Cold Reading Demos at TAM2

In my experience a scientist cannot bias the results for one group by picking a certain test. An inappropriate test may be chosen for the measurement scale used, or something similar, but it seems unlikely to bias the analysis..
If I have a 2 X 2 design with binary data, it won't change the outcome if I use Fisher's Exact test or Chi squared.
The point at which statiticians can screw up is in data entry. So it is better, in all probability, to have someone who is unbiased to do the analysis.
 
thaiboxerken said:


Garrette keeps trying, it seems that Schwartz isn't accomodating.

The truth is, comic-book guy, Garrette didn't even send an email to Dr Schwartz at all! Didn't you know that? You seem to live in a strange world of make-believe.
 
Hi Jeff,

I agree with almost all of what you said. I do think that, although unlikely, it is possible at least that one choice could make the p-value small enough to be statistically significant, and the other method wouldn't. The p-values would be close, but one method would make it technically statistically significant.

For example, consider the table of hypothectical data:

58, 25
25, 25

and the hypothetical alpha level is .02.

Using your example of Fisher's Exact and chi-square test:

From Fisher's Exact, we get a p-value of .0178. From the chi-square test we get a p-value of .0218. With a Yates correction on that chi-square test we get an even higher p-value of .035. Only in the Fisher's Exact case do we technically have statistical significance however.

(in 2x2 tables however I agree that this issue is not too big of a deal, but in other analysis methods it could be)

In any case, the p-values from the two methods would be different since some methods are approximations or modifications/adjustments of others. If the evaluator was biased against something, it is possible he'd consciously or unconsciously report the method that gave the larger or smaller p-value (or larger or smaller estimate in general) to reflect his bias.
 
CFLarsen said:

How? You just state that they could. Please explain how.


It depends on the exact experiment, of course. Bias creeps in the same way it does in any experiment.


Sorry, but your own posts show this to be true.

No matter how many times you say it, it still won't be true.
 
thaiboxerken said:
Schwartz refuses to disclose all the data of his experiments before, why should anyone believe that Schwartz would allow all the data to be released in his "challenge"?


To play devil's advocate here a bit, when was the last time JREF disclosed all of its data? Peer review? On a website? In press releases? In a book? Where? In an accessible database?
 
BillHoyt said:

Has nothing to do with the hypothesis test, sir. This is a presentational decision.


Has everything to do with it. It is how bias can creep in. Read How to Lie With Statistics for interesting discussions.


That is why Fisher says the statistician should have been there from the beginning.


Ideally, yes. In reality, sometimes not. I've already said this though.


You clearly have no idea what you are talking about with regard to statistics, you poser!


Wah. You'll have to do better to be taken seriously.


Power, sir, depends on the type of test and n.


And on several others things you failed to mention: confidence, variation, the hypothesis, and the values under the hypothesis.


Your sig line, clearly, was disingenuously chosen to help your pitiful pose. You had no freakin' idea what Fisher was talking about, tr'oll


I like Fisher and statistics, that is why I chose it. I have no freakin' idea about experimental design, sure. I've yet somehow managed to BS (and MS) my way through three experimental design classes and consulting people in experimental design! Funny how that works.


Not with regard to the hypothesis being tested. That is decided a priori.


You don't interpret results of an experiment a priori because the experiment hasn't been done yet.


T'ai, you have just clearly demonstrated that a) you are no statistician, b) you have been posing for months here, c) you have never studied experimental design, d) you have no understanding of p-value, e) no understanding of "power" in statistics and f) no understanding of blinding. This is pathetic.


Sorry to dissappoint you, but your lame attempts at personal attacks don't intimidate, impress, or phase me (or anyone) in the slightest. I've dealt with many bullies in the past in real life, and so far, you've not varied from their methods. Phhftt.

You'll have to talk about the issues (not me) actually.
 
Lucianarchy said:


The truth is, comic-book guy, Garrette didn't even send an email to Dr Schwartz at all! Didn't you know that? You seem to live in a strange world of make-believe.

Did Garrette not say that he did? Now are you saying Garrette is just fabricating?
 
CFLarsen said:

No, you have not. You have merely repeated your claim.


See above posts.


But you haven't explained how the statistician unconsciously chooses the method that will give better results for one group.


It varies according to the specifics.


Your answer is pivotal to your claim that you can explain what "tripleblind" is and how it works. So far, you have not been able to answer that.


Wrong.
 
T'ai Chi said:
Riiiiggghhht. I like Fisher, that is why I chose it. I have no freakin' idea about experimental design, sure. I've somehow managed to BS (and MS) my way through three experimental design classes and consulting people in experimental design! Funny how that works.
Great to hear, tr'oll. Please cite the papers that credit you.
 
