Cold Reading Demos at TAM2

[CFLarsen]

I'm sorry, but I just don't know what you are driving at. What is the flawed data that you are talking about here?

Very well, if you want to play games:

You have claimed that Hoyt's data were flawed.

You have used your own data in a Poisson calculation and compared it with Hoyt's Poisson calculation.

So, if Hoyt's data were flawed, what use is the comparison?

 

[Thanz]

Very well, if you want to play games:

You have claimed that Hoyt's data were flawed.

You have used your own data in a Poisson calculation and compared it with Hoyt's Poisson calculation.

So, if Hoyt's data were flawed, what use is the comparison?

I'm not playing games, I just wanted to be sure I knew what you are talking about.

You seem to be confused about whatever comparisons I may be making about the two results.

I am complaining about the method that Hoyt used to collect his data. I point to my count and analysis as being correct. I explain in plain terms why Mr. Hoyt's is wrong. I then point to his results and flawed count to show the effect of his error. I am not doing any sort of meta-analysis on the two results.

Any comparisons are to show the effect of Hoyt's errors. I do not rely on Hoyt's results to say anything substantive about JE and cold reading.

 

[CFLarsen]

I'm not playing games, I just wanted to be sure I knew what you are talking about.

I thought that was very clear. I can't really see where the confusion lies.

You seem to be confused about whatever comparisons I may be making about the two results.

I am asking you what the purpose of the comparison is. There is no confusion, I am merely asking you what the purpose was.

I am complaining about the method that Hoyt used to collect his data. I point to my count and analysis as being correct. I explain in plain terms why Mr. Hoyt's is wrong. I then point to his results and flawed count to show the effect of his error. I am not doing any sort of meta-analysis on the two results.

Any comparisons are to show the effect of Hoyt's errors. I do not rely on Hoyt's results to say anything substantive about JE and cold reading.

No, no, no....how can a comparison be of any use, if one part of it is based on flawed data?

I have asked you this several times. You seem unwilling - or just unable - to answer that. Could I ask you, once again, to explain this:

If Hoyt's data were flawed, what use is the comparison?

You complain about the method that Hoyt used, yet you did not even question his data. So, what use is the comparison?

 

[Skeptical Greg]

Until this analysis question is answered, can we assume there is a possibility that JE and others really do talk to the dead?



P.S.

Billy and Thanz,

You both seem to be too intelligent to be wasting so much time on this..

Why don't you look around and see if there is another discussion or two you could contribute to...

 

[Thanz]

I have asked you this several times. You seem unwilling - or just unable - to answer that. Could I ask you, once again, to explain this:

If Hoyt's data were flawed, what use is the comparison?

You complain about the method that Hoyt used, yet you did not even question his data. So, what use is the comparison?

And I have tried to answer several times. I think that perhaps I just don't understand the question. What comparison are you talking about, specifically?

I think that any comparisons I have done are to show the effect of Hoyt's flaws on the analysis. So, the purpose would be to point out the effect of Hoyt's error on the analysis.

As for not questioning his data, of course I question his daata - if by data you mean his count. I saw his method is flawed and his count is wrong. His count is wrong because his method is flawed, and therefore his result cannot be used to say anything meaningful about JE and cold reading. I thought that my position on this was clear.

 

[Thanz]

Until this analysis question is answered, can we assume there is a possibility that JE and others really do talk to the dead?

Well, I think only in the "anything's possible" sort of way. And I don't think that this analysis would put an end to the question no matter what the results are.

Why don't you look around and see if there is another discussion or two you could contribute to...

I seem to remember something about pruning old threads... perhaps I should start a new thread about that?

 

[CFLarsen]

And I have tried to answer several times. I think that perhaps I just don't understand the question. What comparison are you talking about, specifically?

I have only spoken of one comparison: Your own. The one you did on your own data, using Poisson, compared to Hoyt's data, also using Poisson.

I really, really cannot see why this is confusing to you, Thanz. There is only one comparison here. The one you made.

Do you understand this now? YOUR COMPARISON!

I think that any comparisons I have done are to show the effect of Hoyt's flaws on the analysis. So, the purpose would be to point out the effect of Hoyt's error on the analysis.

But, according to you, the comparison you did was based on flawed data. So, what does it prove? Again, you avoid this question.

As for not questioning his data, of course I question his daata - if by data you mean his count. I saw his method is flawed and his count is wrong. His count is wrong because his method is flawed, and therefore his result cannot be used to say anything meaningful about JE and cold reading. I thought that my position on this was clear.

Not really. Because you have yet to explain how your comparison has any value, if it is based on flawed data!

Why is it that you cannot answer this simple question? You understand it perfectly, Thanz. But you cannot answer it.

 

[BillHoyt]

It would be good if you could be a little bit more specific here. What is a markov chain? Does it make any difference in the decay if it is a primary collision or a secondary collision? What exactly are you studying with the Poisson distribution?

Sorry, but as I said, I don't really know anything about radioactive decay.

