Put simply, some of the advocates of psi are taking the position that if a clear, specific explanation of the finding cannot be demonstrated, then the conclusion must be that psi is the correct explanation. That does not seem to me to be a valid conclusion. (And, to be clear, I am aware that there are many investigators who are not coming to such a conclusion.)

As pointed out elsewhere... psi refers to a phenomena, not to an explanation at this point. In testing for this phenomena, they take into account all known alternative possibilities for information transfer, and control for them. What one would expect would be left is merely noise or chance. When the results are statistically significant, this means that they aren't able to be accounted for due to noise or chance. This is what psi refers to.

Specifically, in the video excerpt that CharlieM recommended to us, Dean Radin in effect says that psi is real, game over, case closed, and it is only the dogmatic close mindedness of the scientific establishment the prevents this reality from being acknowledged.

I've discussed psi research elsewhere, and the same arguments come up over and over again (except for the exceptionally dumb ones, which are always a unique treat). The first part is usually composed of denying that any successful replications exist, which simply shows they haven't gone over the research, or else they would realize this isn't true. The next part is usually composed of pointing out the obvious flaws in the first ganzfeld experiments. Though these tend to be valid criticisms against those studies, but don't acknowledge the better controls and procedures that have been implemented since then, which still produce the same statistically significant results.

Then it usually goes, "the results aren't statistically significant enough -for me-", and/or "I bet there's some flaw in the controls that neither I nor anyone else can point out, but it must be there, because that's what I believe."

If the experience of psi researchers is anything like what I've personally seen, they've probably had this identical conversation hundreds of times. When everything is laid out, we are left with the question, "Why did all these researchers get statistically significant results, and why didn't all these researchers?"

This is where its useful to continue on to the next part of exploring this type of anomalous phenomena... namely, trying to come up with a testable explanation that is consistent with everything else we know about reality. I don't see any other way to get the bottom of the that question, other than finding some way to objectively determine what was different between the successful/unsuccessful replications, when all groups appeared to follow the same experimental protocol.

If the results of the Ganzfeld were more robust, for instance if what they found was 70% of subjects were able to recount in minute detail the sequence of events in the video viewed by the sender, without being given a multiple choice question to elevate the likelihood of a positive finding, then that would demand an explanation of some sort. As it is, I find the result distinctly underwhelming. The fact that there is a result that exceeds what is expected by chance by a mere 7% or less may indicate nothing more than that it is difficult to perform an experimental procedure that is 100% "clean" in neuropsychological research. Which would hardly be an earth-shattering finding.

The effect is quite small. It's so small that saying "this is so tiny that it lacks any sort of useful application, and therefore is a waste of time" is a fairly reasonable stance. That, however, is fundamentally different than saying "psi research is pseudoscience".

What your suggesting was a blatant flaw in the original experiments, that has since been corrected, by using the multiple choice system. "minute detail" is highly subjective, and anyone can disagree with what constitutes "detailed enough to count as a hit".

And finally, again, the controls used to create a "clean" experiment are very simple. Creating two soundproof and electromagnetically shielded rooms some 30-odd meters apart isn't groundbreaking stuff, many facilities already have these types of rooms for other kinds of particularly sensitive experiments.

Vinncent wrote:What one would expect would be left is merely noise or chance. When the results are statistically significant, this means that they aren't able to be accounted for due to noise or chance. This is what psi refers to.

That's a misunderstanding of statistical significance. You posted a link to wiki earlier, yourself. Perhaps that was too technical?
For the ganzfeld experiments it means simply that the chance of getting as many or more hits as you got is less than 1 in 20 if the chance to get a hit in a single trial is 1 in 4.
IOW, it is based on how often you get a certain result if nothing is going on.
By itself, it tells you nothing about whether it can be accounted by chance or anything else.

If the experience of psi researchers is anything like what I've personally seen, they've probably had this identical conversation hundreds of times. When everything is laid out, we are left with the question, "Why did all these researchers get statistically significant results, and why didn't all these researchers?"

A fairly simple answer is mistakes and errors. That's what pretty much any competent scientist would assume faced with such results.

Vinncent wrote:I've discussed psi research elsewhere, and the same arguments come up over and over again (except for the exceptionally dumb ones, which are always a unique treat). The first part is usually composed of denying that any successful replications exist, which simply shows they haven't gone over the research, or else they would realize this isn't true. The next part is usually composed of pointing out the obvious flaws in the first ganzfeld experiments. Though these tend to be valid criticisms against those studies, but don't acknowledge the better controls and procedures that have been implemented since then, which still produce the same statistically significant results.

Then it usually goes, "the results aren't statistically significant enough -for me-", and/or "I bet there's some flaw in the controls that neither I nor anyone else can point out, but it must be there, because that's what I believe."

If the experience of psi researchers is anything like what I've personally seen, they've probably had this identical conversation hundreds of times. When everything is laid out, we are left with the question, "Why did all these researchers get statistically significant results, and why didn't all these researchers?"

This is where its useful to continue on to the next part of exploring this type of anomalous phenomena... namely, trying to come up with a testable explanation that is consistent with everything else we know about reality. I don't see any other way to get the bottom of the that question, other than finding some way to objectively determine what was different between the successful/unsuccessful replications, when all groups appeared to follow the same experimental protocol.

I don't really disagree with any of those points. As I said, my disagreement is mainly with those who say the issue is settled and telepathy (or whatever) exists.

The effect is quite small. It's so small that saying "this is so tiny that it lacks any sort of useful application, and therefore is a waste of time" is a fairly reasonable stance.

Which is pretty much my stance. As I think I've said before, if a drug trial showed a medication achieved a 32% response, vs. 25% for placebo, even if that result was statistically significant, it would still mean nothing of practical value. I doubt many researchers would want to invest time and resources to investigate it further.

That, however, is fundamentally different than saying "psi research is pseudoscience".

Which I haven't said. To my mind, forming a testable hypothesis and then testing it constitutes science, no matter how outlandish the hypothesis may be.

What your suggesting was a blatant flaw in the original experiments, that has since been corrected, by using the multiple choice system. "minute detail" is highly subjective, and anyone can disagree with what constitutes "detailed enough to count as a hit".

But they have deliberately chosen a system where a hit must happen at least 25% of the time, just by chance. Which is vastly inflated compared to what we would expect under "normal" circumstances. (What are the odds of two random individuals thinking about the exact same thing at same time?)

