LOL. My argument is not in any way ad hominem (yours, on the other hand, are full of that fallacy), and I didn't accuse him of lying. I've actually been published in peer-reviewed journals- you're lucky if your parents "published" your stuff on the refrigerator.

I am published and I bet you I was published at a younger age than you. Thank you very much.

your argument is indeed ad hominem, see my edit. It can also be easily proven. we actually keep our grant proposals for this reason.

My argument against Bem was not in any way ad hominem. Do you even know what the term means?

If indeed your argument is not ad hominem then this excerpt from Bem's rebuttal should put your argument to bed. He even does a better job than you at explaining how this sort of error works.

Also, can you explain why his methods didn't show any of the same statistical artifacts (in the direction of psi) in the control groups? surely you can't use chance to explain how his "flawed" methodology resulted in the same effect replication after replication in the same direction of the hypothesis. Yes, such methodology would be flawed if used but the error you need to disqualify this paper has got to be systematic. Your argument does not lead to the observed results even if it were true, which it isn't as is clear from Bem's rebuttals

That's him completely missing the point.. because that's exactly what he did. By having so many arbitrary partitions to begin with, and clearly and obviously making more and doing funky stuff during the experiment itself, and testing many things separately, the odds are much higher than his p-value that SOMETHING will appear "significant". If I ran 100 psi tests with normal people, I'd probably get one at the nominal p=.01 level. That's not a meaningful result.

What he does is even worse, because he decides what to apply statistical tests to only after deciding how to partition his data. It's the equivalent of going through the data, realizing that blonde girls do well at one task between 2 and 3pm, then doing a statistical test on exactly that and presenting it as a p<.01 finding. It's worse than worthless.

As somebody else said, I have no problem with people conducting these experiments competently. If somebody wants to test exactly one hypothesis going in, and analyze the data appropriately, go for it. And if somebody, somehow, gets an actual significant result without data manipulation shenanigans, and other independent people in other places keep reproducing a significant result, then maybe there's something going on. Of course I'm not holding my breath. Z=2.x with obvious shenanigans isn't even a first result to care about.

There is absolutely no grounds for this critique. Yes he does change his variable but in separate experiments which is totally fine given this was done over a long period of time and considering the amount of work needed to convince people with low bayesian prior.

Granted, the paper would have been slightly stronger if he had more exact replications but that's hardly a dent in the study. Moreover, you still haven't explained why his statistical error seems to be consistent over such a large sample and why it doesn't appear in his negative controls. I agree that it is possible for such methodology to render a study useless, in this case, even if there is such methodology flaw in his research being more exploratory than confirmatory, it still wouldn't account for the results. While that maybe the case purely by luck (that it didn't affect this study) it is what it is and unfortunately for you stubborn skeptics, that argument won't work this time.

You seem to forget that the odds of such error resulting in those results time after time are very minute. The raw unadultered data is also out there for people to manipulate themselves. Finally, this argument of yours, if was indeed used to disqualify this paper then it should be used to disqualify most papers in psychology today. you can't use the prior argument here since you already used it when you called for bayesian analysis, which also turned out to be significant when done properly.

here is also a successful confirmatory study just in case you aren't starting to take this seriously
http://papers.ssrn.com/sol3/papers.c...act_id=1715954

Talking to you is absolutely ****ing pointless. You can only parrot phrases, generally incorrectly. You can't follow an argument or even understand the first thing I'm talking about (you accused me of ad hominem attacks on Bem, said I accused him of lying, and now you say I called for bayesian statistics.. seriously, wtf?), and you can't express a coherent argument to save your life.

Furthermore, the "confirmatory study" was only marginally significant in ONE of the two areas (max Z again 2.x), that author did some mining of his own, and isn't even a confirmatory study, since it was done.. 5 years ago. If the 53% number is real, then even a 10000-event trial, which should only take about a 40-hour work week to do, could spit out a z-score of 6. It's not hard to do research that doesn't suck, assuming you know what research that doesn't suck is.

lol ok, all this is total BS and I am not even sure why I bother to dismantle it one by one.

for your information since you really do seem to be clueless about how research work. people exchange their work while they're working on it so it's totally fine for a confirmatory study to be done before the paper is published. more confirmatory studies are currently being done.

you make no point but rather a general hand-waving criticism that doesn't include you showing how it would result in these results being invalidated or how his method biases him to gain a positive result in erotic groups but not in the negative control groups. your critique also becomes ad hominem if you persist on it after reading Bem's rebuttal. I foolishly assumed you actually did read it and try to understand it.

