The skeptics are winning

[Ersby]

“Winning�, of course, being a relative term. There is no “winning� in science. And there’s little to suggest that skeptics themselves have contributed to this result. Furthermore, I fully accept that tomorrow may bring contrary evidence. However:

I would instead say that parapyschology is becoming the science it was always supposed to be: the science of self-delusion, bias, how thinking can go wrong in intelligent people, and the like. Personally, I refound my interest in parapsychology after reading about scientific controversies like Continental Shift, Piltdown Man, Eugenics, Phrenology, Physionomy and homosexuality as a mental disorder. When I read about parapsychology’s recent results, my initial response was that this was a “science� waiting to collapse, and I’d be lucky enough to be around when it happened, rather than read about it in a book years later.

Not exactly an unbaised starting point, I admit, but how many of us can truly claim an unbaised starting point?

I’ve been eagerly bevearing away, trying to put together as many ganzfeld experiments as possible in one spreadsheet. I found this very difficult since many of the early experiments are limited only to passing references, especially if they were unsuccessful. This leaves me with little more than a “vote counting� method of deciding if the ganzfeld is successful. After 132 experiments, you’d expect 66 to be positive (even marginally) and 66 negative. I have 72 positive, and 60 negative. This doesn’t really fill me with confidence that an effect has been established, especially since I’ve made no allowances for flawed protocols [ie, Maimonides (handling clues), PRL (video tape degeneration), Sargeant and Blackmore (both poor randomisation), Dalton (faked results?? I’ve sent some emails out for more info. I’ll post my findings when I get them)].

But my recent bullish feeling about the lack of evidence in parapsychology have not just been bolstered by the fact that, out of the last 17 ganzfeld experiments I found that weren’t in any meta-analysis, 15 had negative results. They’ve also been supported in the last few days by the idea that Dalton’s work may not be as scientifically sound as we might like. Additionally, a search for other recent papers the 2004 PA brings up quotes like:

(a search on google for the titles should bring up a link to the pdf)

“On the false hypothesis of psi-mediated shift of statistical average in test with random event generators� (Pallikari)
The most recent meta-analysis based on all available to date micro-PK data, testing the direct influence of human intention on the outcomes of true random number generators (RNG), does not support this hypothesis. Furthermore, a very large-scale experiment set out by a consortium of independent research groups to replicate the micro-PK hypothesis, has failed to show the effect.

“A test of predictions from five studies on telepathic group communication of emotions� (Westerlund, Dalkvist)
The three explanations denying the occurrence of any telepathic effect were: (1) All significant results obtained in the original experiments were caused by random variation. (2) The replication experiments were better controlled than the original experiments. (3) The original positive results were obtained through systematic selection. All of these three explanations were judged to be plausible.


Exploring the reliability of the “presentiment effect� (Broughton)
This project extends a series of experiments begun by Radin that appears to demonstrate an autonomic nervous system (ANS) response to future emotionally arousing experiences. […] The experiment did not produce overall evidence of a presentiment effect, nor did it demonstrate test-retest reliability. Since this experiment deviated from prior methods of analysis the data were examined without de-trending and automatic artifact removal, but there was still no evidence of a presentiment effect.

“The Invisible Gaze: three attempts to replicate Sheldrake’s Staring effects� (Lobach, Bierman)
We did three studies to attempt to replicate Sheldrake’s staring studies under conditions of sensory shielding. Participant input their responses in the computer. In total 188 sessions resulted in 4784 trials. The over-all hitrates were 50.6% (N=53), 52.1% (N=45), and 49.7% (N=37) respectively. […] From the results of these three studies we conclude that the staring paradigm is not the easily replicable paradigm that it is claimed to be. Apart from one internal effect all hitrates were very close to chance.

I'm not suggesting that there are no positive results in parapsychology at present (that, in itself, would go against chance) but that these important pillars of parapsychology should be doubted at the same time is important, I feel.

At the very most, it indicates that Rhine’s hypothesis (that psi is present in everybody, not just the talented few) is wrong. Which means that parapsychology has to concentrate on those individuals that demonstrate this ability. But this greatly increases the chances of an individual being particularly good at decieving, rather than at psi. In other words, it’s progressed no further than the “Project Alpha� experiments some twenty or thirty years ago.

Ho hum.

[Anders]

Originally posted by Ersby

[snip]

I’ve been eagerly bevearing away, trying to put together as many ganzfeld experiments as possible in one spreadsheet. I found this very difficult since many of the early experiments are limited only to passing references, especially if they were unsuccessful. This leaves me with little more than a “vote counting� method of deciding if the ganzfeld is successful. After 132 experiments, you’d expect 66 to be positive (even marginally) and 66 negative. I have 72 positive, and 60 negative. This doesn’t really fill me with confidence that an effect has been established, especially since I’ve made no allowances for flawed protocols [ie, Maimonides (handling clues), PRL (video tape degeneration), Sargeant and Blackmore (both poor randomisation), Dalton (faked results?? I’ve sent some emails out for more info. I’ll post my findings when I get them)].
[snip]

Great work Ersby, you even got one of the Swedish results!

[Ashles]

Survey published today comparing attitudes in 1954 to 2004:

% who believe in ghosts:
1954: 10%
2004: 42%

:(

[Ersby]

Originally posted by Ashles
Survey published today comparing attitudes in 1954 to 2004:

% who believe in ghosts:
1954: 10%
2004: 42%

:(

Well, there's winning and there's winning. I was talking about winning re. scientific standards.

I personally "believe" in ghosts because it's entertaining: it pleases me to do so. "Ring" wouldn't be half as good a film if I didn't. (The Japanese version, not the US version which I haven't seen) And I enjoy M.R.James' stories etc.

Do you have a link to the survey?

[Ashles]

Sorry, yes it is a bit of an unverified statement as it stands.

It is referring to the UK population by the way.

It was a survey by ICM as a follow-up to a Gallup survey in 1954. (I read about it in er, the Sun newspaper)

These are the guys who carried out the poll (http://www.icmresearch.co.uk/default.asp)
but I can't find the specific poll referred to (it was a general comparison about social conditions compared to 50 years ago - the beliefs section was only briefly mentioned)

[Ersby]

Ah well, not to worry.

Meanwhile, there's a remarkable lack of reaction from the recent wave of people trumpeting their evidence of psi.

I think I'll just bump this thread every now and again. Just to let them know the argument's been lost.

[Anders]

Originally posted by Ersby
Ah well, not to worry.

Meanwhile, there's a remarkable lack of reaction from the recent wave of people trumpeting their evidence of psi.

I think I'll just bump this thread every now and again. Just to let them know the argument's been lost.
Hey Ersby,

Do you have any numbers on quality of studies and their outcome? Can we show that good quality studies show lesser ESP effect than poor quality studies?

[Ashles]

Guaranteed Davidsmith and De'Ville's Advocaat will still claim that this demonstrates nothing but huge evidence that Psi exists.

But it will be fun to watch them try and twist the facts to suit their beliefs.

Is there a good resource which keeps track of all the Ganzfeld results? So we could look at them ourselves without having to hear about them from the believers (who, obviously only pass on any pro-psi news)?

[Ersby]

Originally posted by Anders
Hey Ersby,

Do you have any numbers on quality of studies and their outcome? Can we show that good quality studies show lesser ESP effect than poor quality studies?

