Parapsychology & Psychical Research

Discussions on UFOs, ghosts, myths etc.

Moderators: Calilasseia, DarthHelmet86, Onyx8

Re: Parapsychology & Psychical Research

#21  Postby NucleicAcid » Apr 02, 2010 3:10 pm

See I think that is another commonly held misconception - that the only way to show the existence of psi is rejecting the null hypothesis. Aside from the silliness of the tautology of this statement (really, all of science is rejecting the null hypothesis), it makes the very large assumption that psi is just any effect that is leftover after all other variables are locked down. If you look at the history of psi research, you'll quickly see that this isn't at all the case. Psi is a quirky phenomena, but there are certain aspects of it that are predictable, and this leads to testable hypotheses. You find evidence to support those hypotheses, and you build evidence for psi. Science.

You're confusing the sheep-goat effect with what I call the Bad Joujou effect.

The Sheep-Goat Effect is a specific effect which, at its simplest, people who believe in psi or are psi-positive score better on psi tasks than people who do not believe in psi. If you ask someone to perform a psychic task, and they believe psychic powers are hogwash, they're not going to succeed at the task. It's no different than playing pickup basketball, and someone picks the kid in gym that isn't all that athletic, and he says, "I'm really not that good at basketball," and he ends up failing miserably, as almost a self-fulfilling prophecy. Perhaps he would have done better if he actually tried, but the lack of confidence definitely impedes on his will to do well and play with skill.

Specifically, the Sheep-Goat Effect predicts a positive correlation between belief in psi, and standardized score on a psi-task such as ESP. If guessing were completely random, we would expect there to be no relationship between belief and performance (alas, if psi didn't exist, there would be no relationships between any variables and performance for a sufficiently large experiment).

If it's a legitimate effect, there's no reason why it would be invalid to test for it, no matter how absurd it sounds. Science is built on testable hypotheses. SGE is a testable hypothesis, has been tested, and showed to have a large body of evidence for it.

The Bad Joujou effect is basically a type of experimenter influence in which negative results are "explained away" because of some essentially unknown confounding variable, stereotypically "bad vibes" or the like. There are two sides to this coin. One, there are some cases in which there genuinely is a third-party variable that hasn't been accounted for, and this is responsible for the difference in effect. But the other half is over-reliance on this cop-out, which I think most parapsychologists agree is copping out.

If you have an extremely well built study, and you control for all the psi-conducive variables (or as many as is reasonable), and you get chance results, that would be evidence against psi. Put enough of these together over time and you'll start seeing the ranks of parapsychologists thin out (you'll always have die-hard believers, but the ones that are basing their belief entirely on the data would see that perhaps it was just an error and most likely leave).

But it seems like skeptics do the exact same thing! Any study that is above chance MUST have some sort of flaw in it.

Let's put it this way. IF psychic functioning existed (and I'm talking about real-world effect sizes, e.g. there are definite limits on how much and what kind of information can be transmitted, not superpower mind-reading), how would you go about testing for it?
"It is far better to grasp the Universe as it really is than to persist in delusion, however satisfying and reassuring. "

"Imagination will often carry us to worlds that never were. But without it we go nowhere."

-Carl Sagan
NucleicAcid
 
Posts: 24

Print view this post

Ads by Google


Re: Parapsychology & Psychical Research

#22  Postby Mr.Samsa » Apr 03, 2010 3:15 am

NucleicAcid wrote:
If you have an extremely well built study, and you control for all the psi-conducive variables (or as many as is reasonable), and you get chance results, that would be evidence against psi. Put enough of these together over time and you'll start seeing the ranks of parapsychologists thin out (you'll always have die-hard believers, but the ones that are basing their belief entirely on the data would see that perhaps it was just an error and most likely leave).

But it seems like skeptics do the exact same thing! Any study that is above chance MUST have some sort of flaw in it.


Of course, that's how science is done. Basically assume something is false and is the result of bad methodology/confounds until you have a solid definition of the phenomenon that is falsifiable, a valid mechanism it can work by, and a mountain of research that shows the predictions made by this falsifiable definition are not refuted by empirical research. As far as I can see, psi researchers have none of the above - they only some studies that are arguably 'above chance'. Following the general process of science, the strongest position we can hold here is to argue that this is a statistical anomaly that needs to be explained. This is the most parsimonious explanation as it doesn't require us to posit the existence of psi and we have seen quirky statistical effects occur before. Once we've ruled out statistical effects, then we can look at what the results say to see whether we can posit psi as a possibility or whether we need to rule something else out first.

In other words, when you hear the sound of hoof beats, think horses not zebras. I think this is why a lot of parapsychologists get a bad reputation because on the surface they might actually be discovering something interesting about the world - but they are too busy trying to shove their beliefs into the gap to actually take a step back, do some real science, and discover what's causing this weird effect. If psi is true, then they're going to discover it anyway so why not just go through the motions?

NucleicAcid wrote:Let's put it this way. IF psychic functioning existed (and I'm talking about real-world effect sizes, e.g. there are definite limits on how much and what kind of information can be transmitted, not superpower mind-reading), how would you go about testing for it?


