Merged New telepathy test: which number did I write ?

[AdMan]

Generally speaking, I suspect many of the people here who claim to "represent Science" or to be "guardians of Science" actually have probably very little (if any) real experience of scientific research, and this is a problem, to be constantly taught lessons be people who clearly don't know much about science, but naively believe (perhaps) they know everything because they post on a "skeptical" forum (there have been some good remarks too). One important rule of actual scientific research is that it is necessary to be very careful (not "sloppy" or negligent.)

Michel, I have been very critical in this thread of your extremely sloppy (you used the word, ironically) testing. Perhaps you are criticizing my posts that are critical of your tests when talking about "people who clearly don't know much about science."

If so, you have no idea what you are talking about. I have a bachelor of science degree in biology (plus all the chemistry, physics, etc., required for pre-med). I know what science requires. And I can say that you demonstrate very little actual understanding of the scientific method, or even of basic probability theory to properly analyze your results.

So stop pretending that you know more than people who are actually knowledgeable about things you are posting about. That post was insulting.

 

[dlorde]

The results of this test seem to support my telepathy hypothesis (in my opinion, at least),

Well frankly, your opinion is wrong. With a valid sample size of 3 (less than the number of choices available!), the results are valueless.

For the next experiment, toss three coins and claim that getting two or more heads (or tails) 'seems to support' the coin being biased that way.

 

[Michel H]

Michel, I have been very critical in this thread of your extremely sloppy (you used the word, ironically) testing. Perhaps you are criticizing my posts that are critical of your tests when talking about "people who clearly don't know much about science."

If so, you have no idea what you are talking about. I have a bachelor of science degree in biology (plus all the chemistry, physics, etc., required for pre-med). I know what science requires. And I can say that you demonstrate very little actual understanding of the scientific method, or even of basic probability theory to properly analyze your results.

So stop pretending that you know more than people who are actually knowledgeable about things you are posting about. That post was insulting.

Ah ok, so you have a bachelor of science degree in biology. That's pretty good (in view of your name, I was wondering if you had a master's degree in advertisement, something like that), although it's not a doctorate (Ph.D.). You may have never done scientific research in your life, and this might explain why you may have difficulty understanding what I am trying to do here. Generally speaking, everybody is welcome in my threads, but, if you disagree, or have (possibly serious) doubts, I think you should explain the reason for these (something you did not do in your posts #122 and #140, for example), in order to be constructive, to bring something useful in the thread. In this post, you brought a piece of information to the community (albeit in a rather aggressive fashion), so I think that's good.

 

[Tamlin]

You may have never done scientific research in your life, and this might explain why you may have difficulty understanding what I am trying to do here.

Have you ever done scientific research in your life, Michel?

 

[Michel H]

... If you had had only 2 valid answers (by your reasoning), and one of them had happened to be correct, would you have declared that the results were 50% correct and therefore evidence of telepathy?

That's a good question, Emily. The p-value in this case is p = 44% (link), which is really large. I think I would rather say possible evidence for telepathy. If such a result was repeated many times, it might lead to strong evidence (a "proof").

 

[Michel H]

Have you ever done scientific research in your life, Michel?

Yes, I have a Ph.D. degree in physics from a large U.S. university. I have done some more research in physics at home in recent years, but I haven't had time to publish it yet.

 

[dafydd]

Ah ok, so you have a bachelor of science degree in biology. That's pretty good (in view of your name, I was wondering if you have a master's degree in advertisement, something like that), although not it's a doctorate (Ph.D.). You may have never done scientific research in your life, and this might explain why you may have difficulty understanding what I am trying to do here. Generally speaking, everybody is welcome in my threads, but, if you disagree, or have (possibly serious) doubts, I think you should explain the reason for these (something you did not do in your posts #122 and #140, for example), in order to be constructive, to bring something useful in the thread. In this post, you brought a piece of information to the community (albeit in a rather aggressive fashion), so I think that's good.

We all know what you are trying to do here and you have failed miseably.

