Designing a triple-blind experiment?

CFLarsen

Penultimate Amazing
C
Joined
Aug 3, 2001
Messages
42,371
Here's how Steve Grenard describes a "triple-blind" experiment:

A triple blind experiment in the context of mediumship involves blinding of medium, sitter and investigator. In simple terms, the medium and sitter do not know each other, cannot see/hear each other and have no knowledge of each other; there is no feedback from sitter to medium. The investigator does not know which sitter is assigned to which medium because that pairing is randomly drawn by a fourth, uninvolved party and is only revealed after the trial and ratings (using numbers rather than names for sitter/medium) has taken place.
Click to expand...

Steve, BillHoyt has asked you several times, but I'll chip in, too:

How would you design - in detail - a triple-blind experiment of mediums? Including how the evaluators evaluate.
 
For you viewers just tuning in, here is Steve's assertion:

A triple blind experiment in the context of mediumship involves blinding of medium, sitter and investigator. In simple terms,
the medium and sitter do not know each other, cannot see/hear each other and have no knowledge of each other; there is no feedback from sitter to medium. The investigator does not know which sitter is assigned to which medium because that pairing is randomly drawn by a fourth, uninvolved party and is only revealed after the trial and ratings (using numbers rather than names for sitter/medium) has taken place.
Click to expand...

from the Ian Rowland thread

A bit of background: Randi criticized Gary Schwartz' disasterously flawed JE experiments. Schwartz challenged Randi to improve the protocol. Randi responded with a description of a double-blind protocol. Schartz remarked on what a good design suggestion that was, but that he would go one better and make his experiments triple-blind.

Now Grenard has supported Schwartz' contention about the triple-blind experiment. But he has been asked three or four times already to propose one. The above description is as far as he has offered thus far.

Put some popcorn in the microwave, pop open a microbrew, and enjoy. The show should start any minute now...
 
Interesting that Schwartz did not do it right the first time. This supports my contention that he simply is looking for opportunities to do more "better" research. Completely unethical.
 
Ed said:
Interesting that Schwartz did not do it right the first time. This supports my contention that he simply is looking for opportunities to do more "better" research. Completely unethical.
Click to expand...

I'm particularly tickled by the feint of using grad students as controls. Come on! Compare cold-reading with uneducated guessing? What was the null hypothesis there?

H<sub>o</sub>: My lab can pull the wool over the grant underwriter's eyes?

Cheers,
 
I may have missed something here, but haven't many psi experiments been performed where the mediums, sitters, and experimenters are blinded in this fashion?

~~ Paul
 
BillHoyt said:


I'm particularly tickled by the feint of using grad students as controls. Come on! Compare cold-reading with uneducated guessing? What was the null hypothesis there?

H<sub>o</sub>: My lab can pull the wool over the grant underwriter's eyes?

Cheers,
Click to expand...

And, as I recall, the experimenters, the controls, and the subjects knew each other. They mingled and all of that stuff. It sort of sounds more like a cocktail party game than research. And why Grenard even mentions Schwartz is beyond me.

Tell me Bill, if a guy has (evidentially) superb academic credentials and designs a really lousey experiment, would you not suspect that it was done on purpose?
 
Ed said:
Tell me Bill, if a guy has (evidentially) superb academic credentials and designs a really lousey experiment, would you not suspect that it was done on purpose?
Click to expand...

I am shocked, shocked, I tell you, at such an outlandish speculation! Please go wash your mouth out with homeopathic soap and light a votive candle to appease your dead loved ones who surely have surrounded you and are screaming at you so loudly that JE can't think or sleep!

Here's what Gary has to say about his "triple-blind" offer. It was published on Grenard's site. For some surely amusing reason, all of Randi's comments are noted as "Randi:" and all of Gary's comments are noted as "Veritas:"

VERITAS - The latest single-blind experiments rule out cold reading, guessing, selective memory of hits and misses, rater bias, and experimenter bias, from the findings. However, Randi doesn't believe the data. This is because he is convinced this is all "nonsense." By the way, the idea of a "triple-blind" study came up because Randi did not trust our double-blind procedures! We will do a triple-blind study once the double-blind study is completed. Triple-blind is ever more "bells and whistles" (see below).
Click to expand...

From: Survival nonscience

Oh, but when, oh when will the honored guest of the party show up?

Cheers,
 
Ah, the guest arrived, but he's at the neighbor's house. Here is my reply to what he just posted over there. I wonder what he's doing over at the neighbor's house? We left a note for him on the door.

SteveGrenard said:
I described the triple blind procedure in brief. A detailed description, including rating parameters, will take several pages to describe. I will start dowloading my notes and material on
this tonight. I am not ducking the question as several designs have been discussed but I wasn't sure you all wanted this kind of detail. I will be happy to provide it. AFter this is done , I would hope that it will be read and commented on (not just poo-bahed) with constructive criticisms. Thank you.

Larsen: How would you design - in detail - a triple-blind experiment of mediums? Including how the evaluators evaluate.
Click to expand...

Steve,

The question is very simple. It does not require great detail. It certainly does not require notes. Think about it a bit and answer generally, but focus on the evaluation portion, sir. We're all ears.

I shall cross-post this to the thread Claus created.

Cheers,
 
Bill,

It has been a while since I went over the S. stuff, but as I recall, there was no real blinding, just a hospital curtain seperating E and S. Is there a board of professional conduct at most universities? I seem to recall that Tim Leary was booted by one at Harvard.
 
Ed said:
Bill,

It has been a while since I went over the S. stuff, but as I recall, there was no real blinding, just a hospital curtain seperating E and S. Is there a board of professional conduct at most universities? I seem to recall that Tim Leary was booted by one at Harvard.
Click to expand...

Gary maintains his work has already been run through an IRB and approved. His paper described the physical blind as a screen. It turned out to be a folding screen with, obviously, gaps between each panel. His subsequent experiments now use a solid, floor-to-ceiling screen.

I guess Steve is still trying to find a way around his little design dilemma, isn't he? Ah, Steve! I'll give you a hint, in two parts:

a. Chinese Wall

b. Triple-blind protocol is a misnomer. The formal term is triple-blind analysis. The experiment is conducted as a double-blind, but in the analysis phase a third blind is introduced.

Enough hints. I think you see the problem already and are trying to figure a way around it.

Cheers,
 
Paul C. Anagnostopoulos said:
Bill, what's a Chinese Wall as related to experimental design?

~~ Paul
Click to expand...

To blind a trial, information must be kept from one or more participants. In colloquial terms, it must not be allowed over a Chinese wall. In a blinded drug trial, for example, the patient doesn't know if the bottle marked A contains the drug or the placebo. The information is kept from him, safely behind a "chinese wall."

Cheers,
 
Paul C. Anagnostopoulos said:
Bill, what's a Chinese Wall as related to experimental design?

~~ Paul
Click to expand...

I believe it relates to each participant being unaware of the work or strategies employed by any other participants.
 
Just to elaborate a little, I believe a Chinese Wall relates to the analysis phase, in which case no single person sees all of the data. The data is parsed out to several analysts who are each unaware of the strategies used by the others to score the data, and who are unaware of the data that is available to the others.
 
Luke T. said:


I'll see your Venetian blind and add a duck blind.
Click to expand...

Stevie,

The natives are getting restless. They're already starting in on the Quack jokes.

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

Back
Top Bottom