Like I said, we'd need a falsifiable definition first. Most research I've read in the area have left this fairly vague, which makes it easy to either over-interpret your results as evidence, or to use the ambiguities as a scapegoat. Personally, if I was investigating something like telepathy (and assuming that psi-ability is something innate to humans rather than a special ability or learnt talent etc) then I would define it as something like, "observing the contents of another person's "mind" through no external cues". There would probably need to be some tweaking here, such as how accurate they need to be, what's meant by "mind" etc, but I think it would be an alright start.

Then I'd get around 100 participants and randomly divide them into two groups of 50, and randomly assign them partners from the other group (without them knowing or ever meeting the person). I'd sit Participant A in Room 1, and Participant B in Room 2 and then present A with a series of 10 randomly chosen objects and shapes. They would be fairly simple geometric shapes such as squares and triangles etc so they'd be easy to "transmit and receive". No experimenter would be present in the room at all. When the trial starts, a green light will turn on in both rooms and Participant A will have the first image appear on the screen. After a certain period of time has elapsed, say 30 seconds, the lights will turn orange and Participant B will be required to record what he saw in his "mind" - this will be a one word answer such as "square" or "triangle". Any answer that is longer than one word will be recorded as incorrect. Once he has submitted his answer into the computer, he is prevented from changing it and the lights will turn red to indicate the end of a trial. Then the process is repeated through 10 trials and then the session ends

If participants truly are able to pick up images and sensations from complex video images, then recording a triangle should be easy and we should predict pretty accurate results - say, at least 2-3 correct responses for every group of participants? If we don't see this, then we can take it as fairly strong evidence for the idea that the participants in the Ganzfeld tests are simply exploiting a statistical anomaly.

Out of interest, do you know of any Ganzfeld experiments that use more than 4 samples for the receiver to choose from? Preferably one that uses 20 or more? That would limit their chances of getting a false positive.
Image
Mr.Samsa
 
Posts: 11370
Age: 35

Print view this post

Re: Parapsychology & Psychical Research

#23  Postby NucleicAcid » Apr 03, 2010 2:32 pm

As far as I can see, psi researchers have none of the above - they only some studies that are arguably 'above chance'.


Curious, is this a conclusion you've come to yourself, or is this someone prominent has said? If it is the latter, I'd like to know the source, I'm collecting information for a piece that I'm writing on psi myths.

This viewpoint is really striking, because it seems to say that psi research occurs because someone was doing some test, had some above chance statistics, and the experimenters concluded it must be psychic powers.

It's quite the opposite. Many people reported having experiences such as knowing when a distant relative was in distress, or suddenly answering a question someone else was thinking. This led some scientists to wonder if humans had telepathy. They devised a bunch of experiments to try to test telepathy in the lab. Their preliminary results showed an effect. So they did more tests to confirm. And so they just kept testing and getting positive results and refining the tests so there were less threats to internal validity, and improving the tests to try to increase how much information is transmitted.

I think "Humans have telepathy" is falsifiable in its own right, the tricky part is finding the specific conditions under which it would definitely occur, so you can rate its nonexistence. I think if Forced-Choice Telepathy tasks or Dream Telepathy or Ganzfeld I or Ganzfeld II or Presentiment or Feeling Of Being Stared At, if any of those large-scale analyses had dropped big fat zero, there would be far less of a case. But they all test different aspects of the same mechanism, they all get similar results, and they all point in the same direction.

Then I'd get around 100 participants and randomly divide them into two groups of 50, and randomly assign them partners from the other group (without them knowing or ever meeting the person). I'd sit Participant A in Room 1, and Participant B in Room 2 and then present A with a series of 10 randomly chosen objects and shapes. They would be fairly simple geometric shapes such as squares and triangles etc so they'd be easy to "transmit and receive". No experimenter would be present in the room at all. When the trial starts, a green light will turn on in both rooms and Participant A will have the first image appear on the screen. After a certain period of time has elapsed, say 30 seconds, the lights will turn orange and Participant B will be required to record what he saw in his "mind" - this will be a one word answer such as "square" or "triangle". Any answer that is longer than one word will be recorded as incorrect. Once he has submitted his answer into the computer, he is prevented from changing it and the lights will turn red to indicate the end of a trial. Then the process is repeated through 10 trials and then the session ends


This is actually very similar to forced-choice ESP card study (now run on computers so that there is no chance of card counting and true randomization.) It's been done many times, and comes out above chance most of the time.

If participants truly are able to pick up images and sensations from complex video images, then recording a triangle should be easy and we should predict pretty accurate results - say, at least 2-3 correct responses for every group of participants? If we don't see this, then we can take it as fairly strong evidence for the idea that the participants in the Ganzfeld tests are simply exploiting a statistical anomaly.


Somewhat paradoxically, a triangle is a relatively difficult thing to transmit, and it all has to do with the brain.

