seem hostile, aggressive, absurd or crazy, it would be a monumentally foolish error to not separate out the friendly and serious-sounding answers, with the hope that their hit rate will be higher than the random-chance level of 25%. I generally find that this procedure leads to a significant improvement (see the second test for an example). More generally, observing a correlation between credibility and accuracy is strongly suggestive of a real ESP effect, this is why these considerations are always central in any test of Michel H's alleged propensity to communicate his modest thoughts to others.