I did not claim that you asked for Bayesian analysis. the "you" was refering to the group of absurd skeptics who seem to type and publish faster than they can think.

I seem to parrot stuff because I guess I am not delusional enough to think that I am a more reliable source than Bem, JPSP reviewers and other independent, far more qualified scientists and statisticians. you don't seem to have that tendency and you even pride yourself of that. Another reason for this parroting tendency is that your critiques are pretty unoriginal and have been already made by low-level thinking skeptics and refuted months ago. As much as I wish you did, you simply give me no reason to come up with original rebuttals.

Also, the confirmatory study shows significant effect in one experiment, true, but even in the one that wasn't significant it shows that the effect is significant for the high-boredom subgroup which confirms Bem's claim in his study. Can your rubbish critique also explain that pattern?

you are trying to use the exploratory vs confirmatory argument to disqualify a paper which is absolutely ******ed. you need exploratory research to do confirmatory research. The notion that Bem's work is exploratory is also erroneous. His work is based on several previous works that he cites in his paper.

LOL. Good luck with those. Given the piss-poor sample sizes, if this actually DID sit for that many years, that makes the whole case even more suspicious.

BECAUSE HE DECIDES THE HYPOTHESES HE'S GOING TO "TEST" POST-HOC, AFTER LOOKING AT DATA, AND ADVOCATES EVERYBODY DOING JUST THAT IN A BOOK HE WROTE. If you design your hypotheses to fit your data (well, the noise in it in this case), you're always going to be on the plus side. Given the paper, this is by far, far, the most likely explanation.

Not the case in this paper. What he said or didn't say before in a book that he published seperately is merely a philosophical point of contention about science, which he is totally entitled to given his experience and contribution to science. Using that as evidence to point out he did this in this paper is by definition ad hominem, especially when he clearly states that was not what he did.

the suspicious comment is also a non-argument in this case. Both because the work is very well documented and because it's Bem. But also because the notion that a longer research project makes it less reliable is asinine.

Finally, you clearly don't understand how such flawed methodology would affect data patterns in a sample this large. The odds against that in this paper are astronomical and you really ought to be clueless to not know that. Even if an experiment is flawed, it can still be 100% useful and publishable. Thank god science works that way

Desperado,


I will give you a long response but in it you will see that i will provide you with a way to answer for yourself all this.


Why dont you write your own program and anyone here also interested and test this using even better cleaner statistics than the paper seems to be doing. Its a simple program. Create 2 covered pictures left and right and reveal the picture after you select one and see how often you reveal the picture that has "hot" content and produces for that reason some higher response/interaction (if this is valid claim to begin with ). You can choose to have the computer pick before or after your choice is made. He had 3600 trials. You can do that within 3-4 hours of leisure time over a week or sooner.

In fact a program like vueprint - search it freeware at download . com - can do that if you create a directory with say 1000 hot pictures and 1000 dull ones and have the computer pick one with your eyes closed and you try to guess what it is , or you do not close your eyes, you pick if the next picture will be hot or not according to what you feel at the time and then press the button to reveal the picture. The program i suggested is a freeware and its very convenient for viewing stored images very fast in slideshow mode even or one by one at your pace either alphabetically or randomly . Use the option it has to randomly select a picture (i think "enter" does the random and "space" the alphabetically) for the ones in the directory probing currently. Obviously if you put in that directory an equal number of 2 families of pictures you have your experimental setup ready and do not even need to author a program, but of course assuming you trust the vueprint's random generator. The way i suggested to do it it will only take you 1-2 hours at fast pace to check thousands of trials. Then report your results here. You would have started to convince yourself in doing so. Feel free to define yourself what a hot image is and select it from the internet appropriately (maybe usenet unless its illegal?) then add to that group random dull images of the same number. You could rename the files say from 000 to 999 , 500 and 500 or 100 and 100 whatever to eliminate any chance the computer chooses some starting letters more often than others or the possibility pictures of similar starting names have similar content. However that is not a problem itself really because after you have completed the test you can rerun the program this time just registering if an image is hot or not, without trying to predict it, to test if the computer program indeed serves 50-50.
That will be very fast to test as well (i mean to test its random generator fairness) .