A triangle isn't represented in the brain as "A Triangle." Shapes are very abstract. It is broken down into various components, and there is a neuronal connection for each of those, e.g. you'd probably see things start lighting up as "Slanted line to the left," "Slanted line to the right," "Horizontal line," "Three corners," "Equal lengths" etc. Then, for telepathy, somehow each of these neuronal patterns needs to "jump" from one brain to another. If only a few key bits are missing, it could be interpreted as the letter A, the letter V, a prism, the roof of a house, the cone of a rocket, a spear, a bird with spread wings, on and on. The advantage of using really intense imagery is that you get more pieces to send, and hopefully improve the chances of successfully reconstruct the image. Telepathy isn't like sending a crystal clear image. From a cognitive psychology point of view it's metaphorically like taking a picture of a scene on a crummy camera, writing a paper to describe the picture, ripping up that paper into a bunch of pieces, putting each of those pieces into a bottle, corking the bottles and throwing them into a river, having someone who speaks a different language down river skim up the bottles, reconstruct what they can of the letter, and try to draw the picture while a really loud rock concert is blasting.

It's a signal to noise ratio thing. Plus, telepathy does not occur much more often than it does occur. It's the reason entanglement experiments use coincidence counters - something like 1 in a billion emissions are actually entangled, the rest are normal. What we need is a coincidence counter for thoughts (and trust me, I've been trying to figure out how to do this for a while :-P).

The other problem found with using simple shapes almost from the get-go is that people get bored. People complain that the Rhine experiments exhibited a decline effect over time. That's because these subjects often did THOUSANDS of trials of (focus for ten minutes) "Square. (focus for ten minutes) "Circle". etc.

So therefore, theoretically the more details you have that can be transmitted, coming from all the different parts of the brain, e.g. visual, auditory, semantic, and especially emotional, as long as you have a relatively limited pool of things to pick from, the better your chance of getting a hit.


Out of interest, do you know of any Ganzfeld experiments that use more than 4 samples for the receiver to choose from? Preferably one that uses 20 or more? That would limit their chances of getting a false positive.


No, not particularly that I know of, but that sounds pretty similar to some remote viewing experiments. The GotPsi experiment addresses this somewhat: http://www.boundary.org/bi/articles/GotPsi-public.pdf
They use basically a semi-free response format (full description on page 3) so they had an objective way to count how well a hit was.
"It is far better to grasp the Universe as it really is than to persist in delusion, however satisfying and reassuring. "

"Imagination will often carry us to worlds that never were. But without it we go nowhere."

-Carl Sagan
NucleicAcid
 
Posts: 24

Print view this post

Re: Parapsychology & Psychical Research

#24  Postby Mr.Samsa » Apr 03, 2010 3:11 pm

NucleicAcid wrote:
As far as I can see, psi researchers have none of the above - they only some studies that are arguably 'above chance'.


Curious, is this a conclusion you've come to yourself, or is this someone prominent has said? If it is the latter, I'd like to know the source, I'm collecting information for a piece that I'm writing on psi myths.


Just my own opinion from reading papers. I'd be interested in reading your piece when you've finished though.

NucleicAcid wrote:This viewpoint is really striking, because it seems to say that psi research occurs because someone was doing some test, had some above chance statistics, and the experimenters concluded it must be psychic powers.

It's quite the opposite. Many people reported having experiences such as knowing when a distant relative was in distress, or suddenly answering a question someone else was thinking. This led some scientists to wonder if humans had telepathy. They devised a bunch of experiments to try to test telepathy in the lab. Their preliminary results showed an effect. So they did more tests to confirm. And so they just kept testing and getting positive results and refining the tests so there were less threats to internal validity, and improving the tests to try to increase how much information is transmitted.


Yeah but the problem is that they shouldn't be testing for telepathy - they should be testing to see whether people score above chance. If, as you say, these results come up consistently, then we need to look at possible reasons this could be the case. A faulty assumption or a statistical effect should be the first things that come to mind, and I personally don't think they've ruled either of those options out to even begin speculating psi (and they can't conclude psi until they come up with a plausible mechanism).

NucleicAcid wrote:This is actually very similar to forced-choice ESP card study (now run on computers so that there is no chance of card counting and true randomization.) It's been done many times, and comes out above chance most of the time.


Interesting, any chance of linking me to some?

NucleicAcid wrote:
If participants truly are able to pick up images and sensations from complex video images, then recording a triangle should be easy and we should predict pretty accurate results - say, at least 2-3 correct responses for every group of participants? If we don't see this, then we can take it as fairly strong evidence for the idea that the participants in the Ganzfeld tests are simply exploiting a statistical anomaly.


Somewhat paradoxically, a triangle is a relatively difficult thing to transmit, and it all has to do with the brain.

A triangle isn't represented in the brain as "A Triangle." Shapes are very abstract. It is broken down into various components, and there is a neuronal connection for each of those, e.g. you'd probably see things start lighting up as "Slanted line to the left," "Slanted line to the right," "Horizontal line," "Three corners," "Equal lengths" etc. Then, for telepathy, somehow each of these neuronal patterns needs to "jump" from one brain to another. If only a few key bits are missing, it could be interpreted as the letter A, the letter V, a prism, the roof of a house, the cone of a rocket, a spear, a bird with spread wings, on and on.


