Much peer review and replication needed. Interesting.
Very interesting, replication is the key. Have printed the pdf to read at my leisure. (If I ever get any leisure time that is)
I was expecting something like Dr Who's "psychic paper".
The results will probably not be widely replicated. At least not by the scientific community.
Do you mean that the scientific community won't attempt to replicate them or won't succeed?Originally Posted by Croydon Bob
I was expecting something like Dr Who's "psychic paper".
The results will probably not be widely replicated. At least not by the scientific community.
I rather think Bob means that, assuming the scientific community are sufficiently impressed by this paper to attempt to replicate these results (not being a scientist, I have no idea how kosher the author's academic reputation really is), they won't succeed. Of course not - "sheep and goats effect" and all that.
Without having read this paper all the way through - as I said, I'm not a scientist, and life's just too darn short - I'm reminded of one of Rupert Sheldrake's experiments, which he and his followers claimed, probably in all sincerity, was a resounding success, but which nevertheless failed to impress the skeptics. It was a long time ago so I forget the precise details, and I can't give you a link (though I daresay there must be one - perhaps somebody could provide it?), but basically it went like this.
Sheldrake got a group of people who didn't speak Japanese and had them attempt to memorise a traditional lullaby which has been soothing Japanese babies to sleep for hundreds of years. His control group attempted to memorise a verse of the same length which sounded like Japanese but in fact meant nothing.
Sheldrake's deeply peculiar and pretty much all-embracing theory of morphic resonance predicted that, because the lullaby was firmly entrenched in the minds of millions of Japanese, it was intrinsically easier for other human beings to memorise than a load of nonsense somebody made up a couple of days ago. And sure enough, the lullaby proved easier for the subjects to recall more or less correctly than the nonsense rhyme, because we're all psychically connected to every other member of the human race, including all those dozy Japanese babies. QED.
But wait just a minute! Why does a lullaby catch on in the first place? Is it perhaps intrinsically pleasant, and therefore the kind of thing that sticks in your memory? An equally valid experiment would be to take some utterly bewildered tribal persons who have had no contact with civilisation, and see if they are better able to memorise the Beatles' greatest hits, or some meaningless songs I wrote last week and recorded with a few obscure session musicians who were doing their damndest to sound like the Beatles. I have very little doubt that a Stone Age bloke with a bone through his nose would somehow prefer "Please Please Me" to whatever half-assed Paul McCartney impression I attempted. And I have no doubt at all that Rupert Sheldrake would regard this as conclusive proof that his theory was correct.
So returning to the precognition experiment, do I take it that all the words the students were trying to memorise were equally memorable? And if so, how was this empirically determined? Was their success rate greater when the subsequently chosen word was "jam" rather than "syzygy"? As for the question regarding sexual positions, that's about as emotively loaded as anything possibly could be - absolutely anybody, except maybe an Anchorite eunuch, will automatically choose one far more often than the other. That doesn't of course allow them to somehow predict the future by mistake; but if there's even the slightest error in the experimental procedure, that kind of sloppy methodology will inflate it out of all proportion.
It is perhaps worth mentioning here that, since the only halfway scientific explanation for psi that doesn't involve totally unknown and totally undetectable forces and/or various tenets of Christianity and/or mangled mysticism of a vaguely Buddhist variety assumes that the Many Worlds interpretation of quantum physics is literally true, and the psychic is somehow able to choose which potential universe to be in and drag the rest of us along with him, surely a heavily loaded question like this would lead a fellow who likes taking girls from behind to pass the test by creating a reality in which a disproportionate number of the naughty pictures show that very thing?
Has anybody looked into this possibility? Are the poor ladies who posed for these pictures suffering obscure sexual nightmares in which, like that notorious and much misunderstood cat in the box, they're trying to do two pages of the Kama Sutra at once? Or have they been ruthlessly obliterated in favour of counter-Earth bimbos who happened to be the other way up?
I'm not entirely sure that this is the kind of experiment responsible scientists ought to be doing. If you got a real weirdo producing amazing results, of course you'd encourage him, and where would it end? Before you knew it, the entire population of the world would find themselves mysteriously attracted to bent spoons... (If anybody reading this got The Horn for "bent spoons", it's probably too late already.)
I look forward to having my arguments convincingly refuted by somebody who googled the word "Sheldrake" and cant spel to wel but knows he's a genius, therefore I'm wrong, and furthermore, I will spend my next 3,000 incarnations as something that survives by marinading itself in pus.
I don't really have time to go through the whole paper, but the headline one seems like nonsense.
Precognitive recognition of erotic stimuli:
100 volunteer (50m 50F).
820 photos rated on a 9 point scale for arousal
Erotic, negative and neutral.
40 sessions 12 erotic; 12 non erotic; 12 neutral
60 sessions erotic or nonerotic
Shown blank screens left and right on a computer, asked to predict which (left or right) would reveal the stated type of picture.
First trials negative - women seemed more likely to predict erotic but not men. So study modified to increase erotic explicitnedd for male subjects. how second trials re-run not stated, but lets assume rerun of orignial design.
Also note un stated number of "placebo' trials used to settle down subjects between real tests.
