I had been thinking about writing this post for awhile, and I begin to fight my standard inertia bit more when I saw this article on Slate, dated May 17, 2017: "Daryl Bem Proved ESP Is Real. Which means science is broken."

Then, in the comments on this blog, Leo MacDonald referred to a related post on NEUROLOGICAblog: "Follow Up on Bem's Psi Research."

We'll be talking about these in a moment, but first my thesis:

Skeptics will never be compelled to accept the existence of psi because laboratory research involves difficult statistics that can be argued about ad infinitum, and exceptional individual cases of psi can be dismissed as "anecdotes" one by one.

Let's look at both of these issues in turn.

Psi laboratory research

Here's the money quote in the Slate article: "Bem had shown that even a smart and rigorous scientist could cart himself to crazyland, just by following the rules of the road." Yes, folks, you're cray cray if you believe in psi.

For the most part, researching psi under laboratory conditions means testing a lot of people and seeing if an effect is produced that cannot be explained by chance. One classic experiment is to have people guess which of four symbols is the target. The success rate across a reasonably large sample size should be 25%. If it is much higher than that, then one can in theory call it evidence of psi. One series of lab experiments that many believe showed evidence of psi is the ganzfeld experiments (Wikipedia is extremely biased against psi, but this will give you an intro to the facts at hand).

Skeptics can attack the design and conditions of such experiments, and they do so regularly, but there is a deeper issue at hand: there's always the chance that the results of the psi experiment are due to, well, chance.

In a wide variety of fields, research is performed with the aim of achieving a p-value of .05. That means that there is a 5% chance (p = .05 = 5%) that the positive result achieved is due to chance. Experimental results within this predetermined p-value are often called "statistically significant." Whether it's medical research or the social sciences, p ≤ .05 is typically considered pretty good.

So, just do your psi lab research, get p ≤ .05, and you've proved psi, right? The world is changed forever, right?! Of course not. Because p = .05 is actually pretty lousy. It means there is a 1 in 20 chance that the experiment is, in effect, meaningless. I personally would not change my belief system based upon such a paltry stat, and I wouldn't expect anyone else to do so, either.

But that's fine, since we can get that p-value even lower, right? Eventually they'll have to believe! Well, yes and no. One typically gets the p-value lower by increasing the sample size. The nature of statistics is such that, say, cutting that p-value in half can't be done with double the sample size; the relationship is not linear, and for very small values of p, extremely large samples are needed.

Thus, getting p = .0001 (1 in 10,000 chance of being due to chance) would require a monumental sample size for one's psi experiment, and even then Skeptics could more or less correctly say, "I'm not going to change my worldview based on this evidence when there is still a 1 in 10,000 chance that it's nothing at all. Extraordinary claims require extraordinary evidence, and that still isn't extraordinary enough!"

So, we just have to give up, right? Not so fast! There's a thing called meta-analysis, in which we get that sample size and p-value to where we want them by combining lots of different but related experiments. Sometimes you will read about a meta-analysis of psi experiments that claim very small p-values indeed, and someone may say something along the lines of, "The odds of this meta-analysis result being due to chance are over 3 trillion to 1!"

That's great, but here's the rub: meta-analysis is quite complicated. Different experiments using different methodologies need to be combined, and how that should be done is to some extent a matter of opinion and judgment. The calculations themselves require expert knowledge—this is far from Statistics 101 material—and the experts themselves can disagree on both the methods used and the meaning of the results. And that's exactly what happens. The pro-psi statistician argues with the anti-psi statistician, and only a tiny fraction of the population on either side is qualified to assess what they are saying. It ends up a lot like the climate change argument: "Trust our smart guys. They're really smart and there are a lot of them and those who disagree ought to be ashamed of themselves." When you hear laboratory research on psi has been "debunked," that's typically what has happened: some credentialed person has gone over the numbers and declared them unconvincing. Or some media Skeptic has declared the methods and conditions of the experiment unacceptable. Either or both.

