Cold Reading Demos at TAM2

Garrette,

I don't know if you've read through the thread or not, but its not a discussion of Schwartz's research (and frankly, I've gone through the flaws of that, too, so many times...so I'm glad it's not :) ).

The part of the discussion that I was responding to was "what does triple blinding mean"? I think (but am not positive) that Schwartz has described a plan for triple blinding and I gave that as an example.

That's the point--whether an experiment can be triple blind or not. This was one (proposed) example of it.
 
Clancie,

Yes, I've read the thread, and I, too, do not want to rehash the Schwartz experiments.

I was responding to this comment of yours:

I may be wrong about this, but I thought the way Schwartz used "triple blind" was that the sitters didn't know who the mediums were, the mediums didn't know who the sitters were, and the test administrator didn't know which medium and which sitter were involved in which readings.

Which I took to mean that you were putting forth the Schwartz experiments as an example of triple blinding.

I take it that I was incorrect and you were instead asserting that Schwartz has proposed triple blinding in future experiments?

If so, my bad for misunderstanding.
 
TLN said:
Frankly, I think that was going a little too far (see, fearless),

Perhaps. However, these are the same methods, and they should be pointed out whenever they are used.

TLN said:
but you're still stereotyping without referring to anyone directly. You're both bigoted and cowardly. Sorry.

Never be sorry for speaking the truth.
 
CFLarsen said:

Now, what if the numbers were temperatures? The statistician does not know what the numbers mean.


The statistician would know they were temperatures (just like he knows they are blood pressures), he just wouldn't have access to the group information.


What does he compare them to?


The other groups, other numbers, an "alpha" level, like he normally would in any statistical analysis.


Stop. There are problems.


No, there aren't any. We are just doing 'normal statistics', but are recoding the group information.


Where, in your example, do you present the average blood pressure data?


The point, as already noted, was to show how the statistician would be blinded to group membership information. I really can keep repeating this all day if need be.


Nice evasion.


Nope.


You give an example, and when it is explained that it is thoroughly inadequate, you deem it "a quick example".


Too bad you weren't convincing on why it is supposedly thoroughly inadequate as you believe.


No? When I ask for examples, I am not asking for sloppiness.


Yawn.

As noted already, my example was to show how one could analyze data without the group information. It's not really difficult.


I was not referring to the specific numbers, but to the fact that:

1) You did not know how blood pressure was measured.
2) You had no idea in what range a blood pressure would be.

All in all, you had no idea what you were doing.


Then you are incredibly nieve to think 1) and 2). It is very amusing as both of my theses directly involved blood pressure. :) One was predicting heart disease, and the other was determining the reliability of blood pressure measuring devices. Do let me know if you want autographed copies of either, and I'd be happy to oblige.

But continuing on.. I mentioned that the numbers were entirely made up. I already mentioned that the example was to show how one could analyze data without the group information.


..., you have come up with an entirely inadequate and flawed example that does not explain what triple-blind is.


I showed how one could recode group information to make the real group labels unaware to the evaluator. That is the "guts" of a triple blind. The rest proceeds as any standard statistical analysis would, and therefore doesn't need to be shown, as it is entirely standard after that point.


Of course, when cornered, just claim that people are stupid.


That's what you do to apparently. Observe:

"All in all, you had no idea what you were doing."
"Could it be that you simply don't know?"


How does triple-blinding work?

Already answered. The group information is recoded so the evaluator doesn't know real group membership of individuals.

Oh, and in anticipating your next post:

Already answered. The group information is recoded so the evaluator doesn't know real group membership of individuals.
 
TLN said:
Which you do constantly Clancie. Name names or stop being such a bigoted coward.
Its usually rude to name names, its offensive to name names, and I would consider it unnecessary.

Rather than nitpicking, judge Clancie's comments using fair and objective inferences, dont go off trying to pick a fight.
 
TLN said:
Stereotype implies that most skeptics are like this, which is nonsense.

Name one poster here who fits this mold and tell me why you think that's the case.

Show a little backbone or keep your bigotry to yourself.
You've misread Clancie's comment.

She said:
It's true that most here -aren't- like the stereotype that he mocks.
You are under the impression that Clancie stereotypes skeptics as the extreme form of a "skepticism" !Xx+-Rational-+xX! parodies, however Clancie explicitly says the stereotype is not representitive of most skeptics (hence her usage of "it true most here arent like the stereotype").

Clancie believes the stereotype ludicrous, but its humorous. I would agree with that sentiment.

Just a minor booboo on your part.
 
Nyarlathotep said:
Do you always congratulate yourself when you are pleased with your handiwork, Ian? Either in print or aloud, that is.
Personally, I doubt !Xx+-Rational-+xX! is Ian's sockpuppet.