Result:
Where erotic images to be revealed 53.1% predictive accuracy. (12x40 + 60 x 18 = 1460 tests, 50% probability - 95% confidence interval 2.56; 99% 3.38)
Neutral images 49.1%; Negative 51.3%, Positive 49.4%, romantic non erotic 50.2%.
Heres the rub, we do not have one trial but 5 possible positive outcomes, added to this a rerun of the data after a failed previous attempt - so ten possible positive outcomes. Bayesian correction requires that we consider a significance value of (0.005) as the level of significance. Given the confidence intervals at the 0.01 level (+/- 3.38% around the 50% mark) the result is non signficant.
The statistics do not hold up, so the reviewers need to send this to someone with statistical knowledge.
Dalriada posted a link to this article on Facebook.
It's well above my head, but some members might find it interesting.Does psi exist? In a recent article, Dr. Bem conducted nine studies with
over a thousand participants in an attempt to demonstrate that future events
retroactively affect people’s responses. Here we discuss several limitations
of Bem’s experiments on psi; in particular, we show that the data analysis
was partly exploratory, and that one-sided p-values may overstate the
statistical evidence against the null hypothesis. We reanalyze Bem’s data
using a default Bayesian t-test and show that the evidence for psi is weak
to nonexistent. We argue that in order to convince a skeptical audience of a
controversial claim, one needs to conduct strictly confirmatory studies and
analyze the results with statistical tests that are conservative rather than
liberal. We conclude that Bem’s p-values do not indicate evidence in favor
of precognition; instead, they indicate that experimental psychologists need
to change the way they conduct their experiments and analyze their data.
http://www.ruudwetzels.com/articles/...setal_subm.pdf
Thanks quite a useful restatement of 'extraordinary claims require extraordinary evidence' The way one gets around this in practice, is that single centre exploratory results are neve regarded as confirmatory evidence in important advances. Only properly constructed multicentre trials, designed to confirm the magnitude of effect previously shown in exploratory trials can be accepted as confirmation.
But even taking the standard yard stick, this paper fails, since multiple possible outcomes were tested in parrallel and that which by chance seemed to back the authors predjudice was accepted as the post hoc primary endpoint.
Psychologists could do well to read this, as well as parapsychologists.
Although the authors say that "One can argue about the relative merits of classical t-tests versus Bayesian t-tests..." it's pretty clear they take a Bayesian approach. Indeed, it seems unlikely that problems 2 and 3 would have occurred if Dr Bem had used the Bayesian paradigm.
"Classical" NHST (Null Hypothesis Significance Testing) has attracted a lot of criticism, but applied statistics seems to be moving in the Bayesian direction.
Credit where credit is due. I just finished reading this paper and whilst I didn't analyse the statistics as others have done, the methodological flaws I was looking for didn't crop up.
One thing in particular impressed me. I was looking to see if the target was being predetermined to see if there was any way that the subject was reading the information from the present rather than the future. Bem avoided this possibility having the computer randomly select the target after the effect had been measured. However he went further than I would have expected. He pointed out that a computer's pseudo random number generator may be predetermined. So used a hardware RNG in some (but not all) of his trials. The reason for not using the true randomizer in all experiments is that if the psi hypothesis is entertained then we must equally assume that a positive result may be obtained by the subject influencing the "random" result as from predicting it. This of course is not possible if the "random" result is actually predetermined as part of a stream of pseudo random results produced from an initial seed.
If the effect is real (and of course that is a big IF) then repetition of the experiment should yield interesting results. The experiments seem to be simple to repeat. Such repetitions would of course not suffer from the criticism that the investigation was exploratory. There would still be reporting bias and but a halfway decent meta analysis should detect whether smaller trials yielded more positive results than larger trials.
I have asked for a copy of the replication package. So I can explore further.
Then we get to the most contentious part - the speculation at the end especially physics yes you guessed it: Quantum. We have quotes from Dean Radin but I think the following one is important enough to sum of the feelings of many here.
No Dean it isn't. I'd add that hanging a lampshade on this, does not entitle the hanger to wildly speculate on what quantum entanglement might be understood to explain at some future point in our journey towards understanding unless they wish to supply some empirical evidence.Quantum entanglement as presently understood in elementary atomic systems is, by itself, insufficient to explain psi.
Then Bells Theorem is mentioned. The proof of Bells theorem has two possible interpretations for realism and locality.
Realism is the theory that all these quantum values that quantum mechanics regards as probabilities do actually have real values at all time even if we can't know them.
Locality is the theory that an object is only influenced directly by its immediate surroundings.
Bells theorem is that they can't both be true.
If the realism holds then locality must be false and vice versa.
Bem seems to argue against both locality and realism. We don't know which argument Bell Theorum supports but it doesn't support both.
What are the details for getting hold of this, it sounds interesting?
On page 48 of the paper:-
So I dropped him an e-mail at the address on his homepage to take him up on the offer. Of course I'm not an academic and have been quite up front about my sceptical stance so he may well refuse.Replication packages are available on request for Macintosh and Windows-based
computers to encourage and facilitate replication of the experiments reported here.
Thanks. I'll give it a try if I'm not too busy.
Posting Permissions
Reply With Quote
Bookmarks