No. When I put together my spreadsheet about the ganzfeld (if anyone wants it, they can email me) I ignored issues like that. There is a decline from early (pre autoganzfeld) top later (post autoganzfeld). Especially if Dalton's work is suspect (which remains to be seen.)

[Interesting Ian]

Originally posted by Ersby
After 132 experiments, you’d expect 66 to be positive (even marginally) and 66 negative. I have 72 positive, and 60 negative. This doesn’t really fill me with confidence that an effect has been established, especially since I’ve made no allowances for flawed protocols [ie, Maimonides (handling clues), PRL (video tape degeneration), Sargeant and Blackmore (both poor randomisation), Dalton (faked results?? I’ve sent some emails out for more info. I’ll post my findings when I get them)].


Could you clarify what you mean by a positive result, and a negative result? Are you just talking about above average result as opposed to below average result, and not taking into consideration statistical significance at all?

[TLN]

Originally posted by Interesting Ian
Could you clarify what you mean by a positive result, and a negative result? Are you just talking about above average result as opposed to below average result, and not taking into consideration statistical significance at all?

But there is no statistical significance; that's the whole point.

[davidsmith73]

Originally posted by Ersby

But my recent bullish feeling about the lack of evidence in parapsychology have not just been bolstered by the fact that, out of the last 17 ganzfeld experiments I found that weren’t in any meta-analysis, 15 had negative results.

Which studies are these?


Additionally, a search for other recent papers the 2004 PA brings up quotes like:

“On the false hypothesis of psi-mediated shift of statistical average in test with random event generators� (Pallikari)


“A test of predictions from five studies on telepathic group communication of emotions� (Westerlund, Dalkvist)

Exploring the reliability of the “presentiment effect� (Broughton)


“The Invisible Gaze: three attempts to replicate Sheldrake’s Staring effects� (Lobach, Bierman)


I'm not suggesting that there are no positive results in parapsychology at present (that, in itself, would go against chance) but that these important pillars of parapsychology should be doubted at the same time is important, I feel.




These are some of the successful papers from the conference procedings:

"Anticipatory skin conductance responses: a possible example of decision augmentation theory"

"The precognitive habituation effect: an adaptation using spider stimuli"

"Physiological correlates of ESP: heart rate differences between targets and non-targets in clairvoyance and precognition forced choice tasks"

"Precognitive avoidance and precognitive déjß vu"

(there might be a few more, but the links weren't working)

These papers can be read at:

http://a1162.fmg.uva.nl/research/PSI/PA2004org.html


So we have a bit of a mix of negative and positive results. It's no wonder meta-analyses are used to try to resolve the debate. However, the presentiment type experiments seem very promising. The Broughton paper that Ersby referenced is a failed example but there are many experiments of this type that were successful.

[Ashles]

Told you it would be fun.

[Anders]

Originally posted by davidsmith73

[snip]
The Broughton paper that Ersby referenced is a failed example but there are many experiments of this type that were successful.

Your use of the word “successful� implicates, in my opinion, that you are biased. In science every good quality experiment is successful even if it shows that the hypothesis was wrong. But then again, you don’t ever accept science as a tool to get to know the truth about the world, or do you?

[Open Mind]

The Invisible Gaze: three attempts to replicate Sheldrake’s Staring effects� (Lobach, Bierman)

But they later got positive evidence in Sheldrake's telephone experiments.

As for dismissing Kathy Dalton's trial ....is that not the common sceptical attitude? If the result is too positive - error or fraud is the assumption, if it is moderately weak - it's experiment error, if it is very weak - it's random variations and if it's negative - it's correct and beyond doubt! :D


At the very most, it indicates that Rhine’s hypothesis (that psi is present in everybody, not just the talented few) is wrong.


I also think it is much stronger in some than others. I'd like to see trials restricting the choice to adult psychics who had psychic experiences as children (but not when sleeping! That doesn't count at all) Those who developed it later (or so they believe) should be dismissed from trial only natural psychics, they don't need to be well known psychics necessarily.

I think PSI is very weak in most humans and it becomes practically non existant when they don't think it can exist.

[Ashles]

If the result is too positive - error or fraud is the assumption, if it is moderately weak - it's experiment error, if it is very weak - it's random variations and if it's negative - it's correct and beyond doubt!
"If the result is too positive " - We don't know. Never seen a replicable example of this

"if it is moderately weak" - It's hard to say. We've never seen a replicable example of this

"if it is very weak - it's random variations" - Well that and statistical manipulation, and possibly reporting of positive(ish) results and ignoring negative. A certain level of experimenter bias. Hardly convincing however you view it. When your positive results are so incredibly slight then it could be just about anything.

I think you left out "no effect" which is odd as it is almost entirely what experimental results actually show.

"and if it's negative" - If it's negative then we are seeing the opposite of the claimed paranormal effect and the claimant must be Uri Geller.

[Joe_Black]

You keep banging that drum Ersby.

[Dr Adequate]

You keep collecting and reporting as much data as possible, Ersby. And see who complains --- the sceptics or the woowoos.

[Ashles]

You keep banging that drum Ersby.
Wow!

Yeah, it's really the sceptics who keep banging on about the Ganzfeld, isn't it Joe?

Tell you what, when believers stop making claims about what the Ganzfeld data 'shows' we'll stop analysing that data.
Deal?

[amherst]

Originally posted by Ersby
I’ve been eagerly bevearing away, trying to put together as many ganzfeld experiments as possible in one spreadsheet. I found this very difficult since many of the early experiments are limited only to passing references, especially if they were unsuccessful. This leaves me with little more than a “vote counting� method of deciding if the ganzfeld is successful. After 132 experiments, you’d expect 66 to be positive (even marginally) and 66 negative. I have 72 positive, and 60 negative. This doesn’t really fill me with confidence that an effect has been established...
4.3 The Vote-Counting Debate (http://anson.ucdavis.edu/~utts/91rmp.html)

Vote-counting is the term commonly used for the technique of drawing inferences about an experimental effect by counting the number of significant versus nonsignificant studies of the effect. Hedges and Olkin (1985) give a detailed analysis of the inadequacy of this method, showing that it is more and more likely to make the wrong decision as the number of studies increases. While Hyman acknowledged that "vote-counting raises many problems" (Hyman, 1985b, page 8), he nonetheless spent half of his critique of the ganzfeld studies showing why Honorton's count of 55% was wrong.
Do you want to know why Hyman "spent half of his critique of the ganzfeld studies showing why Honorton's count of 55% was wrong?" I don't think you do Ersby. I think the last thing you want to do is realize that (once again) your own "analysis" should have given you "confidence that an effect has been established."

amherst

[dharlow]

Originally posted by Ersby

I would instead say that parapyschology is becoming the science it was always supposed to be: the science of self-delusion, bias, how thinking can go wrong in intelligent people, and the like.


I'm not sure exactly how you can state this since your conclusion is based, in part, on the data of parapsychologists themselves which suggests no effect. How are they deluding themselves when they have presented this information? Now as to some of the non-experimentalist proponents who go on and on about certain research lines and probably don't even read some of this stuff...that's a different story.

[CFLarsen]

Originally posted by Open Mind
But they later got positive evidence in Sheldrake's telephone experiments.

