The "Process" of John Edward

BillHoyt said:

We are not trying to approximate binomial with Poisson here. We are simply using Poisson. Period. Yes, radioactive decay is a physical system that has a Poisson distribution.
Click to expand...

And radioactive decay has a near infinite N (atoms) and a small p. I wonder why Poisson is applicable?

Bill, did you find out how Poisson was derived yet? Have you addressed dodge's binomial comparison yet?

Lurker
 
Hey! New page - 21!

Since Mr. Hoyt still hasn't answered my question regarding his counting methods, I'll repost it on the new page so he doesn't have to go looking for it.

I would like to know why you think it is appropriate to count "And they're also talking about somebody who would be known as either Richard or Rich, because a big R-connection that comes up connected to you" as more than one guess. In my count, I counted this as one R guess. From your description, I gather you counted it as 3 R guesses.
Click to expand...
 
I once saw a list of all the names thrown out by a medium and tabulated by Michael Shermer. There were something like 40 names and he said these all happened during the course of a seminar several hours long. This was a misleading statement since half of them were the same name and about a 1/4 were phonetically alike. This exercise proves nothing in such cases.

Richard = Ricky = Rick = Dick = Dicky = Rich = Ritchie

Isnt it possible someone with the name Richard has been called by different people at different times all of those nicks?
 
I'll repeat my previous post, and add some clarification:
I re-worked the transcripts and came up with different results and a different method. Here it is, in a nutshell:

1. I used the census data figures orginally presented, although these may need tweaking.
2. I excluded the CO show data, and concentrated solely on the available, unedited transcripts from LKL, etc.
3. I looked at JE's style and adjusted the counting procedure as follows:
o I counted all of his name guesses
o Whether he stated them as names or initials, I counted them
o I excluded impossible-to-deal-with things such as "a B softened by a vowel," and chalked that up to a "B" guess.
o I included even bizarre names such as "pepper", "salt", "brooklyn" and other nickname guesses, except that
o I only counted "Liz", "Elizabeth" type guesses as the full given name, and did not also count an "L". but
o When JE recited a littany of names, I counted each one, whether they had the same initial or differing initials (again excluding the "Liz/Elizabeth, Ronny/Ronald, and Bill/William" type guesses, where I only counted the intial of the full given name.

Sound a bit complicated? You should read the transcripts. I could not see another way to approach things fairly given that sometimes he was all over the board. My hypothesis was, that, if there is a JE mediumship process, i should honor as much of it as I could figure in making the counting rules.
Click to expand...
Regarding the multiples, I realize this post didn't full explain my reasoning. Sometimes, JE calls out a name or an initial for a person who passed, sometimes for somebody else still living, sometimes he uses the same initial for possibly two different people. Sometimes the names he tosses out are impossible to categorize. "Helen , Ellen or L-name" is a prime example. What the heck do you do there? The first one was H, the second E, the third goes to L. Does he mean L preceded by something? Or L followed? He does the same thing with the "B softened by a vowel" thing. Sometimes he mentions simply a vowel in front. Other times he precedes that vowel with a consonant.

There are three choices as I see them: drop such guesses entirely and thereby lose most of the data, pick one letter and run with it (but which one?) or make a uniform rule allowing him as much leeway as you can. I said I tried to honor his process as much as possible, and I decided to grant him the leeway. The specific leeway I did not grant was "Rich or Richard", where one is clearly a nickname for the other. In those cases, I counted one guess. If he said "Rich or Richard or R" I counted two guesses.

With 26 available letters, and with a focus on just "J", the major distortion I expected was a blossoming denominator. (Clearly, this did happen.) The overall effect I expected was the "J"s would either barely hold onto their significance or that they would be washed out by all the multiple Ds and Rs and Ms and Bs, etc.
 
Bill,

Have you had a chance to look into the derivation of the Poisson yet?

For fun, try calculating the odds for getting a heads by flipping a coin three times. List the odds for each total heads as 0 times, 1 time, 2 times and 3 times.

Here is what I got via a probability tree which has zero error:

0 0.125
1 0.375
2 0.375
3 0.125

Making it a cumulative we get

0 0.125
1 0.5
2 0.875
3 1.0

Now I plug this problem into Poisson using mu=np=3*0.5=1.5

0 0.22313016
1 0.5578254
2 0.808846831
3 0.934357546


But wait, Bill, the numbers are different. Why is that? I have provided the EXACT solution and good old Poisson is not very close. Why is that?

Lurker
 
Lurker said:
But wait, Bill, the numbers are different. Why is that? I have provided the EXACT solution and good old Poisson is not very close. Why is that?

Lurker
Click to expand...

monkey.gif

Surprising that you misapply Poisson and see an error?
 
BillHoyt said:
Regarding the multiples, I realize this post didn't full explain my reasoning. Sometimes, JE calls out a name or an initial for a person who passed, sometimes for somebody else still living, sometimes he uses the same initial for possibly two different people. Sometimes the names he tosses out are impossible to categorize. "Helen , Ellen or L-name" is a prime example. What the heck do you do there?
Click to expand...
I focussed on how wide a net he was casting. We have already debated whether it is appropriate to consider a guess of simply "John" as equivalent to "J". For the purposes of my count, I considered them equal, although I think a strong argument exists that they are not.