So randomly give the files names (ie rename them unless you do the above fairness test in the end) to make sure there is no correlation, then use the "enter" button while viewing them to select a next picture randomly from the group. Now of course you can either imagine what it would be and then press it and reveal it (dull or not) or close your eyes, press it, decide what it is with closed eyes, then open eyes and see what the computer had selected. This way you test 2 things , guessing the computer choice that has already happened or guessing its choice that will happen after the guess.

On a side note i dont believe any of this will reveal any effect but of course i remain open minded to be convinced otherwise with proper methods though not complicated change along the way tests and test over test and basically somewhat "flexible" procedures that are not exactly transparent. Keep in mind these tests and news about them were out over 6 months ago or more years even. I would imagine there would be currently a worldwide fever if reproduction was already confirmed in several places. It would explode like the worse epidemic ever and would involved physics labs too. Why isnt that the case? Could it be that many tried (say 100 departments in different countries ) to replicate and only those lucky enough to register a high deviation reported say 2-3 of them? What is needed here is for him to invite several different scientists including mathematicians to reproduce together under supervision his tests with 100-1000 subjects even and see what happens. Wont take more than 1 week and the cost can easily be less than 20k , i bet he can find companies to sponsor it , just to settle the debate . He needs to describe to them how to do it but have others perform it and collect/process the data under supervision. If i had discovered something that major my most pressing concern would be to replicate it in front of others. I mean anything other than that shows serious lack of confidence and urgency to uncover the truth given how major the implication is for physics.


Additionally why does one even need to go to human brains? If indeed nature behaves as if some information about future events is available to present then the implications for Quantum Mechanics are hugely significant. If there exists an observer that has available information about future spin measurement of say a sequence of photons it should have ramifications to Bell type experiments it seems. If an observer knows the result before the collapse of the wave-function there is an element of reality to that value observed later. I believe experiments exist to test exactly that possibility (of the existence of hidden elements of reality) . Maybe it is not as simple as i make it and i need to go over it better (exactly what has been done i mean) but it would seem to me there is no need to go to human brains and all this complexity. Just pick up simple quantum systems and examine the ramifications of the existence of an observer that knows already the result of a measurement that is supposed to be random. I am pretty sure you will arrive at some major contradiction if you assume the existence of such an observer in terms of experiments already performed over the decades. But again lets think about it a bit further. I am confident you can use QM to counter such psi tests very effectively.

masque, can you please give cliffs on just the differences between your proposed protocol/software compared to the one in the study? I thought that one was much better than most people can program. Hardware based RNG's were also used for alot of these studies.

If I really wanted to replicate the experiment myself I can ask for the software and the same images provided by Dr. Bem.

For more on implication of psi in quantum mechanics see Dr. Radin's blog and his books.

http://deanradin.blogspot.com/

he tends to touch on entanglement and Bell's theorem quite a bit but I am not sure how I feel about that. I think it's fine to make parallels but I don't like the fact that he proposes lots of models some of which based on quantum physics. I prefer the classic "I don't know" that Bem opts to use when we really don't know.

I also think he claims your hypotheses have been tested and supported. haven't looked at or for research myself but will probably do so later. Radin has been immersed in this field alot longer than Bem so you might find him a little more biased but it's mainly based on alot of previous and personal research which would become clearer if you read "The Conscious Universe". I also suspect that he himself is highly sensitive to these phenomena (or at least he thinks so) since I can't fathom why someone as smart as he is would choose to perform a career suicide like this. I can see why Bem wouldn't mind risking it but I really do commend Radin for his commitment and testes.