Hmm.. not quite. What you get is called a generalisation gradient. If the target stimulus is a triangle, then you will get the neural pattern corresponding to "A triangle" firing the most in the brain. Then what you'll find is that other areas will fire as well, the strength of which decreases a function of similarity. That is, the neurons corresponding to "slanted lines" will fire, but less than "triangle" but more than "horizontal line" (as there are two slanted lines and only one horizontal). The problem with your explanation of neural patterns is that it would strongly suggest that "triangle" would be a much more likely response than anything else you've suggested, although those responses may appear occasionally.

NucleicAcid wrote:The advantage of using really intense imagery is that you get more pieces to send, and hopefully improve the chances of successfully reconstruct the image. Telepathy isn't like sending a crystal clear image. From a cognitive psychology point of view it's metaphorically like taking a picture of a scene on a crummy camera, writing a paper to describe the picture, ripping up that paper into a bunch of pieces, putting each of those pieces into a bottle, corking the bottles and throwing them into a river, having someone who speaks a different language down river skim up the bottles, reconstruct what they can of the letter, and try to draw the picture while a really loud rock concert is blasting.


Sure, but in your analogy, your solution to the problem of the crummy camera/ripped up essay scenario is to film an entire movie on a camera obscura, then write a novel the size of War and Peace describing all the thematic aspects of the movie, and then burn the book to ashes, stick it in a few bottles, throw them in a river to someone who might not be able to read, and ask them to guess out of four movies which one likely corresponds to the ashes.

If we are talking about complexities of neural networks, then using these "emotive" images will only complicate the matter! If you imagine how many neurons are fired in response to a triangle, imagine how many are fired in response to a wildlife documentary or whatever.

NucleicAcid wrote:The other problem found with using simple shapes almost from the get-go is that people get bored. People complain that the Rhine experiments exhibited a decline effect over time. That's because these subjects often did THOUSANDS of trials of (focus for ten minutes) "Square. (focus for ten minutes) "Circle". etc.


Sure. But the solution here would be decreasing the amount of trials within a session and increasing the overall time to complete the study. I know that resources are thin in science, and probably more so for parapsychology, but skimping on methodology is not something that should be done by a field trying to establish itself as a credible field.

NucleicAcid wrote:So therefore, theoretically the more details you have that can be transmitted, coming from all the different parts of the brain, e.g. visual, auditory, semantic, and especially emotional, as long as you have a relatively limited pool of things to pick from, the better your chance of getting a hit.


I agree with the latter half of that sentence - as long as you have a relatively limited pool of things to pick from, the better your chances of getting a hit. Or rather, the better your chances of getting a significant result. The fact of the matter is simply that you won't get a stable rate of 25% accuracy, no matter what theoretical statistics you employ here. Sometimes it will be higher, sometimes it will be lower and often it will be significantly different. And obviously the less rigorously the methodology, the more likely it is that subjects will score a few extra hits. However, increase the pool for them to select from; say, at least 20 options, then them scoring 3-4 hits "above chance" won't give a statistically significant result.

If you could find a study that used many more sample stimuli than 4 and still got a significant result, then I will be a lot more interested in this field.

NucleicAcid wrote:

Out of interest, do you know of any Ganzfeld experiments that use more than 4 samples for the receiver to choose from? Preferably one that uses 20 or more? That would limit their chances of getting a false positive.


No, not particularly that I know of, but that sounds pretty similar to some remote viewing experiments. The GotPsi experiment addresses this somewhat: http://www.boundary.org/bi/articles/GotPsi-public.pdf
They use basically a semi-free response format (full description on page 3) so they had an objective way to count how well a hit was.


Thanks, I'll give it a read. I'm not too confident given it's a Dean Radin paper, and the methods for determining a hit seem a little worrying..
Image
Mr.Samsa
 
Posts: 11370
Age: 35

Print view this post

Re: Parapsychology & Psychical Research

#25  Postby ersby » Apr 03, 2010 7:03 pm

NucleicAcid wrote:Many people reported having experiences such as knowing when a distant relative was in distress, or suddenly answering a question someone else was thinking. This led some scientists to wonder if humans had telepathy. They devised a bunch of experiments to try to test telepathy in the lab. Their preliminary results showed an effect. So they did more tests to confirm. And so they just kept testing and getting positive results and refining the tests so there were less threats to internal validity, and improving the tests to try to increase how much information is transmitted.


Didn't parapsychology begin with spiritualism? The horrors of the Great War lead people to search for an connection to the dead, which lead to seances, which lead to testing mediums, which lead to etc etc.

I think if Forced-Choice Telepathy tasks or Dream Telepathy or Ganzfeld I or Ganzfeld II or Presentiment or Feeling Of Being Stared At, if any of those large-scale analyses had dropped big fat zero, there would be far less of a case. But they all test different aspects of the same mechanism, they all get similar results, and they all point in the same direction.


