The popular website Slate.com just published a long article about research into paranormal phenomena. The piece by Daniel Engber is an example of the unfair and misleading treatment this topic gets in the mainstream press.
The article is entitled “Daryl Bem Proved ESP Is Real” and is subtitled “Which Means Science Is Broken.” It's a kind of a “trojan horse” title, because the article was clearly written to try to debunk Bem's research. The article discusses experimental research by Cornell University emeritus professor Daryl Bem. The research was published in a peer-reviewed scientific publication, the Journal of Personality and Social Psychology. The widely discussed paper was entitled, “Feeling the Future: Experimental Evidence for Anomalous Retroactive Influences on Cognition and Affect.”
In Experiment 1 described in the paper, subjects sat in front of a computer screen that displayed two images of a screen. The 100 subjects were told behind one of the screen was an image, and behind the other screen was nothing. The subjects were asked to guess which screen had the image behind it, during a series of trials running 20 minutes. When an erotic picture was used as the image behind the screen, subjects were able to guess correctly somewhat more often which screen had the image behind it. With erotic pictures, they guessed correctly 53% of the time, much more than the 50% expected by chance. With pictures that were not erotic, the subjects got results very close to the result expected by chance, 49.8%. Other similar experiments reported in the paper also got more statistically significant results.
Schematic depiction of ESP
Skeptics were outraged by these results, claiming they would never be replicated. But they were replicated. The meta-analysis here discusses many successful replications of Bem's surprising results. The meta-analysis discusses 90 experiments from 33 laboratories in 14 different countries. The analysis reported an overall effect of p=1.2 X 10-10. Roughly speaking, this means the results had a probability of about 1 in 10 billion. This is a very impressive result, showing statistical significance millions of times stronger than what is shown in typical papers reported by mainstream media. A typical paper that gets covered by the press will have an effect of only about p=.01 or p=.05.
In his article at slate.com mainstream media writer Daniel Engber clearly attempts to disparage and belittle this research finding. He uses some common techniques used by those denying or belittling evidence for the paranormal.
The first technique is to deprive the reader of the most relevant information the reader needs to decide on this manner. Engber makes no mention of the extremely compelling effect of p=1.2 X 10-10 reported in Bem's analysis. And so it almost always is when mainstream media reports on ESP experiments – typically a writer will report that a researcher reported “statistically significant” results without telling us that the degree of statistical significance reported is many times greater than what we find in typical scientific papers (which typically have statistical significance of only about p=0.5 or p=0.01).
The second technique Engber uses is a kind of isolation technique. His long article about ESP research gives no summary of the long history of ESP research dating back to the nineteenth century. The reader is almost kind of left with the impression that Daryl Bem's research was some weird fluke like a green sky appearing one day. But that's the opposite of the real situation. Scientists have been producing experiments showing very convincing evidence for ESP and other inexplicable human abilities for more than 100 years. The research started being done systematically about the time the Society for Psychical Research was founded in the late nineteenth century. One high point of the experimental research were the results of Joseph Rhine at Duke University during the 1930's, which actually produced results far more statistically significant than Bem's research. More recently, very compelling results for ESP were very often produced using a sensory deprivation technique called the ganzfeld technique. At the 2014 Parapsychological Assocation meeting, Diane Hennacy Powell MD presented extremely compelling evidence for ESP in an autistic child. Engber doesn't mention any of these things (see the table at the end of this post for the specifics).
Describing a moment in the past, perhaps about 1980, Engber tells us, very inaccurately, that “the laboratory evidence for ESP had begun to shrivel under careful scrutiny.” That is not at all correct. Classic ESP research such as Joseph Rhine's has never been successfully debunked. Before making this claim, Engber cites an example of what he apparently thinks is something that had discredited ESP research. It is the fact that James Randi hired associates to deceive some paranormal researchers, by acting as “fake psychics.” But while that incident reflected badly on Randi, it did nothing to discredit paranormal researchers, since they weren't the ones who were doing the faking.
Engber also throws in some “poisoning the well” rhetoric. He calls Bem's research results “crazy-land,” which is just vacuous disparagement. I may note that some of the most respected research results in scientific history (such as the double-slit experiments) were originally regarded as “crazy” results. Engber states:
Daryl Bem had seemed to prove that time can flow in two directions—that ESP is real. If you bought into those results, you’d be admitting that much of what you understood about the universe was wrong.