One point of contention about your hypothesis, you assume that ESP is not a form of observing the wave function. I am not sure how I feel about this assumption. I feel that if you record precognition of a wave-function before it collapses and record the results, you have already influenced it in some way. I could be wrong though

Edit: in "The Limitless Mind" Targ describes an experiment that shows light behaving deferently as wave/particle in correlation to whether meditators are focusing on it through psi (clairvoyance in this case, not precognition). Obviously hardly a proof of anything given the nature of the citation but might shed some light on what would be expected in such psi experiments.

I thought i made clear how to do it without even programming and how to also test the random generator for fairness. You can use vueprint. You can install it within 5 minutes and its a cool little application to have anyway in your desktop. Its not a very advanced photo editor but its pretty simple and fast. I think if you follow my instructions you will reproduce his first experiment very easily using only yourself and spare time. He claims that people tend to select the hot picture more often because it produces a stronger response to their brains . Well if you use vueprint probing a directory of equal mix of hot and dull pictures , in the 2 ways i proposed (before or after the computer's random choice) and then register how often your prediction if the picture will be hot or not matches the real life result then you will see if you have started to predict things over 50% after thousands of trials (each one lasting maybe 1-5 seconds depending how fast you do it so its over in a few hours or you can do it casually over a week 30-60 min per day ) .

but he did test it for all these. did you read his paper? personally I am not interested in replicating it myself. I actually am pretty convinced about this, mainly because I have a much higher prior about the topic. I think most people would have a higher prior if they were as open-minded and informed about the history of parapsychology research.

Also, testing this on myself may not work since I am completely aware of how the experiment is supposed to work.
my main issue with your idea though, is that a negative result would really mean nothing. The study does not conclude that everyone possesses psi abilities. as a matter of fact it suggests that the ability varies across people so I may just not be good at it.


also, you might want to ask for an invitation to one of these if you're interested in the physics aspect of it all

http://www.stardrive.org/index.php?o...ence&Itemid=82

The official meta-analysis of replicates will be performed after december, 1st 2011 as I pointed out in op (posted a link). Only then we may see the world buzzing. currently having the scientific community buzzing is good enough imo

I assume you haven't watched this Harvard debate about the topic. posted it already itt

http://www.youtube.com/watch?feature...&v=0Tdiu5kwjKs

Yes but nothing beats the truth that exists in self confirmation efforts. I actually trust nobody that doesnt produce in front of my eyes evidence. Unfortunately this is only possible with experiments that i trust are transparent due to their own procedure and great number of involved members (ie CERN) (that is very hard to imagine colluding and who can never escape confirmation efforts) or with direct mathematical proofs or ideas that using math and prior results and hypotheses produce predictions that then others not me will verify independently. I really have very little confidence in people that do not expose in the most ridiculously detailed manner all the steps they took in testing something. Basically i do not trust any paper that simply presents results without the benefit of knowing how they selected the recorded data (was it the full raw data or did they filter out something, even if they tell me they didnt how do i know it) . Yes i take such narrated information as of some value but i assign it very small possibility to be honest unless it makes the most amazing conscious effort in the world to convince me the process is clean (maybe even with an unedited video link of the entire test) . Seeing the prediction one after another without any editing is the only proof that i need and one that is never provided in such papers that what is asked by me is to take the word of others as true.

In papers by Einstein for example i never had to take his word. He derived predictions and experiments verified it later. Or he derived conclusions that explained instantly other problems and offered insight that immediately proved creative because it attracted conviction instantly by its resulting utility (ie his derivation of Planck's black body radiation function or of mc^2 or of gravitational red shift etc ). I fail to get similar reaction by reading such posted papers. That doesn't mean they are lies but it means they are not by construction created in a way that is ultra transparent. The one testing a claim and the one making it must be independent people eventually. Otherwise science has become again story telling...