That's a big claim. Or a lot of small claims all at once. Did they all get similar results? Can you compare the zener card tests to presentiment? Early parapsychology was not about seeing if a particular effect exists, but demonstrating that it did (qv, Rhine's policy of not publishing negative results).

Also, is ESP the same as PK? You've put "feeling of being stared at" in the ESP group, but couldn't it be micro-PK? Except on a human mind, not on an RNG. DMILS (Direct Mental Influence on Living Systems) I think they call it.

Somewhat paradoxically, a triangle is a relatively difficult thing to transmit, and it all has to do with the brain.

A triangle isn't represented in the brain as "A Triangle." Shapes are very abstract. It is broken down into various components, and there is a neuronal connection for each of those, e.g. you'd probably see things start lighting up as "Slanted line to the left," "Slanted line to the right," "Horizontal line," "Three corners," "Equal lengths" etc. Then, for telepathy, somehow each of these neuronal patterns needs to "jump" from one brain to another. If only a few key bits are missing, it could be interpreted as the letter A, the letter V, a prism, the roof of a house, the cone of a rocket, a spear, a bird with spread wings, on and on.


But if the other targets are nothing like a triangle, or an A, or a V, or the roof of a house, then there's no problem.

The other problem found with using simple shapes almost from the get-go is that people get bored. People complain that the Rhine experiments exhibited a decline effect over time. That's because these subjects often did THOUSANDS of trials of (focus for ten minutes) "Square. (focus for ten minutes) "Circle". etc.


This is true. And understandable. Zener cards don't have much to do with spontaneous psi reported in ordinary life. Statistically, it makes sense to use simple shapes, but it's terribly dry.
ersby
 
Posts: 32
Male

Country: UK
Print view this post

Re: Parapsychology & Psychical Research

#26  Postby NucleicAcid » Apr 03, 2010 9:32 pm

Ersby, I know spiritualism definitely played a large role in the development of psi research, but I'm not sure if it came before or after. I'll have to check into that.

I'll also try to dig up some more modern forced-choice experiments.

If we are talking about complexities of neural networks, then using these "emotive" images will only complicate the matter! If you imagine how many neurons are fired in response to a triangle, imagine how many are fired in response to a wildlife documentary or whatever.


The thing about emotions is that instead of complicating the matter, there is reason to believe that emotions potentiate the "psi signal." The majority of spontaneous telepathy cases involved death, mortal danger, or serious injury.

If you could find a study that used many more sample stimuli than 4 and still got a significant result, then I will be a lot more interested in this field.


But doesn't decreasing the MCE mean that it takes less hits in order to reach the same level of significance?

It really don't matter what the MCE is. You are going to have a bell distribution of hits for any given test. Some are going to seem, by themselves, very significant. Others are going to be negative significant. Most are going to hover around the mean. That is the whole purpose of experimental statistics - does Data Set A come from Population X?

Parapsychologists that believe the psi hypothesis is supported say that all things considered, if you look at the whole picture, that the amount that the data set of psi experiments differs from what would be expected by chance is greater than what can be accounted for by coincidence, experimental setup, biases, filedrawer, and anything reasonably conceivable. Skeptics believe that all of those things add up to the psi effect. That's really all there is to it.

Furthermore, while the statistical analysis by parapsychologists is by and large well put together, and they do a good job of controlling for extraneous variables, skeptics have yet to come up with any solid proof that these sources of error that they claim are causing the effect, causes that effect. It seems like it's just a matter of parapsychologists continuing to apply more and more controls, and continue what they are doing, and producing decent results.
"It is far better to grasp the Universe as it really is than to persist in delusion, however satisfying and reassuring. "

"Imagination will often carry us to worlds that never were. But without it we go nowhere."

-Carl Sagan
NucleicAcid
 
Posts: 24

Print view this post

Re: Parapsychology & Psychical Research

#27  Postby jerome » Apr 04, 2010 11:05 am

NucleicAcid wrote:Ersby, I know spiritualism definitely played a large role in the development of psi research, but I'm not sure if it came before or after. I'll have to check into that.



I write extensively on the history of psi research and spiritualism. I'm doing stuff today - well spending time with friends playing boardgames - but I'll catch up tonight and comment...

j x
Yours sincerely, Jerome -- a threat to reason & science

I am an Anglican Prejudice declared - My blog: http://jerome23.wordpress.com/
User avatar
jerome
THREAD STARTER
 
Name: CJ
Posts: 2047
Age: 51
Male

Country: UK
Denmark (dk)
Print view this post

Ads by Google


Re: Parapsychology & Psychical Research

#28  Postby ersby » Apr 04, 2010 12:19 pm

NucleicAcid wrote:[...] skeptics have yet to come up with any solid proof that these sources of error that they claim are causing the effect, causes that effect. It seems like it's just a matter of parapsychologists continuing to apply more and more controls, and continue what they are doing, and producing decent results.