Would this be the experiments where Sheldrake didn't even think about preventing the old "Synchronize the clock"-routine?

Wasn't it Paulie the Greek who asked him about this and got a "Well, I nevah thought of that" reply?

[Paul C. Anagnostopoulos]

Yup, that would be me. Then he said he went back and checked for clock synchronization leaks and determined that they were not a factor. Righty-o.

Lesson: Don't do experiments with the subjects sitting in their homes, even if you do have video cameras.

Has anyone read any results from the email telepathy experiments?

~~ Paul

[CFLarsen]

Originally posted by Paul C. Anagnostopoulos
Yup, that would be me. Then he said he went back and checked for clock synchronization leaks and determined that they were not a factor. Righty-o.

Can you remember what reason he gave?

Originally posted by Paul C. Anagnostopoulos
Lesson: Don't do experiments with the subjects sitting in their homes, even if you do have video cameras.

Lesson: Don't do experiments with the subjects being in control.

I still remember how Jamie Ian Swiss fooled us at TAM2, up close and personal. Had it been any closer, we would have been arrested for indecent behavior. Skeptics were swarming all over him, and he still managed to trick us.

Damn, that guy is good! May he never turn to the dark side, because he will make trillions! If not billions!

[Open Mind]

Eva Lobach & Dick J. Bierman telepheon trial

http://www.sheldrake.org/articlesnew/pdf/Lobach.pdf

[Ersby]

Originally posted by Interesting Ian
Could you clarify what you mean by a positive result, and a negative result? Are you just talking about above average result as opposed to below average result, and not taking into consideration statistical significance at all?

Because a lot of the earlier unsuccessful experiments have little to no information about them, other than a passing reference that "no psi was found" I couldn't take into account statistical significance. I merely counted unsuccessful (25%and below) and successful (above 25%).

It is not a very good way of doing any kind of statistical analysis, but given the figures I have, I don't really have much else.

[Ersby]

Originally posted by Open Mind
As for dismissing Kathy Dalton's trial ....is that not the common sceptical attitude? If the result is too positive - error or fraud is the assumption, if it is moderately weak - it's experiment error, if it is very weak - it's random variations and if it's negative - it's correct and beyond doubt! :D


I'm not dismissing because the results are too high, but because people (other parapsychologists) have said they think the results may have been tampered with. I'm still waiting for a reply from the people involved.

[Ersby]

Originally posted by davidsmith73
(there might be a few more, but the links weren't working)

These papers can be read at:

http://a1162.fmg.uva.nl/research/PSI/PA2004org.html



I've just found the link for those papers that aren't working.

http://a1162.fmg.uva.nl/research/PSI/nopapers/

[CFLarsen]

Originally posted by Open Mind
Eva Lobach & Dick J. Bierman telepheon trial

http://www.sheldrake.org/articlesnew/pdf/Lobach.pdf

And? What do you want to say about this?

[Drooper]

I'm afraid I can't feel so optomistic about winning the argument for critical thinking. Especially when my health insurer has just announced with much fanfare that they are to extend thier benefits to yet another "complementary therapy" - reflexology not less!!


And I have to pay for it AAAHHHHHHH!!!!!:mad: :mad:

[spelling grrrr:mad:]

[amherst]

Originally posted by Ersby
Because a lot of the earlier unsuccessful experiments have little to no information about them, other than a passing reference that "no psi was found" I couldn't take into account statistical significance.
Yes you could have. 11 of the 40 studies reported in the "Victim of Its Own Success" paper, 2 of the 11 PRL studies, 3 of the 11 Storm and Ertel studies, and 12 of the 28 studies in Honorton's original meta-analysis were statistically significant. So let's just say that the rest of the 132 studies that you're aware of were not. 28 is 21% of 132. Chance alone expects 5%, or 6.6 studies out of 132, to be statistically significant. Needless to say, this is a significant effect. And if you were to limit the analysis to just standard studies, it would obviously be even more so.
I'm not dismissing [Dalton's experiments] because the results are too high, but because people (other parapsychologists) have said they think the results may have been tampered with. I'm still waiting for a reply from the people involved.
Even if it is found that Dalton committed some sort of fraud, and that her work should be put into question, this doesn't mean the ganzfeld results "fall to pieces" or even that the standardness analysis fails. With Dalton's two experiments excluded, the Stouffer Z of the standard replications becomes 2.27, which is still statistically significant, p= .0116, one tailed.

amherst

[Dr Adequate]

Originally posted by Open Mind
As for dismissing Kathy Dalton's trial ....is that not the common sceptical attitude?
The Dalton results have been left in until evidence, rather than rumor and conjecture, can be produced for fakery.

This would indeed be "the common sceptical attitude".

[Interesting Ian]

Originally posted by Ersby
Because a lot of the earlier unsuccessful experiments have little to no information about them, other than a passing reference that "no psi was found" I couldn't take into account statistical significance. I merely counted unsuccessful (25%and below) and successful (above 25%).



When they say no psi was found they won't mean the result is 25% or less!

And successful does not mean anything above 25%!

Sorry, I don't trust your figures at all.

[Anders]

Originally posted by Interesting Ian
When they say no psi was found they won't mean the result is 25% or less!

And successful does not mean anything above 25%!

Sorry, I don't trust your figures at all.
So what does mean success and what does mean failure?

[CFLarsen]

Originally posted by Anders
So what does mean success and what does mean failure?

Whatever "God allows".

[Anders]

Originally posted by CFLarsen
Whatever "God allows".
Yes, that's it. I forgot!

[Interesting Ian]

Originally posted by Anders
So what does mean success and what does mean failure?

I don't know. Isn't it a probablity of 1 in 20 or less? Obviously that's arbitrary.

[Anders]

Originally posted by Interesting Ian
I don't know. Isn't it a probablity of 1 in 20 or less? Obviously that's arbitrary.
Funny thing you know what success and failure is NOT, but you have no idea what it IS.

Dependes on the significance intervall, and usually 5%, or 1%. But that also depends on the number of objects.

[davidsmith73]

Originally posted by Anders
Your use of the word “successful� implicates, in my opinion, that you are biased. In science every good quality experiment is successful even if it shows that the hypothesis was wrong. But then again, you don’t ever accept science as a tool to get to know the truth about the world, or do you?

I accept the philosophy of science as it stands and all of its standards. By successful I mean in terms of statistically refuting the null hypothesis. I don't see what's biased about that. What term would you use?

[Interesting Ian]

Originally posted by Anders
Funny thing you know what success and failure is NOT, but you have no idea what it IS.

Dependes on the significance intervall, and usually 5%, or 1%.



I do not know what "significance intervall(sic)" means. And 5%/1% of what??

No one "knows" what it is. I said that the decision here is to a certain extent an arbitrary one. Did you miss that word in my very short post??



But that also depends on the number of objects.

Number of objects? You mean number of trials? Number of potential targets?? How do you think the probability is calculated? :rolleyes:

[Anders]

Originally posted by Interesting Ian
I do not know what "significance intervall(sic)" means. And 5%/1% of what??

No one "knows" what it is. I said that the decision here is to a certain extent an arbitrary one. Did you miss that word in my very short post??



Number of objects? You mean number of trials? Number of potential targets?? How do you think the probability is calculated? :rolleyes:
I can just give you one piece of advice: Study some basic statistics, so have I. But I might have got the english terms wrong, here, and that could be the reason you don't understand. In that case, sorry about that.

