The "Process" of John Edward

Just to clarify for those who could miss it, by P(17), Lurker means
SUM(P(x)|0<=x<=17), not just evaluating the Poisson probability density function at 17.

I know that the Poisson is used often when studying counts, but just because we are studying counts, does that necessarily mean the Poisson is appropriate to use? I dunno. We are assuming that the distribution of J counts is distributed approximately Poisson. Could someone discuss the evidence for that?

If instead the mean was 11.65 instead of 11.356 we'd end up failing to reject the null hypothesis. That seems fairly close. Not sure what that means though.
 
T'ai Chi said:
I know that the Poisson is used often when studying counts, but just because we are studying counts, does that necessarily mean the Poisson is appropriate to use? I dunno. We are assuming that the distribution of J counts is distributed approximately Poisson. Could someone discuss the evidence for that?
Click to expand...
I might be able to help more if I understood your question. Poisson is used for events with low probability. Classic uses are events over a geographic region, events over a span of time. Things like cars arriving at an intersection in a five minute interval. Horse deaths per acre. Defects per square foot of sheet metal. In this case, names beginning with "A" or "D" or "J".

An alternative choice might be binomial, but therein lies a catch. The advantage of Poisson is that the moment-generating functions are such that its mean and its variance are the same and therefore, we only need to know the mean, which is about all we have here.

If instead the mean was 11.65 instead of 11.356 we'd end up failing to reject the null hypothesis. That seems fairly close. Not sure what that means though.
Click to expand...
Close only counts in horseshoes. It means nothing. The rejection criterion is always binary: you accept or you reject. If you the significance is .049 or .019 or .0499, it still meets the criterion.

Cheers,
 
BillHoyt -

I see that you have made several replies, but not to my latest post. Do you agree with what I have said?

Are not even going to bother trying to defend your hypothesis?
 
Thanz said:
BillHoyt -

I see that you have made several replies, but not to my latest post. Do you agree with what I have said?

Are not even going to bother trying to defend your hypothesis?
Click to expand...
Thanz,

My hypothesis needs no defense. You have confounded it with the test of the hypothesis. These are separate concepts scientifically. Please refer to my post about the hypothesis and the null hypothesis. They specify no test. They need not specify a test.
 
Mr. Hoyt -

I have read your post on the hypothesis and the null hypothesis. You are completely missing the point.

Let's start at the beginning. Let's say that you are a researcher, and you want to see if JE was cold reading. You come up with the idea that if he was cold reading, he would probably guess common initials a lot, and less common initials very rarely.

Now you have to figure out the best way to test this theory. You decide that you need to compare it against the normal population instead of just using raw numbers of guesses.

Without regard for whatever data is currently available, what test do you want to do? Do you want to compare one letter? Or do you want to compare ALL letters?
 
Thanz said:
Mr. Hoyt -

I have read your post on the hypothesis and the null hypothesis. You are completely missing the point.

Let's start at the beginning. Let's say that you are a researcher, and you want to see if JE was cold reading. You come up with the idea that if he was cold reading, he would probably guess common initials a lot, and less common initials very rarely.

Now you have to figure out the best way to test this theory. You decide that you need to compare it against the normal population instead of just using raw numbers of guesses.

Without regard for whatever data is currently available, what test do you want to do? Do you want to compare one letter? Or do you want to compare ALL letters?
Click to expand...

Thanz,

You keep trying to shift the focus away from your assertion about not being able to work with the available data. You keep trying to focus the attention on this question of "wouldn't more data be better?"

I have said, repeatedly, yes. You continue to question me on this. It is irksome.

Now you try an angle of attack based on my hypothesis being wrong. It is not. You are really questioning the test of the hypothesis. Do you understand the distinction? If you have read the hypothesis and null hypothesis, why do you still confuse them with the test?

If I am a researcher in this area, and am confronted with the medium's feint, I am left with no alternative but to uncover whatever untainted evidence is available and see if it is workable. JE plays the medium's feint all the time. His show is heavily edited, he won't subject himself to testing (except by very friendly researchers) and leaves little evidence to scrutinize. With little data available, I look at the hypothesis and the available data and ask myself how can I test the hypothesis?