 

[Michel H]

that was a vague suggestion of a possible thing to do. Not a requirement.

Well, it was recommended. And you did it quite well.

 

[dafydd]

Well, it was recommended. And you did it quite well.

When will you be carrying out a telepathy test quite well?

 

[Michel H]

When will you be carrying out a telepathy test quite well?

Quite well? Well, that, I don't know... I believe I have done a reasonably good job so far, but it is perhaps difficult to judge oneself objectively. Also, I prefer not to go too fast. I believe in the importance of team work, so I want to make sure I have answered everybody properly before launching a new test. I am also thinking of starting a new thread (with a poll), asking people the kind of test they would prefer.

 

[Michel H]

Well frankly, your opinion is wrong. With a valid sample size of 3 (less than the number of choices available!), the results are valueless.

For the next experiment, toss three coins and claim that getting two or more heads (or tails) 'seems to support' the coin being biased that way.

Well, dlorde, the p-value of this example you take is p=0.50=50% (this link), which is fairly large. In my analysis of post #149, I got p=15.6%. This roughly means that I had less than one in five chance of obtaining a result this good, and yet I got it. Of course, such evidence must be repeated before statistical significance is reached.

 

[AdMan]

Yes, I have a Ph.D. degree in physics from a large U.S. university. I have done some more research in physics at home in recent years, but I haven't had time to publish it yet.

It's amazing, then, that you apparently fail to understand the scientific method.

Your "credibility rating", which you assign arbitrarily, is not science. Neither is changing the protocol after the test in order to change your results to your liking.

I'd ask for my money back on that degree.

 

[Michel H]

It's amazing, then, that you apparently fail to understand the scientific method.

Your "credibility rating", which you assign arbitrarily, is not science. Neither is changing the protocol after the test in order to change your results to your liking.

I'd ask for my money back on that degree.

Credibility rating: I try to use it here within scientific procedures, so I would call it scientific. It doesn't matter if it cannot be defined with great precision (this is psychology, not physics). I think it is a useful concept if there is a correlation between the credibility ratings (with their uncertainties, and preferably chosen in a "blinded" way) and correctness in the numerical answers.
You may find this article by Dan Gillmor, on the Guardian website, interesting: http://www.theguardian.com/commentisfree/cifamerica/2011/jun/08/andrew-breitbart-anthony-weiner .
I have never posted it, although it talks about "negative credibility". I quote:

How can someone raise his credibility to zero? By having negative credibility in the first place.

In my book Mediactive, I have a chart, a version of which is reproduced here, that I call a "credibility scale". It's essentially a meter I keep in mind when evaluating what and who to trust in the media I follow. You will note that the scale does not start at zero and run upward. It starts strongly negative and extends past zero to strongly positive.

I put anonymous comments on random blogs in deep-negative territory. They would have to work incredibly hard just to get up to zero.

On the other end of the scale you'll notice, as an example, the BBC. It, like the Guardian and Telegraph and New York Times and a number of other organisations, starts in generally positive territory – not uniformly so, but with my presumption that the reporting is much more likely to be accurate, thorough and fair, among other qualities, than not. A presumption of credibility extends beyond our best traditional media organisations. I follow some bloggers who rate even higher, because in domains where I'm knowledgeable, they've proved to me that they consistently get things right and in context.

Protocol of the test: I didn't change it after the test. What I did do, however, was to accept answers by participants who posted no MD5 hash, without knowing who (if anyone) had answered correctly (only one person had posted a hash but that person never posted the text used to generate the hash, so it was useless).

 

[dlorde]

... Of course, such evidence must be repeated before statistical significance is reached.

Until then, you cannot reject the null hypothesis (that you are not telepathic). The p-value says nothing at all about the probability of the alternate hypothesis (that you are telepathic) being true.

 

[TheSapient]

Credibility rating: I try to use it here within scientific procedures, so I would call it scientific. It doesn't matter if it cannot be defined with great precision (this is psychology, not physics).

"Scientific" implies repeatability. Anyone should be able to apply your credibility protocols to the same information and get the same, or similar, results.