Thus, to recap, any given psi experiment is not enough to sway opinion, and trying to combine multiple experiments leads to arguments over recondite statistics. There simply is no way around it.

Going back to the Slate article, Daryl Bem performed a variety of experiments under rigorous conditions and came up with what appeared to be statistically significant results. Instead of merely saying that he was wrong (which of course they do), scientists and other members of the cognoscenti have pointed out that, yeah, the kind of statistical reasoning that has held sway in many fields for over a century is actually pretty lame. Hence the subtitle of the article: "Which means science is broken."

Hearing this kind of thing is gratifying to an extent, as we proponents of psi have been pointing out the double standard for a long time: "How can you call things like psychology and sociology science when the designs of their experiments are a joke compared to what parapsychology has going on now, and they're often satisfied with p = .10 or worse?" Cognitive dissonance can take quite a long time to resolve itself, but it seems that Skeptics and members of the scientific elite will not be moving forward by recognizing psi (duh) but instead by trashing the social sciences. For example, here's an LA Times article from 2012: "Why psychology isn't science." And now the reaction to Bem.

Psi laboratory research: my take

When I discussed the article on Bem with Michael, he said:
I read it. Mixed feelings. On one hand, it would be easy to dismiss the whole thing as an example of knee-jerk skepticism that would rather reject established scientific methods than accept any evidence for the paranormal.

But on the other hand, they may have a legitimate point in saying that these methods are faulty, not only when practiced by parapsychologists but by conventional scientists, as well. I suspect there is a lot of bad science out there, and that statistical analysis can be badly flawed.
I think this is exactly correct. I'm glad to see some consistency finally being applied, and the reality is that it's difficult to prove things like drug efficacy, human behavioral traits, the existence of social trends, and psi effects using statistics.

Eric Newhill had this to say:
The effects in any single experiment are just statistical probabilities. Usually the stat significance is not that strong. Skeptics like to bring up the file drawer effect (studies that showed no stat significance are filed away, not published, not talked about, etc.).

Your comments on meta-analysis are spot on ... very complicated to do it right. Much opportunity for disagreement even among experts and, IMO, should only be done when the different experiments are using the same design (though at least small differences can always be found and raised as confounds).
Eric brings up a good point about the so-called file drawer effect. With respect to a given meta-analysis, Skeptics will claim that the pro-psi side is hiding negative results that would otherwise weaken the effect of the meta-analysis. One response to this is ask whether the number of negative experiments required to cause such a reversion to a theoretical mean could even have been performed. The pro-psi side will say that hundreds of such experiments would be required, and thus the meta-analysis is correct. And on and on the arguments go.

Eric also raised this interesting point:
One last thing, and is usually passed over by even the best researchers. I have read some of the details of some of these experiments. The p-value for the cohort may be a little significant (or not), but there are often individuals that score very very high, well beyond chance. Whereas there are usually not individuals in the cohort that score very very low. This is meaningful in itself. I have always thought that a series of experimental replications should be run, and the individuals in each cohort scoring the highest should be put together into a super cohort and tested. Basically, take the most talented people, those showing unique psi abilities, and focus on them only. This should demonstrate psi well beyond chance and should produce results that raise the eyebrows of many Skeptics.
Right, some of the individual results are mind-blowing, but these become invisible within the larger statistical picture of the experiment.

Personally, I have read a decent amount about psi laboratory research, and I think the experiments, in the aggregate, do provide substantial evidence that psi exists. Yet I most likely would not be convinced by that evidence alone. Rather, it dovetails nicely with exceptional individual cases of psi and my own experience of psi. Nothing in the laboratory research contradicts this other evidence but is congruent with it. It is quite similar to how various sources of information on the Afterlife (NDEs, ADCs, channelings) match up very well.