Well, think for a second - how would skeptics demonstrate that problem x lead to result y? Take as an example the audio-leakage problem that the PRL trials had. In this, it was found that if the white noise was turned off, it was possible to just hear the sound from the video targets in the receiver's headphones. This lead people to wonder if this was a subliminal clue under the sound of the white noise during an experimental trial. (Remember, only the dynamic trials got a statistically significant effect)

So, there are two ways to test for this, as far as I can see. Either you remove it, or introduce it to a different experimental set up, and see what results you get.

The audio-leakage problem was removed and the results didn't show any drop, so parapsychologists said that it can't have made any difference. But was towards the end of the experiments (around 75% of trials had been completed by the time this problem was discovered) and the later experiments included Study 302 which had a problem with response bias - for whatever reason the computer strongly tended to chose targets with humans (well, Bugs Bunny, actually) or water in, and people and water are what people tend to report "seeing" in the ganzfeld. So there's an issue in comparing the first 75% of trials which had one problem, with the rest of the trials which had another problem.

What about deliberately introducing the problem? In 1982 Roney-Dougal published an experiment with audio targets, whereby half of the trials included the target being played subliminally under the white noise to the receiver. The results were that both conditions (ESP and subliminal) were statistically significant with no difference between the two.

Roney-Dougal, “A Comparison of psi and subliminal perception: a confirmatory study”, Research in Parapsychology 1981, 1982
“The major result from this experiment is the lack of difference in the overall psi and subliminal hit rate. On both an objective and subjective basis, neither participant nor experimenter were able to distinguish between the two phenomena during or after the session. A possible explanation for this is that within a free-response experiment of this type, the way in which we become aware of the target is through primary-process mentation: dreamlike imagery, associative thought patterns, and "feelings." This mentation then needs to be "judged" at the conscious level in order to associate it with the target. In the design of this experiment, this process was identical for both phenomena, and the similarity of hit rates indicates that the subliminal information did not emerge into the primary-process”


So there's one experiment showing that subliminally introducing the target appears to inflate the hit rate beyond chance. What does this entail for the PRL trials? Has a link between audio-leakage and success been established?
ersby
 
Posts: 32
Male

Country: UK
Print view this post

Re: Parapsychology & Psychical Research

#29  Postby NucleicAcid » Apr 05, 2010 2:44 am

That's good, I'm glad that people have made an effort to try to line up any variance with potential extraneous variables. So we definitely know that subliminal influence can transmit information, but I think everyone suspected that all along. It also makes sense that if extant, psi would be indistinguishable from other knowledge sourced from the subliminal- since it is believed nearly universally by parapsychologists that psi must percolate up from the subconscious before it is consciously perceived and processed.

The question is, was it a factor during the PRL experiments? I doubt anyone knows at this point, but I believe it wasn't - as the receiver spent most of their time in the Ganzfeld, and this in theory would block out any noise that would have made it through the soundproofing.

Wasn't one of the major changes after this was discovered to remove any sound from the dynamic targets? I think all of the Autoganzfeld trials conducted had this feature.

So like you said, we throw out the PRL studies, but the effect persists. It doesn't mean that psi is now definitely confirmed, but it definitely bolsters the case. That's at least one less variable that could be accounting for the effect. My question is, if it is an extraneous variable, what could we be missing?
"It is far better to grasp the Universe as it really is than to persist in delusion, however satisfying and reassuring. "

"Imagination will often carry us to worlds that never were. But without it we go nowhere."

-Carl Sagan
NucleicAcid
 
Posts: 24

Print view this post

Re: Parapsychology & Psychical Research

#30  Postby ersby » Apr 05, 2010 7:48 am

NucleicAcid wrote:The question is, was it a factor during the PRL experiments? I doubt anyone knows at this point, but I believe it wasn't - as the receiver spent most of their time in the Ganzfeld, and this in theory would block out any noise that would have made it through the soundproofing.


I just realised I may have chosen a mis-leading term when I said "audio-leakage". Wiseman wrote a lengthy paper on the soundproofing at the PRL trials and used the term "sound-leakage". I wasn't referring to that. I'm talking about the technical problem discovered by George Hansen that existed in the equipment itself. So the problem I'm talking about wasn't linked to the soundproofing and (potentially) occured while the receiver was in the ganzfeld. Apololigies for any confusion.

Wasn't one of the major changes after this was discovered to remove any sound from the dynamic targets? I think all of the Autoganzfeld trials conducted had this feature.


I don't believe that's the case, no. [edited to add: I just checked. They modified the equipment so the leakage no longer happened. The targets themselves weren't changed.]

So like you said, we throw out the PRL studies, but the effect persists.


One argument that's never convinced me is to prove a variable didn't have an effect by removing a set of experiments from the database and say because the overall effect is still there, the variable can't have had an impact. For a start, the database is large enough that removing a handful of any experiments wouldn't make a difference. Secondly, it's a very blunt instrument, unable to take into account clear (to my mind) methodological flaws.