In the control population, "J" sings out as the, by far, most frequently seen initial. I chose that to test. I chose .05 significance. That poisons the well for further dipping afterwards as has been discussed before. JE's data were remarkable. In a small set (n=85), we would have expected 11 hits on "J". Yet he had 18. That is significant at the .05 level. I reject the null hypothesis.
 
Not to divert this, but I'm just curious about something. If "J" showed up far beyond chance in JE readings...and if JE was 100% correct every time he said it, wouldn't it give more credence to the idea that he's not cold reading?

Apart from the argument about the too-small sample size (which now even Bill agrees with, apparently, finally), why would excessive use of "J" support JE as a cold reader if, hypothetically, all the "guesses" were correct? :confused:
 
Clancie said:
Apart from the argument about the too-small sample size (which now even Bill agrees with, apparently, finally), why would excessive use of "J" support JE as a cold reader if, hypothetically, all the "guesses" were correct? :confused:
Click to expand...
Clancie,

Where did I say that? Show us the post where you think I said that.
 
BillHoyt said:
You keep trying to shift the focus away from your assertion about not being able to work with the available data. You keep trying to focus the attention on this question of "wouldn't more data be better?"
Click to expand...
No, I am not shifting the focus to "wouldn't more data be better". I am keeping the focus on "is this data adequate".

I do not feel that a test of one letter alone, whatever the results, is adequate to shed light on whether JE is cold reading. An analysis of all letters is adequate. And we don't have enough data right now to do it. Because the sample size is too small.

Now you try an angle of attack based on my hypothesis being wrong. It is not. You are really questioning the test of the hypothesis. Do you understand the distinction? If you have read the hypothesis and null hypothesis, why do you still confuse them with the test?
Click to expand...
Not that the hypothesis is wrong per se. But that the hypothesis is not adequate to tell us anything meaningful about cold reading. It is fine as far as it goes - I just don't think that it goes far enough to tell us anything meaningful. There is just not enough information there. It leaves too much unquestioned.
 
Thanz said:

No, I am not shifting the focus to "wouldn't more data be better". I am keeping the focus on "is this data adequate".

I do not feel that a test of one letter alone, whatever the results, is adequate to shed light on whether JE is cold reading. An analysis of all letters is adequate. And we don't have enough data right now to do it. Because the sample size is too small.
Click to expand...
This is patently false. Let us say that, in 85 guesses, he guessed "J" 85 times. This would be extraordinary. There would be no other letters whose frequencies we test. Does that mean we don't have enough data? No way.

You cannot make the assertion you made. The adequacy of n depends on the hypothesis, the distribution model, the desired significance and the observations.

I am done discussing this with you.



Not that the hypothesis is wrong per se. But that the hypothesis is not adequate to tell us anything meaningful about cold reading. It is fine as far as it goes - I just don't think that it goes far enough to tell us anything meaningful. There is just not enough information there. It leaves too much unquestioned. [/QUOTE]
Once again, you have confused the hypothesis with the test of the hypothesis. I am not repeating myself here either. I am done discussing this with you.
 
BillHoyt said:

I am done discussing this with you.
Click to expand...
Fine with me. Keep your leg in the puddle and keep insisting that you are swimming.
 
Posted by Bill Hoyt
You keep trying to focus the attention on this question of "wouldn't more data be better?" I have said, repeatedly, yes.
Click to expand...

Okay, re-reading your post, I guess you only saying "more data would be better", but the very small sample size is still enough for what you were doing, in your opinion.

But I'm still interested if you -really- think your results stand up, Bill? Do you really feel that this "hypothesis about 'J'" has added something statistically of value to the discussion of cold reading and JE?
 
Clancie said:
But I'm still interested if you -really- think your results stand up, Bill?[/b]
Click to expand...
"Stand up" how, Clancie? Do I really think that further data will support the cold reading hypothesis? Yes. Absolutely. The problem will be to find such data. CO cannot be used because it increases the likelihood of warm- or hot-reading.
Do you really feel that this "hypothesis about 'J'" has added something statistically of value to the discussion of cold reading and JE?
Click to expand...
Yes. It adds a rejection (at the .05 level) of the hypothesis that JE's name and initial guesses match the names in the population. He overused the most frequently found forename initial, "J" significantly. If JE is cold-reading we would expect him to concentrate on those initials that are most abundant in the population. We see exactly that in this instance.

Cheers,
 
You must log in or register to reply here.

Back
Top Bottom