Following is the e-mail I sent on 25 January of this year, copied from my Sent folder, which explains why there is no header for the second sending. No response yet, but I would not necessarily expect one so soon.

A complication is that in about two hours I am leaving on an extended trip that will probably prevent me accessing my e-mail for up to three weeks. Any response before then may go unnoticed.

------------------
------------------

Hello, Professor Schwartz:

I am resending my e-mail of 2 December 2003 in the hopes you will receive it and respond. I understand the holidays are very busy for everyone.

I am also sending it to two similar but different e-mail addresses as, after checking to be sure I had the correct address, I am not certain I sent it to the correct address the first time.

Please read the enclosed e-mail. I must, however, make one amendment: my availability has been changed to at least late summer.

Thanks for your time.

Patrick Vessels


--------------------------------------------------------------------------------

----- Original Message -----
From <[email protected]>
Date Wed, 03 Dec 2003 19:13:04 +0300
To [email protected]
Subject Replication of John Edward Success in Afterlife Experiments
Dear Professor Schwartz:

My name is Patrick Vessels. I am a member of the James Randi Educational Foundation internet forum. My username on that forum is Garrette.

Steve Grenard, whom I believe you know, is also a member of that forum and gave me your contact information. He indicated that he has already communicated with you regarding me, but it is possible I am mistaken.

I am contacting you concerning your Afterlife Experiments, particularly those funded by HBO and conducted at HESL. More specifically, I am contacting you because Steve suggested that you would be interested in my claim that I can replicate the 'success' of John Edward in that experiment.

If you are interested, then I wish to discuss the possibility of recreating that experiment. My contact information is at the end of this message (or in the address portion of this e-mail). Currently, I am a soldier deployed to Iraq; there are times when I have relatively ample access to the internet, but there are also times when I have virtually none at all. My redeployment date is not specified but should be in mid-to-late spring of 2004.

I expect no remuneration for my participation in any agreed upon experiment. Steve has graciously offered to cover my expenses, at least in regard to transportation costs. I can personally bear some lodging costs.

In considering your response, it is important that you understand some things about my claim and about me.

First, about my claim:

1. I claim I can replicate what you call John Edward's success during the HESL experiment; that means any experiment with me will need to have the same design features--the same strengths and the same flaws. It is in duplicating the flaws where I foresee difficulty in setting up this experiment.

2. I do not claim to be able to replicate John Edward's success in this experiment while I am placed under more stringent controls; i.e., I am not offering to participate in a new experiment with new designs, new protocols, and new aims.

Second, about me:

1. I am an amateur magician, but strictly amateur. I have never been paid for a performance and have, in fact, conducted very few performances at all--none for many years.

2. I am, however, an avid researcher of magic.

3. My personal opinion of the quality of your Afterlife Experiments is very low. I have written on the internet about what I consider to be extensive and fatal flaws in their design and in your analysis.

4. I respect your academic achievements and degrees. However, because the degrees are so impressive and the flaws in the experiments are (to me) so obvious, I find the only reasonable conclusion is that you consciously overlook the flaws. I cannot believe that a man so highly educated and intelligent (as you must be to have received your Harvard degree) could honestly miss flaws that are so basic and glaring.

5. Because of #4 above, I have, on occasion, used the term 'fraud' in describing you. I have stated that the flaws in the experiments are fact, but that it is my opinion that your presentation of the experiments as indicative of mediumship is fraudulent.


I do not say this last to antagonize you, though I am sure it must to some degree. Rather, I say it in the interest of disclosure so that any relationship we form can be open and professional. I do not wish to falsely present my stance.

I am sorry that this message is so long, but I do not know how I could have shortened it and still covered the essentials. If you have an interest in pursuing this further, there will be much more communication of much greater length and detail. Protocols will need to be agreed upon and obstacles overcome, some of which, I admit, I do not yet know how to overcome.

In any case, please feel free to e-mail me at any time. My e-mail is [email protected], and I will respond as quickly as circumstances allow.

Best wishes,

Patrick Vessels
(Garrette on the JREF Forum)
 
BillHoyt said:

Great to hear, tr'oll. Please cite the papers that credit you.

You'll have to actually address the issues of experimental design, and not the issue of me to be taken anywhere near seriously.
 
T'ai Chi said:
It depends on the exact experiment, of course. Bias creeps in the same way it does in any experiment.

But the statistician comes in after the experiment is done.

T'ai Chi said:
No matter how many times you say it, it still won't be true.

I don't need to say it many times, it is true. Your own posts show this.

T'ai Chi said:
It varies according to the specifics.

Then pick those specifics that show that a statistician can unconsciously choose the method that will give better results for one group.

T'ai Chi said:

Nope. You have not been able to explain how "tripleblind" works. You pointed to a supernatural ability to predict the method that will give biased results, that's all.