[Interesting Ian]

Originally posted by Anders
I can just give you one piece of advice: Study some basic statistics, so have I. But I might have got the english terms wrong, here, and that could be the reason you don't understand. In that case, sorry about that.

Why should I need to study basic statistics? I'm wondering what proportion of parapsychological experiments yield statistically significant results. Let's define statistical significance as a probability of 1 in 20 of getting obtained result should no anomalous factors be in operation, and no artefacts are present skewing the result. What's the problem?

And if you increase the number of trials, but the subjects get the same proportion right, then this will decrease the probability of getting the obtained result - right? And if over many experiments with varied protocols and different experimenters, we still, by meta-analysis, still get statistically significant results, then this would tend to suggest that artefacts are not responsible - yes?

So what's the data? What does Ersby data actually mean. Is anything over 25% (presumably getting correct target out of 4) supposed to be statistically significant? Obviously that is crazy. I don't understand his figures. Do you??

[Anders]

Originally posted by Interesting Ian
Why should I need to study basic statistics? I'm wondering what proportion of parapsychological experiments yield statistically significant results. Let's define statistical significance as a probability of 1 in 20 of getting obtained result should no anomalous factors be in operation, and no artefacts are present skewing the result. What's the problem?

And if you increase the number of trials, but the subjects get the same proportion right, then this will decrease the probability of getting the obtained result - right? And if over many experiments with varied protocols and different experimenters, we still, by meta-analysis, still get statistically significant results, then this would tend to suggest that artefacts are not responsible - yes?

So what's the data? What does Ersby data actually mean. Is anything over 25% (presumably getting correct target out of 4) supposed to be statistically significant? Obviously that is crazy. I don't understand his figures. Do you??
Yes I do understand his figures. 25% is the result inticipated by chance alone. Anything over or lower than that would implicate something else than chance. Although I don't see how it could fall below 25%, and it seems, looking at the figures, it don't. Dalton got about 33% which is pretty much.

But chance is a funny thing. To be sure, the result has to be a lot over 25%, like 30% or 34% to implicate something else than chance. But that has to do with the number of test subjects. The more test subjects you have, the more certain can you be on the numbers, and the lesser margin you can have.

Some result over 25% has been shown to be introduced by errors in the set up or the process of the experiment.

[Interesting Ian]

Originally posted by Anders
Yes I do understand his figures. 25% is the result inticipated by chance alone.


You'd need to give a range e.g. 20%-30%; unfortunately that's no good because it will also depend on the number of trials. Hence the need to introduce this concept of statistical significance. For example a result of 30% may not be very impressive and will not be statistically significant with very few trials. On the other hand, maybe with a colossal number of trials, but only getting 27%, this might well be statistically significant (i.e. the chances of getting such a result are 1 in 20 or less).




Anything over or lower than that would implicate something else than chance. Although I don't see how it could fall below 25%, and it seems, looking at the figures, it don't. Dalton got about 33% which is pretty much.

But chance is a funny thing. To be sure, the result has to be a lot over 25%, like 30% or 34% to implicate something else than chance. But that has to do with the number of test subjects. The more test subjects you have, the more certain can you be on the numbers, and the lesser margin you can have.



My comments just above address this.





Some result over 25% has been shown to be introduced by errors in the set up or the process of the experiment. [/B]

An experiment result over 25% has been shown to be in error? Ummm . .thanks for the meticulous detail ;)

[amherst]

/

[Anders]

Originally posted by Interesting Ian
You'd need to give a range e.g. 20%-30%; unfortunately that's no good because it will also depend on the number of trials. Hence the need to introduce this concept of statistical significance. For example a result of 30% may not be very impressive and will not be statistically significant with very few trials. On the other hand, maybe with a colossal number of trials, but only getting 27%, this might well be statistically significant (i.e. the chances of getting such a result are 1 in 20 or less).

I could give you specifics, but I'm too lazy to get up, get my statistics textbooks, find the right numbers and post them here.

But you are basically correct! The more trials the lesser interval.



My comments just above address this.

An experiment result over 25% has been shown to be in error? Ummm . .thanks for the meticulous detail ;)
Hey! no need to get ironic on me! But we might both be of the ironic generation, so it's ok! :)

[Aussie Thinker]

Ersby

I appreciate your attempt to dispel the last line of woo defence that “psi� exists.

It will fail though because these people REFUSE to accept a scientific result that shows a “psi� effect does not exist..

Examples..

1. Results do not and never will = chance.. therefore psi exists (remember these loons do not think statistically = to chance IS = to chance)
2. ONE positive result is enough to prove psi exists
3. Negative and positive results show that chance is not happening therefore psi exists.

For the life of me I just cannot understand them.

Why do they maintain a belief in something that requires so little evidence to maintain. What evidence there is is EXTREMELY vague and its arguable wether it exists at all. The effect if it is exists is indistinguishable from chance and produces NOTHING.

The ONLY time an experiment shows arguable “positive� results is when it is a vague indeterminate useless test of some “power� that has no use, is not relevant and has no application or continued line of research or development.

[flyboy217]

Originally posted by Anders
But chance is a funny thing. To be sure, the result has to be a lot over 25%, like 30% or 34% to implicate something else than chance.

This is patently false. If you were to flip a (perfect) coin one billion times and get 51% heads, you'd have something startling indeed.

[flyboy217]

Originally posted by Ashles
Survey published today comparing attitudes in 1954 to 2004:

% who believe in ghosts:
1954: 10%
2004: 42%

:(

http://www.belfasttelegraph.co.uk/news/features/story.jsp?story=585419


Faith in the paranormal, though, is on the increase, with 42 per cent of us claiming to believe in ghosts, a 10 per cent rise over half a century.

[Interesting Ian]

Originally posted by flyboy217
This is patently false. If you were to flip a (perfect) coin one billion times and get 51% heads, you'd have something startling indeed.

But not at all startling to skeptics. Nothing short of lifting a large concrete block just by using ones mind would convince them psychokinesis exists. This is why these discussions on here are such a waste of time.

[Aussie Thinker]

Flyboy,

This is patently false. If you were to flip a (perfect) coin one billion times and get 51% heads, you'd have something startling indeed

But if you tossed the coins a few hundred times (more statistically inline with “psi� testing) then a few % difference would be meaningless.

Then you have to add in the ambiguity of the “testing�.. with a coin you have heads or tails. With ‘psi� no one has a friggin’ clue.. an “effect� is usually only “found� after the fact. I would agree with the argument that I would want to see a SIGNIFICANT result better than chance before entertaining the idea of “psi�

[Aussie Thinker]

Ian

But not at all startling to skeptics. Nothing short of lifting a large concrete block just by using ones mind would convince them psychokinesis exists. This is why these discussions on here are such a waste of time.

You KNOW that is not true…

We would be satisfied if you could lift a FEATHER with your own mind.

But NO ONE ever has ?

All we ask is a small but real manifestation of “psi�.. It NEVER shows up..

[Anders]

Originally posted by flyboy217
This is patently false. If you were to flip a (perfect) coin one billion times and get 51% heads, you'd have something startling indeed.
You are correct my friend! (who said skeptics are all negative). But I was refering to the number of trials that are usually carried out in the "average" Ganzfeld trial. If you made a billion trials, got a 0.5% above 25%, then we might have a nice result, worth publishing in Nature!