For this specific example, I considered it to be 3 guesses - as he would accept any similar H, E, or L name as a "hit". In this instance, he is using 3 letters to guess one person, so I felt it appropriate to count all three.

The specific leeway I did not grant was "Rich or Richard", where one is clearly a nickname for the other. In those cases, I counted one guess. If he said "Rich or Richard or R" I counted two guesses.
Click to expand...
This is where we differ. I consider that to be one guess - as only one intitial letter is involved. He will accept any R name, and he is hoping to be close with Richard. But he is not making two guesses - the net stays only as wide as "R". As we are considering only the letters involved, I considered this to be one guess of the letter R. He is NOT casting 2 R nets here, which is what your count would seem to say.

With 26 available letters, and with a focus on just "J", the major distortion I expected was a blossoming denominator. (Clearly, this did happen.) The overall effect I expected was the "J"s would either barely hold onto their significance or that they would be washed out by all the multiple Ds and Rs and Ms and Bs, etc.
Click to expand...
However, that clearly did not happen. Adding in extra "J" guesses has more of an effect than adding in extra letters to the denominator. We have radically different counts. I don't know exactly what transcripts you used, but I used the LKL transcripts posted by Renata and got 9 J guesses out of a total of 43, compared to 18 and 85 for you.

When Kerberos did his count, he counted 78 guesses and 14 Js. The difference between his and yours are 4 Js and 7 total guesses. But the 7 extra total only move the expected by about .6, but the extra 4 Js makes all the difference in whether we can reject the null hypothesis.
 
BillHoyt said:


Surprising that you misapply Poisson and see an error?
Click to expand...

Hey, now we are getting somewhere! Finally! Now, how about you tell us when Poisson will be accurate and when it will not.

Thanks!

Lurker
 
SteveGrenard said:
I once saw a list of all the names thrown out by a medium and tabulated by Michael Shermer. There were something like 40 names and he said these all happened during the course of a seminar several hours long. This was a misleading statement since half of them were the same name and about a 1/4 were phonetically alike. This exercise proves nothing in such cases.

Richard = Ricky = Rick = Dick = Dicky = Rich = Ritchie

Isnt it possible someone with the name Richard has been called by different people at different times all of those nicks?
Click to expand...

Can we see this list, Steve?
 
Bill:

I would have thought a math genius like yourself would have been happy as a clam to dive into the derivation of the Poisson Distribution. I wonder why you haven't?:wink:

Could it be because of its relation to the binomial...:D

Lurker
 
Lurker said:
Bill:

I would have thought a math genius like yourself would have been happy as a clam to dive into the derivation of the Poisson Distribution. I wonder why you haven't?:wink:

Could it be because of its relation to the binomial...:D

Lurker
Click to expand...

Lurker,

You are an absolutely waste of time on this issue. Poisson can be derived from binomial. It can also be derived from geometric. It can also be derived by solving the differential equation:

f(t) = q*(1 - f(t))

The last derivation is from basic radioactive decay.

So what if they can be derived from one another? Don't you understand the Central Limit Theorem?

Now go away.
 
I guess we should all use the same readings, and we all should count the J-name occurances, and see what we all get.

Oh, in my opinion Bill probably thinks that JE just does cold readings with the letter J or something.
 
BillHoyt said:


Lurker,

You are an absolutely waste of time on this issue. Poisson can be derived from binomial. It can also be derived from geometric. It can also be derived by solving the differential equation:

f(t) = q*(1 - f(t))

The last derivation is from basic radioactive decay.

So what if they can be derived from one another? Don't you understand the Central Limit Theorem?

Now go away.
Click to expand...

My, touchy, aren't you. :) And what does one assume when one derives it from those equations? Hmm? You always seem to avoid that question. Why is that, Bill?

This is getting humorous as Bill frantically tries to avoid giving me credit for any shred of knowledge. It is so much easier for him to debate with ad homs.

Lurker
 
Come on, Bill. Enlighten us as to when it is applicable to use Poisson? Others have asked. Seems pretty simple. Why don't you answer?

Is it valid for all values of p?

Is it valid for any population size?

Simple questions for someone as knowledgable as you clearly are. And I do mean that. You clearly understand more about statistics than I do. I fully grant you that. So tell me when Poisson is applicable and when it is not. Two simple questions above.

Lurker
 
Lurker,

Do you think it would be prudent for people to have a chance to be able to get back to you, before you chastise them for not answering?

Not everyone is on JREF 24/7, you know.
 
CFLarsen said:


I'll await the statistical experts.
Click to expand...
Why the cop out, Mr. Larsen? You were all over me with stats questions earlier in the thread.

You don't have to be a stats expert to answer some of the basic questions here.

1. Do you think that an analysis of only one letter is meaningful to ascertain whether or not JE is cold reading? If someone were to ask you what you thought was a good test for cold reading comparison, would it be test of one letter or multiple letters?

2. What do you see as the proper counting technique for the following JE (style not a quote) guess: "I am getting a John or Joe or some other J, J-o connection here". How many J guesses would you count out of that?
 
You must log in or register to reply here.

Back
Top Bottom