If senders were able to consistently "transmit" a list of 300 random words with perfect accuracy, I don't think anyone would doubt that was a very significant and startling result.

And finally, again, the controls used to create a "clean" experiment are very simple. Creating two soundproof and electromagnetically shielded rooms some 30-odd meters apart isn't groundbreaking stuff, many facilities already have these types of rooms for other kinds of particularly sensitive experiments.

I'm not talking about "clean" in terms of design, but in terms of execution, as VK-machine has pointed out.

Vinncent wrote:

lobawad wrote:

Vinncent wrote:

The most obvious problem with the Ganzfeld experiment is that it cannot produce a negative or falsifying result.

But it can, and sometimes does.

Can you explain how? Remember that the failure to produce a positive is not necessarily the same as a negative.

That would depend on their hypothesis.

A lot of the earlier experiments were plagued with bad controls... a falsifying result would be an experiment which took any flaws in the controls into account, reran the experiment, and found only chance expectation.

If you're talking about the later experiments (or any experiment, really)... a falsifying result would involve finding a flaw in their controls, and rerunning it again taking that flaw into account, and seeing whether or not they still obtained statistically significant results (or not). Without identifying any new flaw in the controls, for an experiment that has gotten a number of successful replications (and a number of unsuccessful replications), you are left with either:

A. One "side" of the researchers are lying. This is particularly complicated, when taking account successful replications from skeptics, and unsuccessful replications from those who primarily study anomalous phenomena.
B. There are still unidentified factors which influence how well a person is able to "send" and how well the other person is able to "receive", that no one is taking into account.

You still haven't explained the mechanism used to falsify. The Ganzfeld experiments have no mechanism for concluding that the phenomenon in question does not occur. They are exploratory, capable only of providing evidence that the phenomenon might occur.

We could do Ganzfeld experiments every day for the next century without a convincing positive result, and we could still say they have not disproved the phenomenon they are intended to test. A century without positive results would only be convincing evidence that the phenomenon does not occur in Ganzfeld test environments.

Science does not work that way. Look at luminiferous aether. Looking for aether didn't turn up anything, but aether was not rejected on those grounds. It was rejected like this: let's say there is aether. If this is true, then such-and-such will occur to something else other than aether.

If there's aether, then matter must do such-and-such
if matter doesn't do such-and-such, then there is no aether
matter does not do such-and-such
therefore there is no aether

P>Q
~Q
therefore ~P

All you've got in Ganzfeld experiments is "maybe P?!" Until you have a Q, such that not-Q will give you not-P, you can't go about convinced the phenomenon tested by Ganzfeld experiments has been "proven", even if you've got statistically significant deviations from chance.

I don't really disagree with any of those points. As I said, my disagreement is mainly with those who say the issue is settled and telepathy (or whatever) exists.

I'd also disagree that the matter is "settled"... after all, this debate has been going on for decades, and this thread alone is five pages in discussing the issue. My main point was for psi researchers themselves... they are convinced by the current body of evidence that "telepathy" and similar phenomena is real. Despite all the design modifications and improvements on controls, the hit rate of those experiments remains somewhere around 30%. It doesn't appear that more trials will lead to a greater hit rate, which means that running more trials is unlikely to convince anyone who doesn't find that rate convincing.

Moving forward to further identify relevant factors that make participants hypothetically "send" or "receive" better, or other things that can be done (besides some sort of sensory deprivation) to improve this, might make for some more convincing experiments. (Though, if they're correct about things such as "greater levels of creativity improve participants scores"... I have no idea how one would go about training such a thing, or quantifying "creativity".)

Which is pretty much my stance. As I think I've said before, if a drug trial showed a medication achieved a 32% response, vs. 25% for placebo, even if that result was statistically significant, it would still mean nothing of practical value. I doubt many researchers would want to invest time and resources to investigate it further.

That's fair. If it turns out there isn't anything, either in training,technology or procedure, that can legitimately improve this "ability", it's not of any practical value.

Which I haven't said. To my mind, forming a testable hypothesis and then testing it constitutes science, no matter how outlandish the hypothesis may be.

No, but that was the original topic, which is what I've been arguing against... as although their (psi researchers) subject matter is strange, their experimental design seems legitimate.

But they have deliberately chosen a system where a hit must happen at least 25% of the time, just by chance. Which is vastly inflated compared to what we would expect under "normal" circumstances. (What are the odds of two random individuals thinking about the exact same thing at same time?)

If senders were able to consistently "transmit" a list of 300 random words with perfect accuracy, I don't think anyone would doubt that was a very significant and startling result.

Well, -expected- to happen, not -must- happen, but that's a minor detail. The controls involved with most experiments in general make it so that what happens in the experiment are not entirely consistent with what we would expect "in real life" (without those controls present), particularly due to noise. No one knows with any real certainty how ESP operates, so its uncertain to what extent ESP is lost in the "noise" of everything else...

...I was originally going to say that I bet "a lot" is lost in the noise involved with everyday life, but this isn't directly implied by the ganzfeld experiments themselves. Most of the controls are only there to discount other forms of information transfer for the sake of the experiment, only the sensory deprivation is used to increase the effect, and I'm not certain whether sensory deprivation is necessary for ESP (or how much it increases the effect, if at all.)

I'm not talking about "clean" in terms of design, but in terms of execution, as VK-machine has pointed out.

I don't have much more to add on this point. It's certainly possible (even less so with the computer assisted autoganzfeld procedure), but the studies I've read have made a point to explain any problems that have arisen during the experiment, as to why they didn't include certain trials (technical problems, power outages, etc). So it seems to be an "either/or", rather than a "to what extent was it properly executed" thing, with runs encountering technical problems thrown out.

Vinncent wrote:this debate has been going on for decades...

Moving forward to further identify relevant factors that make participants hypothetically "send" or "receive" better

As you say, the attempt has been going on for decades, and there isn't anything in the existing methodology to suggest ways to improve the result. You make some suggestions implying you think sensory deprivation is useful in developing mentalism. Remember, there's only so much sensory deprivation you can apply, and to be fair, you should apply it to the sender, so that the sender has no sensory input to suggest to him the 'information' to send out. Just fucking send it.