[Ersby]

Originally posted by amherst
Yes you could have. 11 of the 40 studies reported in the "Victim of Its Own Success" paper, 2 of the 11 PRL studies, 3 of the 11 Storm and Ertel studies, and 12 of the 28 studies in Honorton's original meta-analysis were statistically significant. So let's just say that the rest of the 132 studies that you're aware of were not. 28 is 21% of 132. Chance alone expects 5%, or 6.6 studies out of 132, to be statistically significant. Needless to say, this is a significant effect. And if you were to limit the analysis to just standard studies, it would obviously be even more so.

amherst

The whole reason I started collecting ganzfeld experiments was to see them all together, using the same statistical measure. Honorton's doesn't use the same method as the Victim of its own success. So I chose Cohen's d and effect size r as the best measures. A "large" effect according to Cohen's d is a 0.8, but that gives only 14 or so significant experiments out of the 135 (three more added in the last day!) Even I think that can't be right, and there've been more than that. Anyone know what p=0.05 is when translated in chohen's d?

And, of course, we'd have to look at those experiments above that threshold and see if any are flawed.

[CFLarsen]

Originally posted by Aussie Thinker
But if you tossed the coins a few hundred times (more statistically inline with “psi� testing) then a few % difference would be meaningless.

You can say that again.

From Ersby's extensive list of RV experiments, the highest number of trials I could find was 206. Then, 120, 116 and 100. The rest were between 1 and 100.

A far cry from a billion, me thinks.

I just flipped a coin 206 times, and came out 116 Heads and 90 Tails. A better result than a lot of RV experiments.

[dharlow]

Originally posted by Ersby
Anyone know what p=0.05 is when translated in chohen's d?

Any effect size calculation will be dependent on the number of trials. A small number of trials that achieves p=0.05 will have a higher effect size than a large number of trials that reaches p=0.05.

[amherst]

Originally posted by Ersby
The whole reason I started collecting ganzfeld experiments was to see them all together, using the same statistical measure. Honorton's doesn't use the same method as the Victim of its own success. So I chose Cohen's d and effect size r as the best measures. A "large" effect according to Cohen's d is a 0.8, but that gives only 14 or so significant experiments out of the 135 (three more added in the last day!) Even I think that can't be right, and there've been more than that. Anyone know what p=0.05 is when translated in chohen's d?

And, of course, we'd have to look at those experiments above that threshold and see if any are flawed.
You can tell if an experiment was statistically significant by looking at its Z score. A Z of 1.65 translates into a one-tailed p of .05. I'm aware of 28 ganzfeld studies which have Z's that meet or exceed this criterion. If there are 135 studies, only 6.75 would have achieved significance by chance alone. The vote-counting method doesn't support your skepticism.

You also don't understand that a phenomenon can have a small or medium effect size and still be real. Bem & Honorton (1994) pointed out that the psi effect in the ganzfeld falls just short of what Cohen considers a medium effect size to be:

Cohen (1988, 1992) has also categorized effect sizes into small, medium, and large, with medium denoting an effect size that should be apparent to the naked eye of a careful observer. For a statistic such as pi, which indexes the deviation of a proportion from .5, Cohen considers .65 to be a medium effect size: A statistically unaided observer should be able to detect the bias of a coin that comes up heads on 65% of the trials. Thus, at .62, the psi ganzfeld effect size falls just short of Cohen's naked-eye criterion. From the phenomenology of the ganzfeld experimenter, the corresponding hit rate of 35% implies that he or she will see a subject obtain a hit approximately every third session rather than every fourth.

It is also instructive to compare the psi ganzfeld effect with the results of a recent medical study that sought to determine whether aspirin can prevent heart attacks (Steering Committee of the Physicians' Health Study Research Group, 1988). The study was discontinued after 6 years because it was already clear that the aspirin treatment was effective (p < .00001) and it was considered unethical to keep the control group on placebo medication. The study was widely publicized as a major medical breakthrough. But despite its undisputed reality and practical importance, the size of the aspirin effect is quite small: Taking aspirin reduces the probability of suffering a heart attack by only .008. The corresponding effect size (h) is .068, about one third to one fourth the size of the psi ganzfeld effect (Atkinson et al., 1993, p. 236; Utts, 1991b).
I think it would be wise of you to leave criticisms of the ganzfeld to skeptics like Hyman and Wiseman. If anyone is going to disprove the results of these experiments, it's not going to be you.

amherst

[Ashles]

From Flyboy's link (http://www.belfasttelegraph.co.uk/news/features/story.jsp?story=585419)
Interest in formal religion is declining: 90 per cent of adults owned a Bible in 1954, a figure now at 65 per cent. Faith in the paranormal, though, is on the increase, with 42 per cent of us claiming to believe in ghosts, a 10 per cent rise over half a century.
Well that's not as bad as the 32% increase quoted in the Sun. Their reporter obviously misread that it was a 10% increase and just wrote that it used to be 10% in 1954.
Poor journalism in the Sun - who would've thought it?

Even the slight increase in the paranormal makes me think the X-Files has a lot to answer for.

[Interesting Ian]

Originally posted by Aussie Thinker
Ian



You KNOW that is not true…

We would be satisfied if you could lift a FEATHER with your own mind.

But NO ONE ever has ?

All we ask is a small but real manifestation of “psi�.. It NEVER shows up..

Lifting a feather up is probably not possible. Try asking someone to influence a random number generator so that 51% of 1's and only 49% of 0's are generated, and vice versa, for a billion 0 and 1's outputted.

[Ashles]

Try asking someone to influence a random number generator so that 51% of 1's and only 49% of 0's are generated, and vice versa, for a billion 0 and 1's outputted.
Yeah, and see what they say.

"You want me to sit in front of a machine for HOW long?"

EDITED TO ADD:
Ian, why do you think the feather is not possible but affecting a random number generator is? What distinction do you draw between these two abilities?

[CFLarsen]

Originally posted by Interesting Ian
Lifting a feather up is probably not possible. Try asking someone to influence a random number generator so that 51% of 1's and only 49% of 0's are generated, and vice versa, for a billion 0 and 1's outputted.

How long would that take?

1 second per number?

More than 31 years, non stop.

[Interesting Ian]

Originally posted by CFLarsen
How long would that take?

1 second per number?

More than 31 years, non stop.

A computer can only output 1 number per second. WOW!! :confused:

[Dr Adequate]

Who mentioned computers. What is needed is a random number generator. The so-called RNGs which computers use are pseudorandom --- deterministic but in a complicated way. To influence them, you'd have to psychically change the value stored in some part of the computer's memory. This would be a lot easier to test for that any statistical effect.

As far a speed is concerned, consider also that the question would be how long it takes to influence the computer, not how fast the computer can turn out the numbers. Can a person with psi powers perform a million acts of psi per second?

[CFLarsen]

Originally posted by Interesting Ian
A computer can only output 1 number per second. WOW!! :confused:

No, I was thinking about how long a person has to focus on a number, before the number can be changed.

Perhaps you have a better design? (I will ask you why you chose your design, so you better be prepared)

[Jeff Corey]

Originally posted by CFLarsen
You can say that again.