You are studying radioactive decay events. Most of the time, a solo unstable atom emits. Some of the time, that results in a collision with another atom, which may or may not emit, and so on. The reult is a mini-Poisson process within a Poisson process. In aggregate, there is no problem modeling this as a single Poisson process.

 

[Thanz]

Why is it that you cannot answer this simple question? You understand it perfectly, Thanz. But you cannot answer it.

No, I really really don't know what you are nattering on about, but I will attempt again to answer.

Does a comparison of my Poisson results and Hoyt's Poisson results tell us anything of value for the ultimate question - that is, whether JE is cold reading? No, it doesn't, for precisely the reason you state - Hoyt's data (his count) is flawed.

Does a comparison tell us anything about the effect of Hoyt's error on the analysis? Of course it does. How else could we assess the impact of the error if we don't compare it to the correct version?

So, a comparison of flawed data and correct data can give you information about the effect of the flaws. But I agree that it won't give you any valuable information about whatever the underlying study pertains to (JE cold reading, here) because of the flaws.

Does that answer your question?

 

[Thanz]

You are studying radioactive decay events. Most of the time, a solo unstable atom emits. Some of the time, that results in a collision with another atom, which may or may not emit, and so on. The reult is a mini-Poisson process within a Poisson process. In aggregate, there is no problem modeling this as a single Poisson process.

Can we apply this sort of reasoning to my coin example? I'll expand it so that I think it is parallel.

Let's say we have two guys doing coin flips. Adam is doing the flips, and then he calls out the result to Bob. Bob is a little hard of hearing, and his hearing aid battery isn't so hot either. On some occasions, which are random, he has to ask again what the call on the coin was. Sometimes he'll say "What?" sometimes he'll repeat the call, and sometimes he'll just write the call down. It happens with both heads and tails.

We are interested in coin flips. Does it matter if we write down all of the mentions of the coin flips, and use that, or does it make more sense to grab Bob's recording of the flips? How much does your answer depend on N?

 

[CFLarsen]

No, I really really don't know what you are nattering on about, but I will attempt again to answer.

Does a comparison of my Poisson results and Hoyt's Poisson results tell us anything of value for the ultimate question - that is, whether JE is cold reading? No, it doesn't, for precisely the reason you state - Hoyt's data (his count) is flawed.

Does a comparison tell us anything about the effect of Hoyt's error on the analysis? Of course it does. How else could we assess the impact of the error if we don't compare it to the correct version?

So, a comparison of flawed data and correct data can give you information about the effect of the flaws. But I agree that it won't give you any valuable information about whatever the underlying study pertains to (JE cold reading, here) because of the flaws.

Does that answer your question?

In a fashion, yes. It tells me that you made the comparison based on flawed data. It tells me that you are arguing both sides: Your comparison has value, and it does not have value. Your comparison tells us something, and it tells us nothing.

Pathetic. One simple question, all these evading posts from you, all these excuses. Pathetic.

One simple question is all it takes. One simple question, that goes to the heart of your argument, and you fold. After so many pathetic excuses ("I just wanted to be sure I knew what you are talking about" - ◊◊◊◊◊◊◊◊!), you fold.

Pathetic. You are bluster, and nothing else.

This thread is saved. You really are something.

 

[Thanz]

In a fashion, yes. It tells me that you made the comparison based on flawed data.

Please, tell me what specific comparison you are referring to. What post? What point do you think I was trying to make? I don't think that you actually understand what I am arguing here. If you are trying to say that Hoyt's flawed data makes one of my points invalid, at least have the decency to tell me which one.

It tells me that you are arguing both sides: Your comparison has value, and it does not have value. Your comparison tells us something, and it tells us nothing.

Not arguing both sides of anything. Just stating the obvious - a comparison done for purpose A may not be valuble if one is trying to answer a question about B. It is not two sides of one issue, it is two separate issues.

Pathetic. One simple question, all these evading posts from you, all these excuses. Pathetic.

One simple question is all it takes. One simple question, that goes to the heart of your argument, and you fold. After so many pathetic excuses ("I just wanted to be sure I knew what you are talking about" - ◊◊◊◊◊◊◊◊!), you fold.

Pathetic. You are bluster, and nothing else.

This thread is saved. You really are something.

What the heck are you talking about? I mean really, what the heck are you talking about? How have I "folded"? Folded on what issue? How does your question go to the heart of anything?

In what way have I relied on flawed data to say anything other than the data is flawed?

I'm glad the thread has been saved. Maybe you can read it one day and actually understand it.

 

[T'ai Chi]

Originally posted by CFLarsen (edited by T'ai Chi)

One simple question is all it takes. One simple question, that goes to the heart of your argument, and you fold. After so many pathetic excuses ("I just wanted to be sure I knew what you are talking about" - bullsh*t!), you fold.

Saved for posterity.

 

[T'ai Chi]

Until this analysis question is answered, can we assume there is a possibility that JE and others really do talk to the dead?

Of course if what they do is real they aren't talking to the actual dead, but rather the spirits of the people who were alive, or whatever.

 

[CFLarsen]

Please, tell me what specific comparison you are referring to. What post? What point do you think I was trying to make? I don't think that you actually understand what I am arguing here. If you are trying to say that Hoyt's flawed data makes one of my points invalid, at least have the decency to tell me which one.