Take for example one experiment in which the person who chose the target was the same person who took notes of the receiver's mentation used for judging at the end of the session. To my mind, this is clearly a flaw. But its results were about the same as the other ganzfeld experiments. Does this mean the flaw isn't important?
ersby
 
Posts: 32
Male

Country: UK
Print view this post

Re: Parapsychology & Psychical Research

#31  Postby NucleicAcid » Apr 06, 2010 8:50 pm

All flaws are important. And it is a blunt instrument, but an instrument nonetheless, that you take away the potential sources of error (either before or through post-hoc weeding) and you still see an effect. It's not valid for confirming the existence of psi to the world, but it bolsters the case that we're not all crazy confirmation-biasing anomaly chasers. We're at least marginally less crazy.
"It is far better to grasp the Universe as it really is than to persist in delusion, however satisfying and reassuring. "

"Imagination will often carry us to worlds that never were. But without it we go nowhere."

-Carl Sagan
NucleicAcid
 
Posts: 24

Print view this post

Re: Parapsychology & Psychical Research

#32  Postby ersby » Apr 13, 2010 10:27 pm

NucleicAcid wrote:All flaws are important. And it is a blunt instrument, but an instrument nonetheless, that you take away the potential sources of error


But you don't, though. Look at my example again. An experiment that is flawed (to my mind, fatally flawed) would remain in a ganzfeld meta-analysis by this method.
ersby
 
Posts: 32
Male

Country: UK
Print view this post

Re: Parapsychology & Psychical Research

#33  Postby ersby » May 07, 2010 9:05 am

It looks like we won’t get an answer from Nucleic Acid on this, but I recently posted the results from a meta-analysis using criteria of standardness and methodological rigour on another forum, so I thought I’d pop them on here too. The criteria were:

No unpublished papers, media demonstrations, or classroom demonstrations
No experiments labelled pilot, informal or exploratory
No papers without pre-defined numbers of trials
No unfinished experiments (one exception – Smith, Savva’s 2008 experiment fell just ten short of its target of 128 trials. It felt churlish to exclude it)
Sargent’s work at Cambridge excluded due to security issues
No experiments that allow contact between someone knowledgeable about the target and the person judging the session
Only experiments that used random number generators or random number tables to chose the target set and target
Only experiments that had duplicate target sets for judging purposes
Only experiments that used red light as the visual stimulus
Only experiments that used white noise or pink noise (ie, static) as the audio stimulus
No experiments of very short duration (15 minutes or under) or very long (45 minutes or longer)
Only visual targets
Only one target per session
Only sessions with a sender
Only sessions judged by the receiver


Below, I’ve included the barest data – just number of trials and number of hits – so that if I’ve made any mistakes in my formulae for p numbers z-scores etc, they don’t get replicated. There’s one in there which didn’t use the usual 1 in 4 direct hit scoring method, so I’ve calculated what hit rate it would need to give the same p number as reported in the paper (hence the odd looking 6.5 hits).

I don’t know how to do tables on this forum, so cut and paste this into Word. Select the drop-down menu “Table” and then “Convert” and then “Text to table”. It should suggest 5 columns, which is right, and then put everything into a table. This can then be Pasted into any spreadsheet.

Author,title,year,trails,hits
Delanoy,The training of psi in the ganzfeld,1982,72,18
Bierman,The effect of GF stimulation and Feedback (series 1),1983,32,11
Sondow,Exploring hypnotisability,1987,60,12
Delanoy,An examination of subject and agent mentation in the ganzfeld ,1988,40,10
Murre ,A GF psi experiment with control condition,1988,41,13
Bem Honorton,PRL Novice 1,1989,50,12
Bem Honorton,PRL Novice 2,1989,50,18
Bem Honorton,PRL Novice 3,1989,50,15
Bem Honorton,PRL Novice 4,1989,50,18
Bem Honorton,Study 301,1989,50,15
Bierman ,Anomolous information access in the GF I ,1993,50,13
Bierman , Anomolous information access in the GF II,1993,50,12
Morris et al,Towards replication and extension of autoganzfeld results Study 2,1993,32,13
Morris et al,Towards replication and extension of autoganzfeld results Study 1,1993,32,8
Bosga Gerding Wezelman,Target affinity,1994,50,10
Kanthamani et al,series 1,1994,30,6
Kanthamani et al,series 3,1994,40,8
Kanthamani et al,series 6,1994,60,18
McDonough et al,EEG in a ganzfeld psi task,1994,20,6
Williams Roe Upchurch Lawrence,Senders and Geomagnetism in the Auto-Ganzfeld,1994,30,5
Morris Dalton Delanoy Watt,Comparison of the sender/no sender condition in the ganzfeld,1995,64,20
Bierman,Series 3,1995,40,16
Bierman,Series 4,1995,68,18
Broughton Alexander,first timers I,1997,50,12
Broughton Alexander,first timers II,1997,50,9
Broughton Alexander,emotionally close series,1997,51,19
Dalton,Exploring the links,1997,128,60
Parker et al,study 3,1997,30,11
Wezelman Bierman,series 5 non cannabis,1997,10,1
Wezelman Bierman,series 6 non cannabis,1997,10,4
Du Tullia Gimeno,Esperimento di Ganzfeld,1997,60,12
Alexander Broughton,CL-1 ganzfeld study: a look at brain hemisphere differences,1999,50,18
Simmonds,Sender personality and Psi performance in the ganzfeld and a waking ESP control,2002,26,6.5
Parra Villaneuva,Personality factors and ESP,2003,30,10
Roe Holt Simmonds,Considering the sender as a PK agent in ganzfeld esp studies,2003,40,14
Morris Summers Yim,Evidence of anomalous information transfer with a creative population in ganzfeld stimulation,2003,40,15
Roe Sherwood Holt Wilson,Interpersonal psi: Exploring the role of the experimenter in Ganzfeld GESP tasks,2003,38,8
Stevens,Testing a model for dyadic esp,2003,50,12
Parra Villaneuva, Personality factor and psi-ganzfeld session, 2003,138,57
da Silva Pilato Hiraoka,Ganzfeld vs. No Ganzfeld: An Exploratory Study of the Effects of Ganzfeld Conditions on ESP,2003,54,18
Roe Holt,A Further consideration of the sender as a PK agent,2004,23,6
Fox Simmonds,Ruling Out Items in the judging set,2004,32,10
Lau,A Bayesian approach to the ganzfeld procedure series 1,2004,20,9
Lau, series 2,2004,20,8
Lau, series 3,2004,20,4
Lau, series 4,2004,20,5
Lau, series 5,2004,20,6
Lau, series 6,2004,20,4
Howard Delgado, series 7,2005,20,7
Howard Delgado, series 8,2005,20,8
Howard Delgado, Finding and Correcting, 2005,54,7
Smith Savva, Experimenter Effects in the Ganzfeld, 2007,114,39