From Ersby's extensive list of RV experiments, the highest number of trials I could find was 206. Then, 120, 116 and 100. The rest were between 1 and 100.

A far cry from a billion, me thinks.

I just flipped a coin 206 times, and came out 116 Heads and 90 Tails. A better result than a lot of RV experiments.
Claus, your result has a probability of .0813 according to a two-tailed Binary Test.
So discard it and keep trying until you achieve a p value of less than .05 and then publish that. The file drawer effect is the usual suspect in these studies.

[CFLarsen]

Originally posted by Jeff Corey
Claus, your result has a probability of .0813 according to a two-tailed Binary Test.
So discard it and keep trying until you achieve a p value of less than .05 and then publish that. The file drawer effect is the usual suspect in these studies.

I'd rather count the total number of sidewalk tiles on Manhattan than flip a coin for that long again.... :)

[steenkh]

Originally posted by Dr Adequate
Who mentioned computers. What is needed is a random number generator. The so-called RNGs which computers use are pseudorandom --- deterministic but in a complicated way. To influence them, you'd have to psychically change the value stored in some part of the computer's memory. This would be a lot easier to test for that any statistical effect.

Indeed. If it was possible psychically to alter the output of a pseudo-random number generator, it would be a lot more interesting how the psi power located the right bits inside the CPU than the actual change itself! Very often, if you change any bits inside a processor or in the RAM where the program is stored, you would cause the computer to crash. If it had been possible to cause computers to crash psychically, we would know by now.

I am aware that some people cause computers to crash all the time, but they always interact with the computer in the conventional way, so that does not count.

[Ersby]

Originally posted by amherst
You can tell if an experiment was statistically significant by looking at its Z score. A Z of 1.65 translates into a one-tailed p of .05. I'm aware of 28 ganzfeld studies which have Z's that meet or exceed this criterion. If there are 135 studies, only 6.75 would have achieved significance by chance alone. The vote-counting method doesn't support your skepticism.


Thanks, I'll see if that makes a difference to my sums


I think it would be wise of you to leave criticisms of the ganzfeld to skeptics like Hyman and Wiseman. If anyone is going to disprove the results of these experiments, it's not going to be you.

amherst

Well, honestly, if anyone here has that kind of attitude about anything, I advise them to just go to bed and not come out again!

"Leave it to the experts" is exactly the kind of attitude I abhor.

[Ashles]

I think it would be wise of you to leave criticisms of the ganzfeld to skeptics like Hyman and Wiseman. If anyone is going to disprove the results of these experiments, it's not going to be you.
So it's okay for anyone who feels like it to promote Ganzfeld studies data, but only certain people are allowed to analyze that data?

Sounds like some believers don't have too much confidence in that data.

"You can look at it but you're not to think about it too hard. Leave it alone! Don't criticise it! You're spoiling it all now. I want to believe in the magic!"

[davidsmith73]

Another study from the 2004 convention with significant effects. It's a study using EEG to measure a response in "receivers" brainwaves to the stimulation of "senders" visual field with light flashes.

Are the sceptics really winning?

http://a1162.fmg.uva.nl/research/PSI/papers/06.pdf

[scribble]

Originally posted by Interesting Ian
Lifting a feather up is probably not possible. Try asking someone to influence a random number generator so that 51% of 1's and only 49% of 0's are generated, and vice versa, for a billion 0 and 1's outputted.

Now THIS is a claim we can TEST.

I'll let you pick all the components. You can pick the RNG we're going to use, you can pick the seed value, and you can pick how many iterations to perform before you begin to influence it.

If you alter the sequence by so much as a SINGLE BIT, I'll mail you $20. I'd offer you more, but I'm not good for it. We'll need to have some way to verify which numbers were output, of course, and check that against the sequence with an identical seed... can you get a witness I'd trust?

[Interesting Ian]

Originally posted by steenkh
[B]Indeed. If it was possible psychically to alter the output of a pseudo-random number generator,

I said a random number generator, not a pseudo-random number generator.

[Interesting Ian]

Originally posted by scribble
Now THIS is a claim we can TEST.

I'll let you pick all the components. You can pick the RNG we're going to use, you can pick the seed value, and you can pick how many iterations to perform before you begin to influence it.

If you alter the sequence by so much as a SINGLE BIT, I'll mail you $20. I'd offer you more, but I'm not good for it. We'll need to have some way to verify which numbers were output, of course, and check that against the sequence with an identical seed... can you get a witness I'd trust?

I never said I could do it. I don't know if I could or not. Probably not under conditions where one feels obliged to perform.

[Ersby]

Originally posted by davidsmith73

Are the sceptics really winning?


Okay, I admit it! I wrote the title because I was tired of people posting a handful of experiments and demanding that all skeptics are delusional/afraid/part of some anti-psi experiment.

So when I saw 2004's paltry crop of positive experiments, I posted this thread.

Even you have to admit, the negative outweighs the positive in these papers.

[davidsmith73]

Originally posted by Ersby
Okay, I admit it! I wrote the title because I was tired of people posting a handful of experiments and demanding that all skeptics are delusional/afraid/part of some anti-psi experiment.

So when I saw 2004's paltry crop of positive experiments, I posted this thread.

Even you have to admit, the negative outweighs the positive in these papers.

I'll admit that the ganzfeld experiments presented are very disappointing. I also sympathise with you. Uncritical acceptance of psi is not a good thing.

The experiments that use physiological correlates as the primary measure show much better results (although the Broughton paper failed). The Kittenis paper is the most impressive results section I've seen yet. I wonder if anyone here can find methodological problems with it? I would also like to see a meta-analysis of these kinds of experiment, if they can be judged similar enough that is. I have the feeling that the number of positive results far outweigh the negative in this case.

[scribble]

Originally posted by Interesting Ian
I never said I could do it. I don't know if I could or not. Probably not under conditions where one feels obliged to perform.

I assume by our age, we've all had to struggle with impotence at some point. Don't worry about it too much and you'll get it back -- the trick is to not freak out about having to perform.

Or, failing that, Viagra works wonders.

[Ashles]

I never said I could do it. I don't know if I could or not. Probably not under conditions where one feels obliged to perform.
Or any other conditions.

Why start imagining you have abilities that you have never observed?

Maybe if I climb Everest the cold and thin oxygen will suddenly turn me into Superman. This is equally likely as you affecting a RNG Ian.

[TheBoyPaj]

Originally posted by CFLarsen
I just flipped a coin 206 times, and came out 116 Heads and 90 Tails. A better result than a lot of RV experiments.

Why did you stop at 206? It's an odd number to choose.

(well, not odd, but you know what I mean)

[Garrette]

Originally posted by CFLarsen:

I'd rather count the total number of sidewalk tiles on Manhattan than flip a coin for that long again....

You're in luck; I RV'd that number this morning. I tried to e-mail you the result but Ian unintentionally messed with my server and it went to Gary Schwartz instead. He's writing a book about it now.

[Garrette]

Winning? I think it's one of those things where there is no winning, only constant vigilance.

The other day at the library there was a group of three gentleman nearby having a discussion. One was leading it by espousing all the "proven" conspiracy theories about the drug companies suppressing the cure for cancer and automobile makers killing people who making cars run on tap water. Luckily, he pooh-poohed one of the other two who insisted that oxygen is not necessary for survival; it just happens to be in the air, but we don't need it. The pooh-poohing didn't work, though; he maintained his position.