If anything, Ian should adopt !Xx+-Rational-+xX! as his very own troll (in the same way Suez and Mr. Skinny have adopted me as their very own son, Ian could have a son/troll of his very own!... [clarification]Note: I'm not a calling myself a troll. [/clarification]).
 
T'ai Chi,

Fine, whatever. You won't admit your errors, we know.

Tell me something: What use is the "blind" statistician? What is the purpose of him/her not knowing the groups? What, in the statistical analysis, could be influenced without it being discovered?

In other words: How can bias influence the result of the statistical analysis in question?
 
Yahweh said:
Its usually rude to name names, its offensive to name names, and I would consider it unnecessary.

You consider smearing people without having the guts to name them neither rude or offensive?

If Clancie wants to complain about something, she should be precise.

Yahweh said:
Rather than nitpicking, judge Clancie's comments using fair and objective inferences, dont go off trying to pick a fight.

Nobody is trying to pick a fight. But someone is trying to get Clancie to be honest about her complaints.
 
I have to agree with Yahweh and Tai Chi, here, at least in part.

Clancie's support of Rational was qualified; the follow-on efforts do seem to me to be 'picking a fight.'

However, even if you are correct, TLN, might I observe that at best you seem to be engaging in those activities that you just stated you dislike Randi for? Counterproductive, I think.

Just my two cents on that...

Regarding the Triple Blinding: I don't know if such a thing is possible, but I understand what Tai Chi is defining as it. I also understood that the numbers in the example were not meant to reflect any reality; it was simply an example to show how there might be Triple Blinding:

1. Give the analysts the raw data, including identifying what the data describes (e.g., blood pressure)

2. Do not give the analysts the information linking the data to either a specific individual or a group (control or test)

If this either

a) Can't be done for some reason I don't see

or

b) Doesn't constitute Triple Blinding even if done

then those points ought to be argued. But Tai Chi has put forth what he calls Triple Blinding.

That last bit was probably just one cent worth...
 
Clancie said:

Well, scoring is subjective, but how does it break the blind? The sitter -has- to score his/her own readings. There's no way around it (although personally I think the subjectivity of scoring, although unavoidable, is a problem in any mediumship testing).

Why? "Medium blind to sitters...sitters blind to mediums....administrator unaware of who did what...." Why is that single blind at best?

I'd be interested to read your list of what they are.

It breaks the blind because the sitter knows the reading was for them.

If you must have a protocol be sitter-centric, then I would suggest that : the medium not know who is sitting, there be neither sound nor visual transfer from medium to sitter and neither sound nor visual transfer from sitter to medium. The readings are recorded, and randomly played back to the sitters. Half will get their readings, half will get somebody else's readings.
 
Garrette said:
Iut Tai Chi has put forth what he calls Triple Blinding.

That last bit was probably just one cent worth...

I agree he has put forth what he calls triple blinding. It is, however, wrong.

What you described differs from what he described. He described statisticians analyzing the data statistically. That is not blinding in any sense. You describe evaluators evaluating the raw data prior to statistical analysis. That is different, and that is blinding. The classic triple-blind experiment includes the following points:

o Patients don't know if they get IND or placebo
o Treating doctors don't know what patients got IND or placebo
o Final results evaluations are not done by treating doctors, but by evaluating doctors who simply review the MRIs, PET scans or whatever.

It is used for cases where the evaluation is more subjective. BP, for instance, may not be a good example because one can establish criteria a priori, and the evaluator, like the statistician T'ai describes, adds no blinding. Where the situation is more complex and therefore subject to judgment calls, adding a blind evaluator does add blinding.
 
Clancie said:
voidx,

As long as there are skeptical fanatics posting frequently on this board.... people who use "skepticism" as a tool to (1) misrepresent their "opponents'" ideas, (2) badger people relentlessly, and (3) make personal attacks without any intention of having an honest discussion on the subject*...then, yes, I think !Xx-Rational-xX! will continue to be, imo, a highly valued member of the JREF community...and someone who is making an excellent contribution to promoting critical thinking (i.e. let the chips fall where they may!)

It's true that most here -aren't- like the stereotype that he mocks. But his comments are, imo, like a Greek chorus to the observations of those who are!

I guess, once again, we can agree to disagree. :)


*edited to add: TLN, all I can say, is this is my definition of the dishonest skeptic. I have my ideas....you have yours. I'm talking about the issue of fanaticism among so-called skeptics. There's no need for name calling and finger-pointing--my point is that !Xx-Rational-xX! is thought-provoking in his parody of these types. (But, perhaps you think no such skeptics exist. Then, we disagree, but...to each his own! )
Well enjoy watching Towelies battle against the minority then. I'll change the channel myself. Always thought that Majority station had some good programs.
 