No, experimental results such as Bem's do not demand that we believe that “time can flow in two directions.” And it also is not true that results such as Bem's require people to believe that much of what they understood about the universe is wrong (although it is true that such research may suggest some people making overly dogmatic assumptions about the nature of time, consciousness and matter might need to reassess their assumptions, and admit their ignorance about such eternal questions).
Engber then goes to a long 8-paragraph discussion of a 2011 scientific paper by Simmons, Nelson, and Simonsohn. It's a paper entitled, “False-Positive Psychology: Undisclosed Flexibility in Data Collection and Analysis Allows Presenting Anything as Significant.” Engber refers to this as the “When I'm Sixty Four” paper. Engber then delivers this very inaccurate statement:
But Simmons, Nelson, and Simonsohn revealed that Bem’s ESP paper was not a matter of poor judgment—or not merely that—but one of flawed mechanics....They’d shown that anyone could be a Daryl Bem, and any study could end up as a smoking pile of debris.
This statement by Engber is hogwash and baloney. In fact, the 2011 paper by Simmons, Nelson, and Simonsohn made no reference at all to Bem's research and made no reference to ESP or any research on paranormal phenomena. Their paper showed how someone might through dubious methods produce a borderline statistical significance along the lines of p=.05 or p= .01. But the paper did nothing to suggest that such dubious methods were used by Bem or any other ESP researcher. In fact, the level of significance claimed by Bem in his meta-analysis is p=1.2 X 10-10. That's a level of significance a hundred million times greater than merely p= .01.
Engber has made a totally inaccurate claim that Simmons, Nelson, and Simonsohn debunked Bem's research, when their paper doesn't even mention Bem or ESP research. What he also fails to tell us is that if Simmons, Nelson, and Simonsohn's paper can be claimed to debunk anything, then it debunks all experimental research claiming results with a significance of around p= .01 or p=.05 – which is basically a large fraction of all research published in neuroscience, medicine, psychology, and physics.
For Engber to claim some great significance for the paper of Simmons, Nelson, and Simonsohn is rather absurd, since the paper basically tells us that we can get false alarms when the results have a significance of around p= .01 or p=.05 – which is something everyone already knew before the paper was written.
Engber's next trick is to start talking about replication failures and research fraud. He says this:
A few months later came the revelation that a classic finding in the field of social priming had failed to replicate. Soon after that, it was revealed that the prominent social psychologist Diederik Stapel had engaged in rampant fraud. Further replication failures and new examples of research fraud continued to accumulate through the following year.
Clearly, Engber is trying to bring into his reader's mind the idea of ESP research being fraudulent. But, in fact, all of the examples cited above refer to work that was being done by researchers working on topics other than ESP or paranormal phenomena. Mentioning such things in the middle of an article on ESP research is not literally inaccurate, but it's extremely misleading, something very much prone to create a false idea in the reader's mind. In fact, ESP researchers have an excellent record of honesty, as good as that of any group of scientific experimenters. Engber does not cite any example of dishonesty by an ESP researcher – he simply deviously leaves his readers with an impression of such a thing.
I can imagine a writer writing in an equally misleading manner, giving some web links that do not actually refer to candidate John Doe, but to some other people. It might go like this:
Doubts have been raised about the candidacy of John Doe. There was a drunk driving arrest (Link 1). Then there was a bank robbery arrest (Link 2). Then there was a grand jury indictment (Link 3). Clearly we must question John Doe's fitness for office.
Of course, if none of these links referred to John Doe, but referred to other people, such a paragraph would be very misleading.
Towards the end of his piece, Engber tries to throw doubt on Bem's findings by referring to a replication attempt to repeat one of the several experiments in his original “Feeling the Future” paper. Engber claims that at the most recent meeting of the Parapsychological Association (which at this date would have been the 2016 meeting), Bem presented a “pre-registered analysis” which “showed no evidence at all for ESP.” Engber claims that nonetheless Bem's abstract of this work, after “adding in a new set of statistical tests,” stated that the replication attempt had produced “highly significant” evidence for ESP. Clearly Engber is trying to suggest that maybe some kind of statistical funny business was going on. But he presents no specific facts to back up such a claim, nor does he give a link that would allow us to check out his insinuations. When I go to the web page that lists the abstracts submitted to the 2016 meeting of the Parapsychological Association, I see no abstracts authored by Bem. The experiment referred to is not even the main experiment by Bem that received the most attention in the press (the experiment I described above). All in all, this does nothing to raise doubts about Bem, but may raise further questions about Engber's hatchet tactics when dealing with the paranormal.