I have a very dear and sweet relative who insists Uri Geller is real and ghosts are real. After all, an actual scientist studied Geller and said he was real and some scientists from Duke University studied a haunted house in Kentucky and said it was real.

I'm walking the line with my own kids. They know my stance on these things, but I keep insisting they are allowed to disagree with me. So far, they haven't disagreed openly with me and have on occasion, without my prodding, called certain claims "hooey."

I don't push them for two reasons: I want them to have the fun of their childhood, and I think their skepticism will be more firmly grounded if it is not forced upon them.

TMI. Sorry.

[flyboy217]

Originally posted by scribble
Now THIS is a claim we can TEST.

I'll let you pick all the components. You can pick the RNG we're going to use, you can pick the seed value, and you can pick how many iterations to perform before you begin to influence it.

If you alter the sequence by so much as a SINGLE BIT, I'll mail you $20. I'd offer you more, but I'm not good for it. We'll need to have some way to verify which numbers were output, of course, and check that against the sequence with an identical seed... can you get a witness I'd trust?

Erm... no. Ian said random number generator. You are proposing a pseudorandom one.

[CFLarsen]

Originally posted by TheBoyPaj
Why did you stop at 206? It's an odd number to choose.

(well, not odd, but you know what I mean)

The highest number of trials in Ersby's list.

[Ashles]

I have a very dear and sweet relative who insists Uri Geller is real
Although he does look like a Muppet I have it from reliable sources that Mr Geller is actually real.

Although nothing he actually says seems to be.

[scribble]

Originally posted by flyboy217
Erm... no. Ian said random number generator. You are proposing a pseudorandom one.

I don't know of any provably random number generators. It's all degrees of pseudorandomness. Some degrees better than others. Yes, you can purchase CDs with white noise on them, but I've yet to see a proof that such output is *random* and not just *evenly distributed*.

Anyhow, that's all irrelevant. What Ian said is equally stupid whether he meant RNG or PRNG. Even discussing the difference at him is pointless; we've been down this road with him before. I just like to pop in to see whether he's still making the same idiotic claims. And sometimes I like to comment on them.

[Joe_Black]

Tell you what, when believers stop making claims about what the Ganzfeld data 'shows' we'll stop analysing that data.
Deal?


I am classed as a believer here. I can't start making claims about ganzfeld data so how can i stop?

[Steven Howard]

Originally posted by flyboy217
Erm... no. Ian said random number generator.

But then later he said computer.

[Ashles]

Quote by Joe Black:
I am classed as a believer here. I can't start making claims about ganzfeld data so how can i stop?
I was referring to when you wrote:
You keep banging that drum Ersby.
I took that to mean that you did not want Ersby to continue to collate and analyse the Ganzfeld data because you felt he had repeatedly done so.
Was I wrong in this reading of your post?
Are you happy for him to continue to look at the data?

Also I took you to be a believer when (referring to the JREF prize) you posted:
jj the prize is bogus.
And then explained your reasoning for this statement:
The are plenty of people who can do what i, but do much better than me. So from my point of view if they have not won it, It looks too be bogus. Do you understand my reasoning?

So, to clarify, are you saying you are not a believer in paranormal abilities? Or you are but you don't want to make claims about Ganzfeld?
I am confused by your post.
No-ome is stopping you claiming whatever you like.

[Interesting Ian]

Originally posted by Steven Howard
But then later he said computer.

I made a mistake.

[Nex]

Originally posted by Interesting Ian
I made a mistake.
We know... :rolleyes:






Ian if you miss the joke in this post too, I swear by the sweet IPU I'll scream.

[Interesting Ian]

Originally posted by Interesting Ian
I made a mistake.

Yea, a computer only deals with algorithms. Not genuinely random numbers.

I made a mistake.

I am mortified.

[Ashles]

Wow, I heard Nex scream from here.:D

[Nex]

YAAAARRRGGHH!!!


OK, I feel better now.

Ashles, your timing's off. :p

[flyboy217]

Originally posted by scribble
I don't know of any provably random number generators. It's all degrees of pseudorandomness. Some degrees better than others. Yes, you can purchase CDs with white noise on them, but I've yet to see a proof that such output is *random* and not just *evenly distributed*.

Quantum measurement is truly random--a qubit in the state (1/sqrt(2)) |0>+|1> has an equal chance of being measured as |0> or |1>.

There are several sites that purport to generate truly random number generators (often from radioactive decay samples). For example:

http://www.randomnumbers.info
http://www.fourmilab.ch/hotbits/

[amherst]

Originally posted by Ersby
Thanks, I'll see if that makes a difference to my sums



Well, honestly, if anyone here has that kind of attitude about anything, I advise them to just go to bed and not come out again!

"Leave it to the experts" is exactly the kind of attitude I abhor.
I advise you, and every other skeptical non-expert, to read Ray Hyman's "Proper Criticism" (http://www.csicop.org/si/2001-07/criticism.html) article. You especially, should take note of this suggestion:
Do not go beyond your level of competence. No one, especially in our times, can credibly claim to be an expert on all subjects. Whenever possible, you should consult appropriate experts. We, understandably, are highly critical of paranormal claimants who make assertions that are obviously beyond their competence. We should be just as demanding on ourselves. A critic's worst sin is to go beyond the facts and the available evidence.
In this regard, always ask yourself if you really have something to say. Sometimes it is better to remain silent than to jump into an argument that involves aspects that are beyond your present competence. When it is appropriate, do not be afraid to say, "I don't know."

amherst

[Aussie Thinker]

Amherst,

I advise you, and every other skeptical non-expert, to read Ray Hyman's "Proper Criticism" article. You especially, should take note of this passage

NO ONE is an EXPERT in this sort of study.

When you are studying something non existent how can you be an expert ?

I am EMINENTLY qualified to tell what is BS and what isn’t.

The entire Ganzfeld gamut of “experiments� is so ridiculously cloudy and impossible to decipher that they should just be scrapped.

Here are the problems..

1. The experiments themselves.. if they were all just simple Zenner card experiments they may have some validity. They often involve “interpretation� by the experimenters.
2. They produce results that are NOT MUCH different to chance
3. We may not get ALL experiments (file drawer stuff) and the ones that are NOT offered up are LIKELY to be the failed ones.
4. Nothing “obvious� ever happens.
5. The experiments NEVER lead to any avenue of further examination

ETC ETC..

Added to this the illogical nature and the scientific impossibility of what they examine and you have a big waste of time.

In short I have NEVER in my life seen so much talked about so little by so many for so few possible gains.

[amherst]

Originally posted by Aussie Thinker
Amherst,



NO ONE is an EXPERT in this sort of study.

When you are studying something non existent how can you be an expert ?

I am EMINENTLY qualified to tell what is BS and what isn’t.
Ersby has been trying to critique the ganzfeld experiments through the use of statistics. While skeptics like Hyman and Wiseman are competent in this respect, it is clear that Ersby is not.

amherst

[Aussie Thinker]

Amherst,

In the Bigger picture.. what is it that you find as such compelling “proof� in these tests.

They (and the results) are so airy fairy and inconsequential that I wonder why you put any faith in them at all ?