Another thing that should be tried, if you think mentalism is subject to development, is doing careful experiments to see how mentalism is extinguished, so that the results of a one-in-four trial return to the expected 25% hit rate within uncertainties. This might involve, if you think mentalism is only a short-range effect, trying to send mental images across continents.

As long as you are eliminating uncertainties, though, you should also eliminate the uncertainties that psi experiments are measuring anything, let alone information transfer outside of recognised sensory channels.

You still haven't explained the mechanism used to falsify. The Ganzfeld experiments have no mechanism for concluding that the phenomenon in question does not occur. They are exploratory, capable only of providing evidence that the phenomenon might occur.

I thought I did.

The experiments are designed to control for every form of information transfer, besides the whole "conscious telepathic sending of information" thing. To prove that this isn't the case, one simply has to show either where some bad control is used, or put forth another testable explanation that doesn't rely on telepathy. If there's a bad control, or alternative explanation, it would be easy to run another experiment that controls for these things, and (ideally) watch the effect disappear (results being only chance expectation).

That may sound weak, but only because no one is claiming much about the mechanism in the first place, as the research hasn't progressed to that stage yet... only being centered on showing the effect exists so far. Without doing the above, which is what you're probably getting at, there is no other mechanism which would explain why one group got positive results, and another one negative.

I suppose an immediate parallel I can draw is various psychology research (given, its definitely one of the softer sciences). For example, the amusing the study about how men are rated more attractive when they wear cologne, even when the judge can't smell the cologne, or even know they're wearing it. The two obvious ways to discount such an experiment would be 1, showing there was a bad control, where the "judges" could either smell the participant, or knew they were wearing cologne, or 2, no third parties being able to replicate the experiment.

Failing those two, how would one falsify such an experiment?

We could do Ganzfeld experiments every day for the next century without a convincing positive result, and we could still say they have not disproved the phenomenon they are intended to test. A century without positive results would only be convincing evidence that the phenomenon does not occur in Ganzfeld test environments.

Science does not work that way. Look at luminiferous aether. Looking for aether didn't turn up anything, but aether was not rejected on those grounds. It was rejected like this: let's say there is aether. If this is true, then such-and-such will occur to something else other than aether.

If there's aether, then matter must do such-and-such
if matter doesn't do such-and-such, then there is no aether
matter does not do such-and-such
therefore there is no aether

There is probably little point in continuing to run the same damn experiment over and over again, without correcting for any hypothetical procedural flaw, as its already been conducted numerous times by numerous research groups, with varying results, under the same procedure. I agree a century of failing to produce statistically significant results would be good evidence against that, but since there is already little motivation/funding in the scientific field for replication, that's unfortunately probably not going to happen.

To your second point, though... it seems like something similar could be said of the ESP hypothesis in terms of the ganzfeld experiments.

If telepathy exists, it should produce this result
If it doesn't, telepathy doesn't exist.
Telepathy [does/does not] produce this result
Therefore telepathy [does/does not] exist
[depending on which "side" of the debate you're on]

There is a notable difference in this case, however... the idea of the "aether" (if memory serves) sought to explain some medium by which light travels through. Psi refers to a phenomena... not currently an explanation for that phenomena. It is also possible that there are various explanations other than specifically telepathy that allow for that type of information transfer, though all sound equally crazy.

On the least crazy end, perhaps there's some physical mechanism unrelated to consciousness that allows for that sort of information transfer, that we simply don't know about yet, therefore can't control for. On the more crazy end, I have an older parapsychology book that gets into the philosophical details as to whether "ESP" takes place through "sending/receiving" thought transfer between two people... or whether or not people are somehow able to either A, see into the future to see what the correct answer was, or B, send information to themselves backwards in time after seeing the correct answer.

Though hypothetically possible, and could be controlled for with the computer programs never telling which results were the correct one (merely noting that a correct hit was made)... I'm not sure whether ganzfeld experiments have controlled for this specifically. However, if these were somehow true (or played a role), it would still be a method of transfer currently not understood by physical mechanisms.

Another fair criticism is that "psi" is simply looking for "what's left when you control for everything else", and is also currently why so little can be objectively stated about the phenomena (at this point). This is where the previous criticism would come into play, in terms of future "explanations" behind how psi phenomena operates. I'll stop before posting a lengthy rant about various poor explanations, though.

P>Q
~Q
therefore ~P

All you've got in Ganzfeld experiments is "maybe P?!" Until you have a Q, such that not-Q will give you not-P, you can't go about convinced the phenomenon tested by Ganzfeld experiments has been "proven", even if you've got statistically significant deviations from chance.

Though I try to be as logical as possible, and know a great deal about cognitive biases, I haven't taken a formal logic class, which might be why I don't fully understand this point.

Are we saying,
P=telepathy, ESP, whatever you want to call it
Q=ganzfeld procedure, or does Q refer to the mechanism which explains how ESP occurs, which no one knows yet?

Vinncent wrote:
The experiments are designed to control for every form of information transfer, besides the whole "conscious telepathic sending of information" thing. To prove that this isn't the case, one simply has to show either where some bad control is used, or put forth another testable explanation that doesn't rely on telepathy. If there's a bad control, or alternative explanation, it would be easy to run another experiment that controls for these things, and (ideally) watch the effect disappear (results being only chance expectation).

That may sound weak, but only because no one is claiming much about the mechanism in the first place, as the research hasn't progressed to that stage yet... only being centered on showing the effect exists so far. Without doing the above, which is what you're probably getting at, there is no other mechanism which would explain why one group got positive results, and another one negative.

It hasn't even been made clear in this thread that the receiver is restricted to describing the 'information sent' unambiguously as in 'A', 'B', 'C', or 'D', for the one-in-four experiments. If the experiment involves interpreting output from the receiver that is any more ambiguous than that, I think a lot of your optimism would seem to be misplaced. You know, 3 is too few, 5 is too many.

I'm just trying to explain to you why mainstream scientists might be laughing at your timid optimism.

As you say, the attempt has been going on for decades, and there isn't anything in the existing methodology to suggest ways to improve the result. You make some suggestions implying you think sensory deprivation is useful in developing mentalism. Remember, there's only so much sensory deprivation you can apply, and to be fair, you should apply it to the sender, so that the sender has no sensory input to suggest to him the 'information' to send out. Just fucking send it.