I appreciate your links and will spend some time to go over them and see for myself because i am not familiar with these scientists but realize that i am trained from physics to trust only people that their work is extremely convincing and eager to be exposed to the maximum to any tests possible as soon as humanly possible. It is the hallmark of a real discovery.

well then, finally a true skeptic. I have the same attitude althought I think of trust in terms of percentages. I can't say I even trust my own experiments 100% though. Even if I could see every piece of evidence. always room for paradigm shifts in my head

I think disqualifying knowledge because you haven't SEEN it or replicated it yourself might be problematic though. I prefer the following approach

“Believe nothing, no matter where you read it or who has said it, not even if I have said it, unless it agrees with your own reason and your own common sense.” – Buddha

how so?

http://www.colbertnation.com/the-col...rn---daryl-bem

hilarious Colbert section, commentary on this work and interview with Bem. Should put concerns about exploratory vs confirmatory to bed.

lol Hynman. he MAD

His experiments were clearly not designed to test a single hypothesis. Ray Charles can see that. If he knew what he were testing for, in advance, 1) he would be testing about only that, and his methodology would be very different, 2) he would be writing about, or overwhelmingly about, only that, and his statistical analysis would not include multiple partitions in every experiment.


You're arguing that he was doing this research- which takes a trivial amount of experimental time- so long ago that there were confirmation studies done in 2006, and none of this saw the light of day until recently. Unless it took him this long to find a way to mangle the data analysis to show something, which I guess is possible, a lag that long simply make no sense.

I told you exactly how it would- you find tons of "significant" patterns in anything if you look after the fact and partition enough, and they don't hold up better than chance going forward. That happens all the time in my line of work, and I've literally spent thousands of hours on that kind of analysis, teasing out the real from the noise to predict going forward. The evidence ITT isn't even close to convincing. It's marginal noise analyzed by somebody making beginner errors.

Ah, the donnicolo defense. Also, do you just hate money? Screen for a good psi-person, win your share of the free million from Randi, win your 100k from me, and retire early. Saying you're convinced this is real, but refusing to spend a tiny amount of effort to win 1.1 million dollars is utterly loltastic. You have to be a complete imbecile or intellectually dishonest (or both) to behave this way.

In psychology, you don't get to do your experiments every day. you do them every semester.

With the amount of experiments he has done and the fact that he probably had other things to do in normal psychology research, and since he was aiming for JPSP he knew he had to pull it off in one go so everything had to be right, perfect and airtight. it is to be expected that a study of this magnitude would take 8 years to publish. And who knows what kind of funds he had for this research. <-- see? making stories up is easy and it's all you do in all your arguments and you do it really badly.


For instance:
Bem, Wagenmaker's rebuttal:
Therefore the rest of your post falls apart.

as for your comments about the 1 mil. I have too much respect for myself to even consider trying to prove something to someone such as randi when the odds against him being convinced are about 1 million to one. There are many, many things I believe and know I can do that I have too much respect trying to prove them against those odds though. plus, evidence suggests psi is a subconscious process, which means it probably needs a laboratory set up and Randi's design is not conducive to that. proof, odds that Bem's experiments being due to chance are about a 100 billion to one shot, never mind the odds that a decorated psychology professor would decide to do these experiments and get funding purely based on no evidence with a lol-prior of near 0, still Randi thinks it's more likely that the experiment is due to luck. well good luck trying to prove ANYTHING to that guy.

"I have believed in PSI before, glad facts took the **** out of me." - plaaynde

LOL, no.



LOLOLOLOL. You obviously have too much respect for yourself to even know what the challenge IS. You don't have to convince randi. You have to demonstrate the ability, under controlled conditions, to some high degree of statistical unlikelihood (in fact, if I had to guess, he'd agree to a protocol for exp 1 similar to the 10k sample z=6 challenge I poster earlier). Randi still probably wouldn't believe you, but he'd pay you.

So I ask again, do you just hate money THAT much? Or are you intellectually dishonest, or an imbecile, or both?

LOL wtf are you talking about? You can get whatever setup you need.

you are really screwed in the head if you think the one million not being claimed is proof of anything. Randi is a liar who will never pay you. plus, I simply don't know how to prove many things against those odds all of which I know to be true . How do you prove mental processes? you'd realize the difficulties if you had any.

on this note, please stop posting in this thread. you're just being an idiot now

Well, besides this being continuous proof of you being unable to read... The fact that YOU are "convinced" (your words) that these things exist, yet YOU make no effort to claim the free million, means that YOU either hate money, YOU are intellectually dishonest, YOU are an imbecile, or some combination of the 3 (I'll go for 2 out of 3).

Horse****.

You, or your collection of people, hit 53% over 10000 trials, and get one mirrion dollars. It's quite easy... if they were actually true.