[Ersby]

Originally posted by amherst
Ersby has been trying to critique the ganzfeld experiments through the use of statistics. While skeptics like Hyman and Wiseman are competent in this respect, it is clear that Ersby is not.

amherst

Whereas you are? It's difficult to talk about the ganzfeld without the use of statistics. If I cannot critique them, can you support them?

It is true I cannot be considered competent in statistics, but I can look at the figures for myself and as long as I remember I have no training in this field, I can keep myself on an even keel and not get carried away. You and Interesting Ian are right to take what I say with a pinch of salt, but to simply insist that I shouldn't be dabbling in the field at all is pretty desperate stuff. If what I'm saying is so dumb that it can be easily debunked, then step right up and do it. I would much rather fail by my own words and my own thinking than constantly relying on quoting what other people said.

Actually, I took another look at the z-scores of the "Victim of its own success" paper, and found when sorted for statistical significance it put the experiments in a different order to mine. This is because part of their calculations involves dividing by the square root of the number of trails. This always struck me as a rather arbitrary way of accounting for an experiment's size, especially since it gives an unrealistically high (or low) effect size to small experiments.

As a neat example, can anyone explain why an experiment of 10 trials and one hit (ie, one hit less than expected by chance) should have a more extreme z-score that an experiment with 51 trials and 19 hits? Makes no sense to me.

[dharlow]

Originally posted by Ersby
As a neat example, can anyone explain why an experiment of 10 trials and one hit (ie, one hit less than expected by chance) should have a more extreme z-score that an experiment with 51 trials and 19 hits? Makes no sense to me.

I'd have to look at the actual paper, but you're right, it seems strange. In order to reach a z of -2 with only ten trials, there must have been some deviance from the standard distribution used in the Ganzfeld. Perhaps the paper's hit rate for the study (which they estimated) is not correct. In any event, this may be in part why the study rated low on the "standardness" judgment.

[Ersby]

Furthermore, the effect size (not effect size r, which I use)reported for the ten trial experiment is more extreme than the effect size for Dalton's 128 trial experiment!

That's simply counter intuitive. If these numbers are supposed to be reporting the size of an effect, then there's something seriously back to front in the figures.

[Ersby]

By the way, those aren't the only examples of strange effect sizes thrown up by the paper. (Just in case anyone thinks I'm taking one single mistake and turning it into a major issue.)

[Interesting Ian]

Originally posted by Aussie Thinker


When you are studying something non existent how can you be an expert ?



Very easily. Let's suppose you're right, and that no paranormal phenomena exists whatsoever. It is still perfectly possible to have a deep understanding of all the research that's been done.

[amherst]

Originally posted by Ersby
It is true I cannot be considered competent in statistics, but I can look at the figures for myself and as long as I remember I have no training in this field, I can keep myself on an even keel and not get carried away. You and Interesting Ian are right to take what I say with a pinch of salt, but to simply insist that I shouldn't be dabbling in the field at all is pretty desperate stuff. If what I'm saying is so dumb that it can be easily debunked, then step right up and do it.
Before your errors were just recently pointed out to you, you believed that if a phenomenon didn't have a large effect size, this meant it wasn't real. You also had no idea what a z score was. How you think you can statistically prove anything about the ganzfeld experiments, when you're this incompetent, is beyond my comprehension.
I would much rather fail by my own words and my own thinking than constantly relying on quoting what other people said.
The only reason why you've been trying to come up with your own statistical arguments is that you realize that the sophisticated ganzfeld skeptics (i.e., Hyman and Wiseman) have themselves failed to discredit the results. So you, without knowing a thing about what you're doing, do your own nonsensical "analyses" and then come in here and proclaim that the believers "argument has been lost." If this isn't "pretty desperate stuff," I don't know what is.

amherst

[a_unique_person]

You forgot 'losing big time re Global Warming'.

source (http://www.theage.com.au/articles/2004/11/26/1101219743320.html)



Most scientists say that global warming is not only real, but is already contributing to extreme droughts, floods and the melting of the polar ice caps. But a few scientists still insist the idea is bunk. With the Kyoto Protocol about to come into force, Melissa Fyfe investigates the doubters, their financial backers and whether they are worth listening to.

At 401 Collins Street on Monday night, 50 men gathered in a room of plush green carpet, pottery and antique lights to launch a book about the science of climate change. Some of them were scientists. But many were engineers and retired captains of industry. Presiding was Hugh Morgan, president of the Business Council of Australia and former Western Mining boss. The master of ceremonies was retired Labor politician Peter Walsh.

Climate change is about science, but not just about science. It's about business and politics and wielding influence. The men - there was just one woman present - were all climate change sceptics, members of an organisation called the Lavoisier Group that argues global warming is nothing to worry about.

The book they launched - the latest weapon in the tussle for hearts and minds over global warming - was by Melbourne climate change sceptic William Kininmonth, former head of the National Climate Centre, part of the Bureau of Meteorology. He argues that global warming is natural and not caused by humans burning fossil fuels.

The book, Climate Change: A Natural Hazard, blasts the models used by climate scientists to predict and simulate what is happening. They are flawed, he says. "Climate change is naturally variable and it poses serious hazards for human kind," he writes. Focusing on man-made global warming is "self-delusion on a grand scale".

The only problem for the sceptics is that the vast majority of scientists think they are the ones that are deluded. "There's a better scientific consensus on this than on any issue I know - except maybe Newton's second law of dynamics", Dr James Baker, of the National Oceanic and Atmospheric Administration in the US, has said.



The scientific process is well regarded in most areas here, but when you talk about GW, it seems that everyone doubts that scientists are correct in their use of science.

What it really comes down to is politicis.



Scepticism, of course, is a hallmark of science. Some global-warming critics are simply suspicious about the idea of consensus in the scientific community. As the Lavoisier Group's Ray Evans points out, the history of science is littered with hard-won battles of one man - such as Galileo - against a flat earth-like consensus. Evans also says he is a "Genesis 1:28 man". That's the passage that says: "God said to them 'Be fruitful and become many and fill the Earth and subdue it, and have in subjection the fish of the seas and the flying creatures of the heavens and every living creature that is moving upon the Earth".

The global-warming doomsayers, says Evans, are anti-development. Moreover, they stem from an environmentalism that has taken the place of Christianity, particularly in Europe. "To put it in its bluntest terms, when you don't believe in God you don't believe in nothing. You believe in whatever is the fashion of the day, and environmentalism has scooped the pool."



The irony escapes this self deluded twit.

Plenty of people believed the earth was round, it was politically incorrect to say it was. Galileo was just the guy who was brave enough to say so.

Now this idiot is using god to justify the quashing of science again.

[Ersby]

Originally posted by amherst

Before your errors were just recently pointed out to you, you believed that if a phenomenon didn't have a large effect size, this meant it wasn't real. You also had no idea what a z score was. How you think you can statistically prove anything about the ganzfeld experiments, when you're this incompetent, is beyond my comprehension.


amherst [/B]

*sigh*

Where my errors pointed out to me, and which errors are you talking about? I neve believed that about effect sizes.

Oh, and I do know what a z-score is. If you read this thread again, you'll see I asked for what level of effect size r corresponded to p=0.05. A bit different, you see.

Your second paragraph is just wish-fulfilment on your part.

Could you explain, while your here, my problems with the z-scores and effect sizes of the "Victim of its own success paper"?