That might be interesting. To my knowledge, only the "receiver" is typically put under sensory deprivation, though I might be wrong. The whole "sensory deprivation" itself is admittedly a matter of confusion for me... I can't see much logical basis for it off the top of my head, other than the results, which I don't know of similar experiments which -didn't- utilize sensory deprivation on the part of the receiver. But without a casual relationship, it still sounds to me like "put a shoe on your head while receiving images from the sender, because that's what the experiment calls for."

Theoretically, it would seem as though the hit rates should depend on both the ability of the receiver to "get" the information, but also of the sender to, well, "send" information. However, in what I've read, it seems as though little time is spent priming the "sender" to perform better. I wonder whether it would make a difference (if 'they' don't already) subject the sender to sensory deprivation.

Another thing that should be tried, if you think mentalism is subject to development, is doing careful experiments to see how mentalism is extinguished, so that the results of a one-in-four trial return to the expected 25% hit rate within uncertainties. This might involve, if you think mentalism is only a short-range effect, trying to send mental images across continents.

It's been theorized that mentalism is a developed ability, but I don't know how much work as been put into establishing this. As far as I can tell, most ganzfeld experiments require participants to have never taken part in a ganzfeld experiment before... I would guess to help with replicibility (if that's a word), so that other researchers don't have to taken into account the ability/training level of their volunteers. At best, they seem to ask for people who "believe in psi" among a few other factors they think are condusive.

However, other psi research has indicated that the effect does not fall off with distance, which is what is expected of practically every other known phenomena. This is why I don't buy into the theory of "psi energy" as an explanation... because it doesn't follow this behavior, and has never been directly detected by any means. A much better treatment of this is laid out in the "air" .pdf I posted earlier in this thread... if you can't find it, I can post it again, but its worth a read.

As long as you are eliminating uncertainties, though, you should also eliminate the uncertainties that psi experiments are measuring anything, let alone information transfer outside of recognised sensory channels.[/quote]

I unfortunately don't understand this point. A qualitative component is usually included, in terms of "describe what you 'saw'" outside of the quantitative hit rate... but why should psi research abandoned measurements? (if that's your implication... again, I don't quite follow this point)

(edit: fucked up the quoting)

It hasn't even been made clear in this thread that the receiver is restricted to describing the 'information sent' unambiguously as in 'A', 'B', 'C', or 'D', for the one-in-four experiments. If the experiment involves interpreting output from the receiver that is any more ambiguous than that, I think a lot of your optimism would seem to be misplaced. You know, 3 is too few, 5 is too many.

I'm just trying to explain to you why mainstream scientists might be laughing at your timid optimism.

I think this point was explained earlier... a problem with the early experiments was that a "hit" was at the discretion of the judge, as to how close they -felt- the receiver was to describing the given picture. This was corrected in later experiments, by giving four choices, so that a "hit" wasn't subject to subjective interpretation.

That is, unless your point was that "four choices is too few", and they should use five, ten, or a hundred different choices instead. Although it would be pretty easy to modify the experiment in this way, there are two immediate problems I can see in adding too many choices:

1. With more choices, the chance that randomly selected images end up looking "too similar" is increased. For example, you might have someone receiving "I see someone traveling... wheels... busy city street... buildings...", and have one image/video clip of someone doing riding a bicycle around town, and another one driving their automobile. With too many choices present, the information "received" is more likely to fit more than one image/clip.
2. The process is fairly time consuming, and more images/clips make it even moreso.

Neither of these really say much about the statistical significance of previous experiments, though... it would simply be a matter of recalculating probabilities based on the (new) number of choices in the experiment. So, although certainly do-able... I don't see how it would improve anything.

And, as an end-note... I'm well aware of how unaccepted these experiments are in the "scientific community". In reality, most people simply don't give a shit, others look at it and say "well that's neat" while continuing to not give a shit (as explained earlier, currently the effect is very small with little useful application), and others outright reject it for various reasons (some valid, some pretty silly).

As a post-end-note... other psi (or related) researchers have claimed that the scientific community has rejected psi research out of paradigmal dogmatism. I don't think this is as significant as some claim. For the large part, from what I've seen and read, most researchers don't give a shit about areas outside their chosen field of exploration, unless it directly applies to what they're doing. I'd venture most researchers don't have an opinion on psi, because nobody gives a shit, rightfully so, because the effect is so small to begin with.

Though some have a valid reason to claim this... a certain video has been making its way around, "RUPERT SHELDRAKE'S BANNED TEDX TALK", where woo supporters have claimed it was taken down due to dogmatism. In reality, Sheldrake's research sucks, because it uses retardedly bad controls. In cases like this, its simply lying to say that certain research is rejected due to "established dogmatism", when the real reason is that some researchers are simply terrible at what they do, as noticed by the rest of the scientific community.

Cito di Pense wrote:It hasn't even been made clear in this thread that the receiver is restricted to describing the 'information sent' unambiguously as in 'A', 'B', 'C', or 'D', for the one-in-four experiments. If the experiment involves interpreting output from the receiver that is any more ambiguous than that, I think a lot of your optimism would seem to be misplaced. You know, 3 is too few, 5 is too many.

My problem is more with the lack of ambiguity. The receiver has to pick one of the four, even if none of them more than remotely resembles anything that passed thru his mind during the experiement. So, for example (and I know this isn't exactly how the procedure is scored), if none of the videos match anything he thought of, he give three of them a 0/5 rating, and one of them 1/5, because maybe it had a bird in it, and was thinking about girls while he was sitting with the ping pong balls on his eyes, and "bird"can be slang for "girl." And if that happened to be the correct choice, then it counts as a hit. When really, the test was a complete miss.

Bascially, the "floor" for the lowest hit rate is artificially elevated to 25%, when it really should be something just above zero.

My problem is more with the lack of ambiguity. The receiver has to pick one of the four, even if none of them more than remotely resembles anything that passed thru his mind during the experiement. So, for example (and I know this isn't exactly how the procedure is scored), if none of the videos match anything he thought of, he give three of them a 0/5 rating, and one of them 1/5, because maybe it had a bird in it, and was thinking about girls while he was sitting with the ping pong balls on his eyes, and "bird"can be slang for "girl." And if that happened to be the correct choice, then it counts as a hit. When really, the test was a complete miss.