In this case Bem is an ace Ivy League psychologist with several decades of statistical research experience, and Engber is neither a mathematician nor a scientist. So if Engber is going to raise doubts on statistical grounds, no one should pay attention to him unless he is very specific in what his precise objections are. A little vague rhetorical doubt-sprinkling does not suffice. In his long article, Engber does not actually provide a single specific well-documented statistical or methodological reason for doubting Bem's research.
Engber's misleading article is typical of the dismal coverage that ESP research gets in mainstream media. A great deal of this coverage is inaccurate. Astonishingly, it has become “politically correct” within today's science culture to make completely false statements about research into the paranormal. It is extremely common for scientists to claim there is no evidence for extrasensory perception, which is entirely false, given the very large body of convincing experimental evidence that has been gathered over more than 100 years, much of it under the auspices of major universities or the US government (see the table below for examples). It is also very common for scientists to say that the experimental evidence for ESP has been debunked or that it was never replicated. Neither of these statements is true. Shockingly, we have a science culture in which highly inaccurate statements on parapsychology research are pretty much the norm. It's rather like the situation we would have if it become popular in American history departments for professors to say that the Americans won the Vietnam War, and that they treated the Vietnamese very nicely while doing it.
Engber's article is entitled “Daryl Bem Proved ESP Is Real” and is subtitled “Which Means Science Is Broken.” A more accurate title for an article would be, “Joseph Rhine and Many Other Researchers Showed ESP Is Very Probably Real, But Science Culture Refused to Accept It, Which Shows Science Culture Is Broken.”
Bem's research was purely experimental. There is an entirely separate reason for thinking that humans can sometimes sense the future in a paranormal way: the large body of episodic accounts supporting such a claim. I will discuss this fascinating evidence in my next post.
Below is a table showing some high points of research into ESP.
Below is a table showing some high points of research into ESP.
|Professor Bernard F. Riess, Hunter College, 1937||Remote card guessing of 1850 cards with woman in another building||73% accuracy rate with expected accuracy rate of 20%||http://futureandcosmos.blogspot.com/2016/02/better-than-smoking-gun-riess-esp-test.html|
|Professor Joseph Rhine and others, Duke University, 1932||Card guessing experiments with Hubert Pearce, 10300 cards, experimenter and subject in same room||36% accuracy rate with expected accuracy of 20%||http://futureandcosmos.blogspot.com/2014/12/when-rhine-and-pearce-got-smoking-gun.html|
|J.G. Pratt, Duke University, 1933-1934||Card guessing experiments with Hubert Pearce, 1850 cards, experimenter and subject in different rooms||30% accuracy rate with expected accuracy of 20%||http://psychicinvestigator.com/demo/ESPdoc.htm|
|Ganzfeld ESP tests, 1997-2008||ESP tests under sensory deprivation, various subjects, 1498 trials||32% accuracy rate, with expected accuracy of 25%||http://www.deanradin.com/FOC2014/Storm2010MetaFreeResp.pdf|
|Rupert Sheldrake, PhD, 2014||63 subjects, 570 trials, test of whether subject could correctly guess a phone caller||40% accuracy rate with expected accuracy of 25%||http://www.sheldrake.org/files/pdfs/papers/ISLIS_Vol32.pdf|
|Rupert Sheldrake, PhD, 2014||50 subjects, 552 trials, test of whether subject could correctly guess who sent an e-mail||43% accuracy rate with expected accuracy of 25%||http://www.sheldrake.org/files/pdfs/papers/ISLIS_Vol32.pdf|
|Diane Hennacy Powell MD, 2014,||ESP tests with an autistic child||100% accuracy on three out of twenty image descriptions containing up to nine letters each, 60 to 100% accuracy on all three of the five-letter nonsense words, and 100% accuracy on two random numbers: one eight digits and the other nine. Data from the second session with Therapist A includes 100% accuracy on six out of twelve equations with 15 to 19 digits each, 100% accuracy on seven out of 20 image descriptions containing up to six letters, and between 81 to 100% accuracy on sentences of between 18 and 35 letters. Data from the session with Therapist B showed 100% accuracy with five out of twenty random numbers up to six digits in length, and 100% accuracy with five out of twelve image descriptions containing up to six letters.||http://dianehennacypowell.com/evidence-telepathy-nonverbal-autistic-child/|