BillHoyt said:
He described statisticians analyzing the data statistically. That is not blinding in any sense.

Precisely.

There has to be a reason to "keep people in the dark", if we can talk about "blinding".

There is no reason to "blind" the statistician. I cannot see how a statistician can fake or be influenced in his analysis, because he must be doing his job according to the "laws of statistics", so to speak.

It's easy to check (and, of course, should be done) whether the statistician has used different methods in his analyses. But there is no reason to "blind" him, because, if he cheats or is influenced unknowingly, it will be revealed by checking his work.

Calling this "triple-blind" is simply perverting the concept of blinding. I could just as easily call it "quatro-blinding", because I did not know which was which, either. Hey, another "layer", right? Sure, I had nothing whatsoever to do with the experiment, but hey, let's add superfluous layers so we can claim that false levels of controls were employed.
 
CFLarsen said:
T'ai Chi,
Fine, whatever. You won't admit your errors, we know.


Interesting belief.


Tell me something: What use is the "blind" statistician? What is the purpose of him/her not knowing the groups? What, in the statistical analysis, could be influenced without it being discovered?


You obviously don't understand the need for blinding.

Let's pretend the statistician was a lifelong coffee drinker. He just loves the stuff. A study is done that is trying to see if coffee is bad for you. The experiment is carried out and he gets the data to analyze. Let's say this is not triple blind, so he knows the real group labelings. He could consciously, but most always unconsciosuly, introduce bias into the analysis methods and interpretation of the results to make it appear that the Coffee group has better health than it really does.

How would he do this? Well, there are many different statistical tests one can use. He might unconsciously choose one that gives slightly better results for the Coffee group than others would. There are also many different things one can say in the interpretation of the results, and if coffee is truly bad for health, he might 'soften the blow' so to speak and be less critical of it than he should be.

With the group memberships blind, he analyzes the data and has to reason to favor (consciously or unconsciously) one group over the other, and therefore the analysis is more sound. He simply addresses things as Group 1 or Group 2, and after his analysis is complete, someone else recodes Group 1 and Group 2 with their real labels.
 
BillHoyt said:

I agree he has put forth what he calls triple blinding. It is, however, wrong.


Wrong. I gave the actual commonly used definition of triple blinding.


He described statisticians analyzing the data statistically. That is not blinding in any sense.


Sure it is. The evaluator has no knowledge of who got what treatment, and therefore bias cannot creep in to the analysis and interpretation of the analysis.
 
T'ai Chi said:


Wrong. I gave the actual commonly used definition of triple blinding.

[/b]

Sure it is. The evaluator has no knowledge of who got what treatment, and therefore bias cannot creep in to the analysis and interpretation of the analysis. [/B]
Hey, you know what, each reviewer also has no knowledge of who got what treatment! Also, the editor who decides to publish has no knowledge. Hey, with three reviewers and one editor we're up to a septuple-blind protocol! Woo-woo!

When will you stop playing churlish tr'oll games?
 
T'ai Chi said:
You obviously don't understand the need for blinding.

Let's pretend the statistician was a lifelong coffee drinker. He just loves the stuff. A study is done that is trying to see if coffee is bad for you. The experiment is carried out and he gets the data to analyze. Let's say this is not triple blind, so he knows the real group labelings. He could consciously, but most always unconsciosuly, introduce bias into the analysis methods and interpretation of the results to make it appear that the Coffee group has better health than it really does.

How would he do this? Well, there are many different statistical tests one can use. He might unconsciously choose one that gives slightly better results for the Coffee group than others would. There are also many different things one can say in the interpretation of the results, and if coffee is truly bad for health, he might 'soften the blow' so to speak and be less critical of it than he should be.

With the group memberships blind, he analyzes the data and has to reason to favor (consciously or unconsciously) one group over the other, and therefore the analysis is more sound. He simply addresses things as Group 1 or Group 2, and after his analysis is complete, someone else recodes Group 1 and Group 2 with their real labels.

How can he unconsciously choose one that gives better results for one group?

How will his bias not show up when his analysis is checked by someone else?

How can a statistician be "less critical"? He merely makes the computations, he does not interpret the numbers.
 
Posted by voidx

Well enjoy watching Towelies battle against the minority then.
Thanks, I will. :) (And let's not forget that it's apparently a very vocal minority....)
 
Garrette, Yahweh...:) (And, Garrette, no problem. I don't think my post was clear enough).
Posted by Bill Hoyt
It breaks the blind because the sitter knows the reading was for them.
Not if they are evaluating all the readings without any medium/sitter communication, and without knowing which reading, if any, was for them. (Schwartz tried something like this on the Internet, btw, but, as far as I know, never has published the results).

I think he recognized that the sitters knowing which readings were theirs when they scored them -was- a flaw in his original mediumship experiments.
 

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