Bascially, the "floor" for the lowest hit rate is artificially elevated to 25%, when it really should be something just above zero.

It's true that given the set-up, one does expect a 25% chance, which is probably unexpected (undetectable, etc) in "normal life". However, the problem with -not- using this set-up, is that (experimentally) it introduces a great deal of subjectivity in analyzing results. Identical results could be interpreted as a hit or miss, depending on who is judging.

Additionally, if there is the possibility that none of choices match what was supposed to be sent, this is also detrimental to the whole "subjectively accurate" interpretation of the receiver. For example, someone reports seeing "A bunch of bricks, stacked on top of one another, some sort of monolith, very dry...". A possible image might be "The pyramids of Giza", but since they didn't happen to pick up some factor, like "sand", might conclude that none of them are close enough to match the description given, because they weren't (slash aren't) able to receive the full picture.

The experimental set-up is designed to minimize the amount of subjective interpretation that is required, so that different groups can't look at the same results, and one conclude that a certain run is a hit, and a different group conclude that the run is a miss. Otherwise meta analyses would be next to meaningless.

It would certainly be impressive if someone could draw a picture, in its entirety, down to the last detail, what was sent... but experimentally, such a result would seem to be an "anecdote" open to interpretation as to "just how accurate it was(n't)".

Vinncent wrote:It's true that given the set-up, one does expect a 25% chance, which is probably unexpected (undetectable, etc) in "normal life". However, the problem with -not- using this set-up, is that (experimentally) it introduces a great deal of subjectivity in analyzing results. Identical results could be interpreted as a hit or miss, depending on who is judging.

Additionally, if there is the possibility that none of choices match what was supposed to be sent, this is also detrimental to the whole "subjectively accurate" interpretation of the receiver. For example, someone reports seeing "A bunch of bricks, stacked on top of one another, some sort of monolith, very dry...". A possible image might be "The pyramids of Giza", but since they didn't happen to pick up some factor, like "sand", might conclude that none of them are close enough to match the description given, because they weren't (slash aren't) able to receive the full picture.

The experimental set-up is designed to minimize the amount of subjective interpretation that is required, so that different groups can't look at the same results, and one conclude that a certain run is a hit, and a different group conclude that the run is a miss. Otherwise meta analyses would be next to meaningless.

It would certainly be impressive if someone could draw a picture, in its entirety, down to the last detail, what was sent... but experimentally, such a result would seem to be an "anecdote" open to interpretation as to "just how accurate it was(n't)".

That gets back to the main point: What you describe at the end there is what most people think of when they imagine what "telepathy" would be like if it exists. The Ganzfeld produced nothing like that, and isn't even designed to demonstrate that. A hit is a hit, no matter how just vaguely similar or eerily accurate the actual description of the "received" image might be.

So the question remains: Just what is the Ganzfeld supposed to be measuring? Scientific methodology is supposed to be a means to an end. An experiment is not run just for its own sake. An experiment is supposed to answer questions about phenomena that occur or exist outside of the experiment. The Ganzfeld seems to be about nothing other than the Ganzfeld.

That gets back to the main point: What you describe at the end there is what most people think of when they imagine what "telepathy" would be like if it exists. The Ganzfeld produced nothing like that, and isn't even designed to demonstrate that. A hit is a hit, no matter how just vaguely similar or eerily accurate the actual description of the "received" image might be.

Without any evidence on this point... I agree with you that's what most people think of in terms of ESP/telepathy/etc. However, it's also a view that's taken to an illogical extreme. There are many people who claim to have a significant experience involving ESP, but it's merely anecdotal, so isn't really evidence of the phenomena. These accounts might be true, or merely a matter of coincidence and cherry-picking experiences, so aren't good evidence of the phenomena.

Playing off your earlier ideas of what you would consider to be "significant, to the extent of being useful", the ganzfeld experiments don't seem to imply that the effect is large enough to be of practical use. Playing off -my- earlier points, it is hypothetically possible that this effect might be increased, if we can learn more about the phenomena, and develop a reasonable explanation that ties in with everything else we know about reality. Or perhaps not, and the effect can't be significantly increased, reducing it to merely an odd curiosity. Time will probably tell.

But still... what most people's impression are of any field of research are wildly different from what it actually entails. I tell people I'm a physical science major, and they imagine I'm blasting lazers at things and building robots, when in reality, I'm mostly just studying textbooks and working on theoretical research applications which don't have any immediate use, other than characterizing various phenomena in a quantitative way. Reality is much less impressive then the idea. ESP can still exist without it meaning that a simple 5 minute meditation will allow one to reach into the future and grab future lottery numbers... which probably isn't the case, considering this doesn't happen.

So the question remains: Just what is the Ganzfeld supposed to be measuring? Scientific methodology is supposed to be a means to an end. An experiment is not run just for its own sake. An experiment is supposed to answer questions about phenomena that occur or exist outside of the experiment. The Ganzfeld seems to be about nothing other than the Ganzfeld.

It is quantitatively measuring the effect that one form of consciousness is able to transfer information to another form of consciousness (at least in the form of "hit rates"... not directly measuring "consciousness" itself, which is another field of exploration I've been looking into with little results), through methods that aren't currently understood by mainstream science. That's pretty much it, at this point. To repeat myself again, it's hardly a "completed" field of research, although various psi researchers claim that "telepathy" and the like has already been well established (one I don't agree with, or we wouldn't be arguing about it). The subtle implication is that this "psi" phenomena is present in everyday life (at least one has hypothesized it as being an evolutionary beneficial ability), but much like entanglement, is impossible to "see" unless its performed in a highly controlled environment which can differentiate it from the natural noise that is present everywhere in cognitive processing.

[edit, fucked up the quotes yet again]

Vinncent wrote:To my knowledge, only the "receiver" is typically put under sensory deprivation, though I might be wrong. The whole "sensory deprivation" itself is admittedly a matter of confusion for me...

... it seems as though little time is spent priming the "sender" to perform better. I wonder whether it would make a difference (if 'they' don't already) subject the sender to sensory deprivation.

As far as I can tell, most ganzfeld experiments require participants to have never taken part in a ganzfeld experiment before... At best, they seem to ask for people who "believe in psi" among a few other factors they think are condusive.

However, other psi research has indicated that the effect does not fall off with distance...