You really are struggling here, aren't you? I made it perfectly clear that I have only talked about one comparison, and that is the one you made. The Poisson one.

Not arguing both sides of anything. Just stating the obvious - a comparison done for purpose A may not be valuble if one is trying to answer a question about B. It is not two sides of one issue, it is two separate issues.

Struggling. Not succeeding.

What the heck are you talking about? I mean really, what the heck are you talking about? How have I "folded"? Folded on what issue? How does your question go to the heart of anything?

In what way have I relied on flawed data to say anything other than the data is flawed?

I'm glad the thread has been saved. Maybe you can read it one day and actually understand it.

Thanz, please. You are really, really struggling here, and you think you can make it look like you are not?

You made a comparison, based on what you have argued was flawed data. For what purpose? You could not answer.

You really think that your hand-waving will do you any good? "Oh, I don't understand what you are saying, so how can I defend myself?"

Stop whining. You are pathetic. You have painted yourself into a corner, and you can whine all you like about it.

Now, please whine about how wrong I am, about how unfair I am. It does not change the facts.

You. Do. Not. Have. A. Case.

 

[BillHoyt]

Can we apply this sort of reasoning to my coin example? I'll expand it so that I think it is parallel.

Let's say we have two guys doing coin flips. Adam is doing the flips, and then he calls out the result to Bob. Bob is a little hard of hearing, and his hearing aid battery isn't so hot either. On some occasions, which are random, he has to ask again what the call on the coin was. Sometimes he'll say "What?" sometimes he'll repeat the call, and sometimes he'll just write the call down. It happens with both heads and tails.

We are interested in coin flips. Does it matter if we write down all of the mentions of the coin flips, and use that, or does it make more sense to grab Bob's recording of the flips? How much does your answer depend on N?

Your examples get wilder and wilder because you can't deal with the statistics involved and because you insist of the presumption that JE has a one-to-one mapping of people and names. The data are abundantly clear, though. In your example, you try to pack two questions into one. You asked "does it make more sense?" There is a choice here, sir, and no it does not make more sense. Can it make sense? Yes. It is no different than random error in any measurement, so long as it is random. And therein lies the statistical question that is giving a charlie horse from dancing so fast: does it reveal a random mini-Poisson, or something decidedly non-Random.

I have stated this before, only to be met with lame and non-statistical responses. The "J"s, if random, would have regressed to the mean. Tr'oll won't handle this issue either. We can only speculate why. But they didn't regresss to the mean. You note that both the numerator and denominator doubled, suggesting there was a non-random coefficient.

Let me whip out, then, the other thing I've brought up time and again, that you chose to ignore: Lurker's claim. He summarized his collection and analysis of a totally separate JE transcript this way:

"Hoyt:
9 J guesses
10 non-J guesses
Poisson says reject null hypothesis

Lurker
2 J guesses
5 non-J guesses"

Note, please the great disparity in growth of the Js versus the non-Js. Js grew by 4.5 -fold, while non-Js doubled.

Clearly, your overcount-pushes-it-into-significance claim has a fatal problem, doesn't it?

 

[BillHoyt]

Tr'oll, Thanz, Kaffee Klatch,

if your next posts do not squarely, directly, accurately and insightfully address all the issues I just raised, I am done with you woo flies. I will be busy spending my time sharpening mallets for the next round of whack-a-woo.

 

[T'ai Chi]

Your examples get wilder and wilder because you can't deal with the statistics involved and because you insist of the presumption that JE has a one-to-one mapping of people and names.

Perhaps you can't deal with the statistics invovled either...

The data are abundantly clear, though.

Fair enough. Do you think that your counting method treats, for example 'I see a grandpa. I'm getting a J name, John, Jacob, Jimmy, Joe, Joseph' as independent counts?

Yes or no??

Tr'oll won't handle this issue either. We can only speculate why.

You only anayze J counts and avoid other important letters. You treat these counts as independent when they aren't. You overcount.

I understand things quite well here.

Clearly, your overcount-pushes-it-into-significance claim has a fatal problem, doesn't it?

I'd like to see you and Thanz analyze, say, 5 JE (or other) transcripts, and

-give counts for each transcript

-as well as the p-value

-and conclusion

We'd see very quickly the differences between approaches.

 

[T'ai Chi]

You. Do. Not. Have. A. Case.

Yes. Thanz. Certainly. Does.

 

[T'ai Chi]

Tr'oll, Thanz, Kaffee Klatch,

if your next posts do not squarely, directly, accurately and insightfully address all the issues I just raised, I am done with you woo flies. I will be busy spending my time sharpening mallets for the next round of whack-a-woo.

Just be sure to leave the ball, Jr.

For the most part, I think the criticism of your analysis method has been very insightful. The statistical issues involved here are not so clear cut. They are actually pretty tricky.

Label your critics all you want; that still won't help your analysis any, nor address any actual issues.

Will you address the issues of independence?

Will you analyze, say 5 transcripts and have Thanz do the same and present your results side by side with his?