Putting this through the meta-analysis software I use, I find that this database gives an unweighted z-score of 3.316, p=0.0003604, weighted z-score of 5.3399, p=0.0000001 (odds of one million to one). But the bulk of this relies on two outliers (Dalton 1997 and Parra, Villanueva 2003) which leaves the database heterogenous (Chi-square=96.2848, df=51, p=0.000131). Remove those so that the database becomes homogenous (chi-square=57.3975, df=49, p=0.191973) and the unweighted z-score becomes 2.1009, p=0.0178, weighted z-score=2.2215, p=0.01315 (odds of 1 in 71).
ersby
 
Posts: 32
Male

Country: UK
Print view this post

Re: Parapsychology & Psychical Research

#34  Postby jerome » Jul 07, 2010 7:01 pm

Thanks Ersby! :) Will you be updating your masterful seven part study of the Ganzfeld soon?
Yours sincerely, Jerome -- a threat to reason & science

I am an Anglican Prejudice declared - My blog: http://jerome23.wordpress.com/
User avatar
jerome
THREAD STARTER
 
Name: CJ
Posts: 2047
Age: 51
Male

Country: UK
Denmark (dk)
Print view this post

Re: Parapsychology & Psychical Research

#35  Postby ersby » Jul 12, 2010 6:06 am

I've no plans to, no. Do you know what happening about the results from Coventry?
ersby
 
Posts: 32
Male

Country: UK
Print view this post

Ads by Google


Re: Parapsychology & Psychical Research

#36  Postby jerome » Jul 12, 2010 11:06 am

My girlfriend is seeing her supervisor on thursday I think. I'll ask her to follow it up, and can email you direct any thing I get back

j x
Yours sincerely, Jerome -- a threat to reason & science

I am an Anglican Prejudice declared - My blog: http://jerome23.wordpress.com/
User avatar
jerome
THREAD STARTER
 
Name: CJ
Posts: 2047
Age: 51
Male

Country: UK
Denmark (dk)
Print view this post

Re: Parapsychology & Psychical Research

#37  Postby jerome » Jul 12, 2010 1:39 pm

BTW, if anyone is reading this thread, I have posted this many times before but I have a great admiration for Ersby one of the most dedicated and intelligent critics or thinkers addressing these issues, and his History of Psi in the Ganzfeld can be found here and is definitely worth reading in full -- http://skepticreport.com/sr/?p=316

j x
Yours sincerely, Jerome -- a threat to reason & science

I am an Anglican Prejudice declared - My blog: http://jerome23.wordpress.com/
User avatar
jerome
THREAD STARTER
 
Name: CJ
Posts: 2047
Age: 51
Male

Country: UK
Denmark (dk)
Print view this post

Re: Parapsychology & Psychical Research

#38  Postby jerome » Feb 23, 2011 7:54 pm

The decline effect is not just for us parapsychologists it seems -- http://www.nature.com/news/2011/110223/ ... 0437a.html

Also a recent wired article http://www.wired.com/wiredscience/2010/ ... ne-effect/

Also the rather useless Skeptics Dictionary entry - http://www.skepdic.com/declineeffect.html

j x
Yours sincerely, Jerome -- a threat to reason & science

I am an Anglican Prejudice declared - My blog: http://jerome23.wordpress.com/
User avatar
jerome
THREAD STARTER
 
Name: CJ
Posts: 2047
Age: 51
Male

Country: UK
Denmark (dk)
Print view this post

Previous

Return to Paranormal & Supernatural

Who is online

Users viewing this topic: No registered users and 1 guest