A qualitative component is usually included, in terms of "describe what you 'saw'" outside of the quantitative hit rate...

The 'sensory deprivation' component seems not to have been rigourously tested, and a qualitative component is accepted, because a psi experiment is more like a seance than it is like a science experiment. The trappings of a science experiment are there, and the trappings of a seance are there.

Really, if it's a capacity of an individual, then an individual who contributes to the higher hit rate will further contribute in repeat experiments. It would be severely damaging to the credibility of the theory that psi is a capability of individual humans if it did not. "Sometimes it happens, and sometimes it doesn't" is not a scientific statement. If psi researchers think the effect does not fall off with distance, then repeat experiments on the same individual will not show diminution if that's the variable.

It's almost embarrassing to discuss seriously principles of basic science with someone who advocates further investigation of psi. I don't do this because I think further research in psi is worthwhile (I don't), I do it to further your scientific education.

Vinncent wrote:1. With more choices, the chance that randomly selected images end up looking "too similar" is increased. For example, you might have someone receiving "I see someone traveling... wheels... busy city street... buildings...", and have one image/video clip of someone doing riding a bicycle around town, and another one driving their automobile. With too many choices present, the information "received" is more likely to fit more than one image/clip.

If that's the case, the confusion between no hit and any hit is already too great. Very poor excuse, because even with only four images, if 'similarity' to the sender's image is what the researcher demands of the receiver in order to pick one of four, the researcher should probably test the receiver against a machine that randomly picks the images, wIthout telling the receiver that there isn't a sender out there. Now that's a double-blind experiment, for you!

Vinncent wrote:2. The process is fairly time consuming, and more images/clips make it even moreso.

Boo fucking hoo. Science is difficult.

Shrunk wrote:My problem is more with the lack of ambiguity. The receiver has to pick one of the four, even if none of them more than remotely resembles anything that passed thru his mind during the experiement. So, for example (and I know this isn't exactly how the procedure is scored), if none of the videos match anything he thought of, he give three of them a 0/5 rating, and one of them 1/5, because maybe it had a bird in it, and was thinking about girls while he was sitting with the ping pong balls on his eyes, and "bird"can be slang for "girl." And if that happened to be the correct choice, then it counts as a hit. When really, the test was a complete miss.

Bascially, the "floor" for the lowest hit rate is artificially elevated to 25%, when it really should be something just above zero.

That's how they do it? No wonder the scientific community shakes its collective head and walks away. These are people pretending to apply the scientific method by playing with statistics. Oh, and ping-pong balls.

Yes, if they lowered the baseline to just above zero, they wouldn't be assuming that psi exists and that it sends an imprecisely specified portion of the 'information' in the receiver's, um, consciousnessness, before they start their experiments.

These folks have an imprecise definition of 'information', an imprecise definition of 'psi', and an imprecise definition of science. But the show must go on, I guess.

Vinncent wrote:It's true that given the set-up, one does expect a 25% chance, which is probably unexpected (undetectable, etc) in "normal life". However, the problem with -not- using this set-up, is that (experimentally) it introduces a great deal of subjectivity in analyzing results. Identical results could be interpreted as a hit or miss, depending on who is judging.

Thanks for presenting the cartoon version of subjectivity/objectivity used in psi 'research'. By definition, the experience being transmitted by the sender is 'subjective' and the information being evaluated by the receiver's experience is 'subjective'.

Vinncent wrote:...they weren't (slash aren't) able to receive the full picture.

Give the man a cigar! The hazy concept of what 'picture' gets transmitted and received is hazy. The baseline for the hit rate is not 25%, but close to 0, which is what a hazy impression of nothing at all is like. If the receiver were actually allowed to admit that he didn't receive any information (which is a possibility, is it not?) then the experiment might tend toward being science. If the receiver can't admit that he didn't receive anything, it's as much as to assume psi exists.

Vinncent wrote:The experimental set-up is designed to minimize the amount of subjective interpretation that is required, so that different groups can't look at the same results, and one conclude that a certain run is a hit, and a different group conclude that the run is a miss. Otherwise meta analyses would be next to meaningless.

It would certainly be impressive if someone could draw a picture, in its entirety, down to the last detail, what was sent... but experimentally, such a result would seem to be an "anecdote" open to interpretation as to "just how accurate it was(n't)".

Damn skippy. The experimental set-up is designed to keep people interested in psi research busy.

The 'sensory deprivation' component seems not to have been rigourously tested, and a qualitative component is accepted, because a psi experiment is more like a seance than it is like a science experiment. The trappings of a science experiment are there, and the trappings of a seance are there.

I agree with the first part. The necessity of sensory deprivation has not, from what I've seen, been verified to produce the results. I disagree with the second, where you mistake both the quantitative and qualitative analysis involved. The later is not put forth as evidence for psi phenomena, rather, is an unrelated addition to that specific effort in order to attempt to understand "What sort of information/imagery is received by the "receiver".

Really, if it's a capacity of an individual, then an individual who contributes to the higher hit rate will further contribute in repeat experiments. It would be severely damaging to the credibility of the theory that psi is a capability of individual humans if it did not. "Sometimes it happens, and sometimes it doesn't" is not a scientific statement. If psi researchers think the effect does not fall off with distance, then repeat experiments on the same individual will not show diminution if that's the variable.

As mentioned before... most experiments I know of involving the ganzfeld procedure utilize participants which have not took part in previous experiments, to aid in reproduction of the experiment. Is there a group that is regularly available to do this experiment repeatedly, and has been tested? I'm not sure. I'm also not sure, in the case of the ganzfeld experiments, whether or not they've been tested over long distances. The last I read was over a distance of 30m separation of rooms. I don't remember the specific studies cited by the AIR.pdf, but other psi experiments seem to imply the effect does not fall off with distance, implying that framing in terms of "energy" is probably barking up the wrong tree.

It's almost embarrassing to discuss seriously principles of basic science with someone who advocates further investigation of psi. I don't do this because I think further research in psi is worthwhile (I don't), I do it to further your scientific education.

If that's the case, the confusion between no hit and any hit is already too great. Very poor excuse, because even with only four images, if 'similarity' to the sender's image is what the researcher demands of the receiver in order to pick one of four, the researcher should probably test the receiver against a machine that randomly picks the images, wIthout telling the receiver that there isn't a sender out there. Now that's a double-blind experiment, for you!