In other words, woowoo.
 
thaiboxerken said:


Did Garrette not say that he did? Now are you saying Garrette is just fabricating?

Garrette said he'd sent one mnths ago, and sat back opining about dr Schwartz's lack of response, without checking his own error in addressing the email, which he now claims he spelt wrong. He only sent a real email this week,
 
Gary Schwartz has my details too and, according to a message Steve Grenard forwarded on to me, is quite glad that he has some cold readers to act as controls.
 
T'ai Chi said:
You'll have to actually address the issues of experimental design, and not the issue of me to be taken anywhere near seriously.
Well, that's the funny thing, there, Tr'oll. I was addressing the issues when suddenly you ejaculated this outburst:
Originally posted by T'ai Chi
I've somehow managed to BS (and MS) my way through three experimental design classes and consulting people in experimental design! Funny how that works.
Do you now wish to withdraw the claim of "consulting people in experimental design"? If not, then please cite the papers where your consultancy was acknowledged by the principal investigators.
 
Clancie, when Penn compared John Edwards to Heroin, he wasn't off. JE has caused horrible anguish and misery. His "belief" that everyone goes to heaven, and is just happily waiting for YOU...

it was a disaster waiting to happen. And there are countless small town JE that are preying on the sad, lonely and depressed. My sister in law, the psychologist can tell such sad stories of people taken for money they don't have, and left wanting more and more. Instead of grieving, they cling, and lose their savings and sometimes even their very life. If you could have seen the video shown of James Van Pragh, and the family that had lost their little girl. You wanted to vomit. OK, do it for "fun" do it for whatever, but don't take advantage of people who have suffered the loss of a child. I always contend that JE is an atheist, because if he believed he would ever be accountable for what he has done in the afterlife, he'd quit and join a monastary NOW.
 
kittynh said:
Clancie, when Penn compared John Edwards to Heroin, he wasn't off. JE has caused horrible anguish and misery. His "belief" that everyone goes to heaven, and is just happily waiting for YOU...

No, no, no, no, no, no.....you got it all wrong.

Everyone goes to Heaven, and is just happily waiting for you to pay JE (or another medium), so the dead can say things like "I love you", "I am OK", and show JE images of letters.

Even though JE sells psychic "do-it-yourself" kits, I have yet to see one single believer say that she did not need her JE-fix any longer after buying one (because now, she could do it herself).

Comparing JE to heroin was - if you pardon the pun - dead-on.

kittynh said:
it was a disaster waiting to happen. And there are countless small town JE that are preying on the sad, lonely and depressed. My sister in law, the psychologist can tell such sad stories of people taken for money they don't have, and left wanting more and more. Instead of grieving, they cling, and lose their savings and sometimes even their very life. If you could have seen the video shown of James Van Pragh, and the family that had lost their little girl. You wanted to vomit. OK, do it for "fun" do it for whatever, but don't take advantage of people who have suffered the loss of a child. I always contend that JE is an atheist, because if he believed he would ever be accountable for what he has done in the afterlife, he'd quit and join a monastary NOW.

He'll have to work pretty hard not to get thrown in hell as it is...
 
CFLarsen said:
Comparing JE to heroin was - if you pardon the pun - dead-on.

*Groan*

Just stick the puns in as you see fit. Smack 'em over the white horse's head. Yep, jes keep needlin' them.
 
Hi kittynh,

Well, I'm going to disagree with much of what you posted ("surprise!"), but I'm not sure if it will be today.

In the meantime, no, I didn't see the clip of JVP that you're talking about. I -did- see another heartbreaking one, though, about a missing boy near a railroad car, etc. He got nothing, should have been working with the parents--at least not showing it on tv and telling them definitely that the boy was dead (not without more to go on) and it seemed quite distressing. So, I think that's probably similar.

My point, quickly, is that JVP will be on Larry King Live tonight. Maybe someone will get through (as with SB) to ask him about the ethics, etc. involved in his work with the parents of missing children. He's taking calls.
 
T'ai Chi said:


For example, consider the table of hypothectical data:

58, 25
25, 25

and the hypothetical alpha level is .02.

Using your example of Fisher's Exact and chi-square test:

From Fisher's Exact, we get a p-value of .0178. From the chi-square test we get a p-value of .0218. With a Yates correction on that chi-square test we get an even higher p-value of .035. Only in the Fisher's Exact case do we technically have statistical significance however.
One problem is that you have cited the one-tailed probability for Fisher's and the two-tailed probabilities for both chi-square tests. The two-tailed p for Fisher's is .027.
Back to Stat 101 for you.
 

ISF - Join now!

Every member here is approved by hand. No bots, no spam, just people who care about evidence and honest debate.

Membership is free!

Create your free account

Back
Top Bottom