Please post protocols you use. I will take them and go back to posts you have already analysed and see how close I come to your ratings.

I have to admit, I don't think you will do this. I don't think you have protocols. I think you create and abandon rationalizations for each individual response to your test.

 

[abaddon]

Credibility rating: I try to use it here within scientific procedures, so I would call it scientific. It doesn't matter if it cannot be defined with great precision (this is psychology, not physics). I think it is a useful concept if there is a correlation between the credibility ratings (with their uncertainties, and preferably chosen in a "blinded" way) and correctness in the numerical answers.
You may find this article by Dan Gillmor, on the Guardian website, interesting: http://www.theguardian.com/commentisfree/cifamerica/2011/jun/08/andrew-breitbart-anthony-weiner .
I have never posted it, although it talks about "negative credibility". I quote:



Protocol of the test: I didn't change it after the test. What I did do, however, was to accept answers by participants who posted no MD5 hash, without knowing who (if anyone) had answered correctly (only one person had posted a hash but that person never posted the text used to generate the hash, so it was useless).

Absolute rubbish. You blatantly manipulated results to get the conclusion you wanted. You blatantly manipulated your idiotic credibility scores to get the result you wanted.

 

[TheSapient]

Yes, I have a Ph.D. degree in physics from a large U.S. university. I have done some more research in physics at home in recent years, but I haven't had time to publish it yet.

A bit of advise. No one really cares that you openly, blatantly, mess with the data to get the results you want when it comes to pretend telepathy. If you try this in physics, people will indeed care. You will kill any chance of a career in the field. Previous work will be closely examined. The world of science is very different than whatever it is you are doing here.

 

[TheSapient]

Michel,

I am still curious about this.

Michel,

Can you describe the mechanism behind your latest objection? You are saying you can send a number into the mind of a remote person.....unless they later send someone a private message in an internet forum that is any different than the wording they used to report their results in that same forum?

How does their PM block your ability? How long must they refrain from sending PMs to other users? If they wrote something else in some other medium, say a shopping list, what would happen to your telepathy?

Does their PM go back in time and stop you from sending your number? Or does your telepathy look into the future and find out that participants have written something?

 

[Michel H]

"Scientific" implies repeatability. Anyone should be able to apply your credibility protocols to the same information and get the same, or similar, results.

Please post protocols you use. I will take them and go back to posts you have already analysed and see how close I come to your ratings.

I have to admit, I don't think you will do this. I don't think you have protocols. I think you create and abandon rationalizations for each individual response to your test.

No, I don't have "credibility protocols".
This is an example of an answer I found credible:

... I do indeed have ESP, and know for a fact that he wrote 2!

, and this is an exemple of an answer that I did not find credible:

The last time I partcipated in this guessing game I was accused of being in a mental institution and therefore my response was invalid.

I'm now in a high-sided elastic banjo with an eskimo parasol, so rest assured that my response is both fluffy and perky.

The number I'm seeing is XX.

I think it is likely I am not the only one who can come to these conclusions.

 

[bruto]

I beg of you. Please respond to this. All of you. I have a bike combo padlock. Each of its 4 dials has numbers 1 thru 6 on it. What`s the combination?

It's a pretty good start, but I would point out that it also has a certain resemblance to a classic bunko scheme, though that scheme does not usually use anything as complex as a lock combination. Still, if you had a large enough population to scam, you might get away with it.

In its simple form, it goes something like this:

You contact a number of people and tell them you're a prophet. You then predict the 50/50 outcome of some public event. Actually you guess one side for half, and the other side for the other half. You're guaranteed that at least half your victims will see a correct prediction. With the remaining half, you do this again, and again you halve your number of victims. You keep doing this until you're left with one victim, to whom you will appear to have made a long string of uncanny predictions. You then ask for lots of money on the next bet. If it comes out your way, you do it again. When you lose, you run with the money.

Of course no internet psychic will take up your challenge because it's too much of a challenge really. Make sure you write that combination down!