I've been having trouble finding the documentation for the autoganzfeld experiments (props to you if you can find it for the sake of discussion), but I'm curious as to the basis for your assumption that a "1 in 4 chance" is somehow not a "1 in 4 random chance" in this experimental setup. The multiple-choice protocol doesn't discount the results... after all, if someone describes something that is incredibly close to one of the images, yet chooses the wrong clip/picture, this isn't somehow counter as a "hit" or "close enough", hence the reason for the multiple choice protocol in the first place, to discount subjective interpretation.

Boo fucking hoo. Science is difficult.

...Really? So do it yourself, if your curious as to how this might change the statistical signifiance. I fail to see how this answers my last point.

That's how they do it? No wonder the scientific community shakes its collective head and walks away. These are people pretending to apply the scientific method by playing with statistics. Oh, and ping-pong balls.

Yes, if they lowered the baseline to just above zero, they wouldn't be assuming that psi exists and that it sends an imprecisely specified portion of the 'information' in the receiver's, um, consciousnessness, before they start their experiments.

These folks have an imprecise definition of 'information', an imprecise definition of 'psi', and an imprecise definition of science. But the show must go on, I guess.

I missed the memo where statistical analysis now qualifies as pseudoscience. You had better inform every statistician ever, as they would love to know. In implying their methodology is pseduoscientific, you should provide a clear response, rather than snark.

Thanks for presenting the cartoon version of subjectivity/objectivity used in psi 'research'. By definition, the experience being transmitted by the sender is 'subjective' and the information being evaluated by the receiver's experience is 'subjective'.

I'm not sure of the point here. The evidence is not in the subjective experience, but in the objective hit rate.

Give the man a cigar! The hazy concept of what 'picture' gets transmitted and received is hazy. The baseline for the hit rate is not 25%, but close to 0, which is what a hazy impression of nothing at all is like. If the receiver were actually allowed to admit that he didn't receive any information (which is a possibility, is it not?) then the experiment might tend toward being science. If the receiver can't admit that he didn't receive anything, it's as much as to assume psi exists.

I don't understand this point either. If the baseline hit rate is 0, that only further speaks to the statistical significance of the results. However, this is blatantly untrue... the expected hit rate due to chance should be 25%, given four potential choices. I'll pass on the cigar, though.

Damn skippy. The experimental set-up is designed to keep people interested in psi research busy.

Those involved with psi research are putting their entire education credentials on the line. Aside from basic scam artists who never went to college, such as Dr. Emoto, there is little to no motivation for anyone to get involved with psi research, other than an earnest desire to explore a certain subset of reality that tends to be ignored by most (reasonably so, since the effect is so slight with no direct practical applications, as explained earlier).

Certainly, people have made far more stupid decisions... but I fail to see how people with Ph.D's in various scientific fields would sacrifice their credibility, continuously, for comparatively little monetary gain, pushing for the reality of something that you claim they know doesn't actually exist. I can think of specific examples to the contrary ('Dr.' Emoto and the like), but this is hardly representative of the majority of scientists exploring this field, and such types tend to be looked down upon.

Vinncent wrote:I've been having trouble finding the documentation for the autoganzfeld experiments (props to you if you can find it for the sake of discussion), but I'm curious as to the basis for your assumption that a "1 in 4 chance" is somehow not a "1 in 4 random chance" in this experimental setup. The multiple-choice protocol doesn't discount the results... after all, if someone describes something that is incredibly close to one of the images, yet chooses the wrong clip/picture, this isn't somehow counter as a "hit" or "close enough", hence the reason for the multiple choice protocol in the first place, to discount subjective interpretation.

...I missed the memo where statistical analysis now qualifies as pseudoscience.

That memo tells you that arbitrarily setting the number at four, rather than more, or less, means that you haven't ruled out a misapplication of statistical modeling of data. A requirement for thoroughness seems to recommend doing studies from two to ten or more and watching how the hit-rate above expected peaks at four. Or doesn't. If it peaks at two, but not significantly more than at four given the definition of statistical significance for that comparison, or somewhere else, nobody seems to know it. Maybe they actually tried this, and just didn't publish the results. If there was a trend in hit-rate-above-expected that depends on the number of choices offered, psi researchers (and the rest of us) could see the method-dependence of their results. Instead, the psi advocates insist on beating the dead horse of hit rate above 25% which tells us nothing about the dependence of hit rate on number of choices. I recommend you learn something about science, Vinncent, instead of just about statistics.

I've got more ideas for these 'researchers', such as testing whether the possession of perfect pitch musically is correlated with psi 'abilities'. And so on. But I have another suggestion for psi fanboys: Either start doing your own research, or get a life. From what I've seen, there's no need to have any scientific training in order to design a psi 'experiment' and speculate that practicing 'meditation' is correlated with increased 'psi' effect, which would require longitudinal studies on subjects from the time they begin practice to the time when their psi abilities peak at 32% hit rate, or whatever it might turn out to be. Oh, you say, these experiments would be too expensive and time-consuming. Welcome to science, Vinncent.

I also do not think the subject ('receiver') should be charged with deciding which image he 'received'. I think a researcher in the room should be charged with the task of deciding whether or not there is anything in the ramblings of the subject that have anything to do with any of the images, and see what the hit rate turns out to be. You know, just as an attempt at a control.

Vinncent wrote:

There is probably little point in continuing to run the same damn experiment over and over again, without correcting for any hypothetical procedural flaw, as its already been conducted numerous times by numerous research groups, with varying results, under the same procedure. I agree a century of failing to produce statistically significant results would be good evidence against that, but since there is already little motivation/funding in the scientific field for replication, that's unfortunately probably not going to happen.

But that is not what I am saying. The problem is that a century of failure to produce statistically significant results would still NOT falsify psi.

Let's say we "believe in psi" and have our century of failed Ganzfeld tests.

Someone says: "Psi almost certainly doesn't happen- the results of Ganzfeld tests do not vary significantly from random guessing."

"Oh, come on," we'd say. "Who walks around with ping-pong balls strapped to their eyes? For all we know the Ganzfeld experiments are really only showing that "plastic is the natural enemy of magic", like they say on that children'ts show."

And, if different labs do the experiment long enough, sooner or later some lab is going to "hit the jackpot" statistically. Then psi believers would say "Aha! That lab must have had the right conditions!"

The Ganzfeld experiment is only capable of showing a positive! But even a positive is based on fallacious reasoning, as you shall see:

Vinncent wrote:
To your second point, though... it seems like something similar could be said of the ESP hypothesis in terms of the ganzfeld experiments.

If telepathy exists, it should produce this result
If it doesn't, telepathy doesn't exist.
Telepathy [does/does not] produce this result
Therefore telepathy [does/does not] exist
[depending on which "side" of the debate you're on]

That does not work, because you've got a formal fallacy called "affirming the consequent" in there.

P = telepathy
Q = result

P > Q
Q
therefore P <<<Fail! fallacy of affirming the consequent

Telepathy produces this result
we find this result
therefore telepathy

This is a fallacy because it is not necessarily true that it was telepathy producing this result. Let's look at modus tollens again:

P > Q
~Q
therefore ~P

If telepathy happens, then we'll have a deviation from random results in this test
We do not have a deviation from random results in this test
therefore telepathy does not happen

This is valid. At first glance, you might think that this gives us the falsification we need to say that Ganzfeld tests of psi are scientific.

But let's write out that argument in a way that is in accord with real life:

If telepathy happens in this test, then we'll have a deviation from random results in this test
We do not have a deviation from random results in this test
therefore telepathy does not happen in this test

So of course we try to get funding for yet another test...

That memo tells you that arbitrarily setting the number at four, rather than more, or less, means that you haven't ruled out a misapplication of statistical modeling of data. A requirement for thoroughness seems to recommend doing studies from two to ten or more and watching how the hit-rate above expected peaks at four. Or doesn't. If it peaks at two, but not significantly more than at four given the definition of statistical significance for that comparison, or somewhere else, nobody seems to know it. Maybe they actually tried this, and just didn't publish the results. If there was a trend in hit-rate-above-expected that depends on the number of choices offered, psi researchers (and the rest of us) could see the method-dependence of their results. Instead, the psi advocates insist on beating the dead horse of hit rate above 25% which tells us nothing about the dependence of hit rate on number of choices. I recommend you learn something about science, Vinncent, instead of just about statistics.

I fail to see how this is an argument against statistical significance. The idea of discounting the results due to the "file drawer" effect was countered elsewhere, by another skeptic site:

http://skepdic.com/ganzfeld.html

If reproducibility is "beating a dead horse", then yes, psi researchers have beat that horse into oblivion, to the point where no amount of further beating will convince anyone who believes that statistically significant results are not statistically significant, if those results don't conform to their preconceived beliefs.

I've got more ideas for these 'researchers', such as testing whether the possession of perfect pitch musically is correlated with psi 'abilities'. And so on.

This was used in some replications, in terms of sending "sound" rather than "imagrey", which simply indicates that they weren't following the experimental procedure, which called for images or video clips to be used.

But I have another suggestion for psi fanboys: Either start doing your own research, or get a life.

Why? The ganzfeld experiments have thousands of trials spanning over many decades, with increasingly improved controls to discount any external contamination, while still producing the same ~30% hit rate. I suppose people could continue to do the same shit over again, but I fail to see what this would establish, if it doesn't correct any hypothetical flaws in the previous experiments. It fails to address why certain research groups get statistically significant results, and others don't, in any way.

From what I've seen, there's no need to have any scientific training in order to design a psi 'experiment' and speculate that practicing 'meditation' is correlated with increased 'psi' effect, which would require longitudinal studies on subjects from the time they begin practice to the time when their psi abilities peak at 32% hit rate, or whatever it might turn out to be. Oh, you say, these experiments would be too expensive and time-consuming. Welcome to science, Vinncent.

Certainly. Anyone can perform an experiment. Its validity depends on its controls and execution. The 32% hit rate doesn't appear to be associated with those experienced in meditation, or experience with participating in ganzfeld experiments, but includes everyone tested. The fact that many psi researchers are post-docs is irrelevant, and hardly supports your implication that they're "amateurs".

I also do not think the subject ('receiver') should be charged with deciding which image he 'received'. I think a researcher in the room should be charged with the task of deciding whether or not there is anything in the ramblings of the subject that have anything to do with any of the images, and see what the hit rate turns out to be. You know, just as an attempt at a control.

You can think that, but you would be introducing unnecessary contamination into an experiment which only needs to rely on randomized computer controls. The hit rate isn't based on the qualitative account of the imagery they "received", but rather on their ability to choose the correct image of the choices given.

Vinncent wrote:The ganzfeld experiments have thousands of trials spanning over many decades, with increasingly improved controls to discount any external contamination, while still producing the same ~30% hit rate. I suppose people could continue to do the same shit over again, but I fail to see what this would establish, if it doesn't correct any hypothetical flaws in the previous experiments. It fails to address why certain research groups get statistically significant results, and others don't, in any way.

lobawad wrote:So of course we try to get funding for yet another test...

@Vinncent:

Please, please try to understand the above statement: There has to be some reason that, in applying for funding, no one ever publishes (or refers to) a Ganzfeld experiment with negative results. Hint: It isn't because we are sure that no Ganzfeld experiment ever yields negative results. This is what is meant by saying that nothing can falsify theories of psi.

Vinncent wrote:

I've got more ideas for these 'researchers', such as testing whether the possession of perfect pitch musically is correlated with psi 'abilities'. And so on.

This was used in some replications, in terms of sending "sound" rather than "imagrey", which simply indicates that they weren't following the experimental procedure, which called for images or video clips to be used.

Oops. Negative result. Bin it. Seriously, though, I meant testing people with perfect pitch using visual images. Why not? Why isn't perfect pitch considered as recommending psi candidates? Afraid you might get a negative result?

Vinncent wrote:The hit rate isn't based on the qualitative account of the imagery they "received", but rather on their ability to choose the correct image of the choices given.

Why include all the patter from the qualitative account? Is that supposed to support their choice of the correct image? What about the qualitative patter from those who did not select the correct image (the 70% who don't)? Couldn't that tell us something about the way psi experiments are conducted? Of course not. No psi researcher wants to focus on the 70% who do not select the 'sent' image, even though there may be something in their patter to investigate. That's why you want to do repeat experiments on the same individual. Lots of them.