Exceptional individual cases of psi

This is much less complicated to talk about. In his original comment on this blog referring to the NEUROLOGICA blog, Leo MacDonald said:
I personally believe the stronger cases comes from the so-called anecdotal claims.
Michael, in the conversation cited above said:
But the really convincing cases are the prodigies who score incredible remote viewing hits or have an obvious, dramatic telepathic link—like the famous ESP experiments carried out by Upton Sinclair and his wife. These cases don't rely on statistical analysis, with its inherent danger of data mining.
And in the comments of the NEUROLOGICA blog post, Ian Wardell (who also comments on this blog), said:
I'm not really interested in such [laboratory] research though. I'm more interested in the spontaneous cases of psi.
Indeed. These cases are compelling. They really leave no escape.

We talk about exceptional cases of psi on this blog all the time. They involve the transmission of information that simply could not have been obtained except through psi, as well as the manipulation of matter itself. When reading the reports, one is left with two choices: believe that psi has been proven, or believe that highly intelligent, observant, and respectable researchers were fooled (as by stage magic) or were themselves simply lying. To provide two examples we discussed here before, Eusapia Palladino and D.D. Home produced or mediated extreme physical phenomena in adequate lighting conditions right in front of the eyes of researchers. These phenomena cannot conceivably be dismissed as "tricks"; the observers could not possibly have been fooled. Either what they say happened actually happened, or they were lying.

But that's OK, the Skeptics have the classic strategy of deny, deny, deny. It doesn't matter who saw or heard what. Their approach to psi is the approach they use on any paranormal phenomenon: deny. A séance that produced amazing information? Cold and/or hot reading and/or lucky guesses. Deny. An NDE that produced veridical information? Explain it away. Deny. Physical phenomena captured as photographs or video? Faked! Deny.

It's really that simple. They think they can "divide and conquer" by simply denying each and every paranormal report in the history of planet Earth. Each of these is merely an "anecdote," and the plural of "anecdote" is not "data"! They think that such an epistemological stance is self-evident instead of self-serving. They are untroubled by the double standard in which scientists who work in the materialist paradigm are capable of "observations" of individual cases, while those who are outside this paradigm are only capable of "anecdotes." They find it reasonable to posit that people throughout history and in every location of the globe have reported psi phenomena of a consistent nature, yet they have all been mistaken. (I personally find it maddening that they dismiss this vast evidence from history yet remain incurious as to what caused such consistent mistakenness. Nothing to see here folks, move along ...)

On rare occasions, a major Skeptic will have an experience that changes his or her thinking on the paranormal. Such happened to Michael Shermer. For his whole life, he was able to dismiss and denigrate the whole world's paranormal experiences, but once something amazing happened to him, well, that's totally different! And of course, the reaction of every "believer" to this conversion is smh...

Exceptional individual cases of psi: my take

What I find interesting is that Skeptics will say that they have never experienced psi or anything paranormal in their entire lives, nor do they know anyone who has had such an experience. This is the furthest thing from my own world.

My guess is that, when most Skeptics imagine a psychic reading, they see a flaky if not creepy psychic sitting in a dimly lit den of fraud, receiving money with greedy fingers, then proceeding to cold- and hot-read the victim and pump him or her full of generalities, superfluities, and, of course, lies and false hope.

In contrast, I am a psychic with many psychic friends. I've given readings and gotten more than a few big hits. I've received readings and have witnessed more than a few big hits. To us, it's nothing unusual, odd, or spooky. We trade psychic advice on virtually a daily basis, in fact. No special setting or mood is required; in fact, I give and receive most readings over Facebook these days. Further, I make no money off of psi at all (I give readings for free on a frequent basis, actually). About half of my psychic friends do charge for readings or other psi abilities, but they do a lot of pro bono as well, and absolutely no one is getting rich from these services. I can also observe that my psychic friends are extremely normal and down-to-earth, and none of them fits the stereotype of the New Age flake (OK, we mostly don't fit that stereotype!). I can assert without equivocation that I have never heard a friend refer to doing anything psychic in a fraudulent or less than sincere way.

In short, psi works for us consistently and on certain occasions amazingly. What incentive would we have to make it a part of our lives if it didn't? I'm not naïve: Skeptics could certainly cite a range of potential psychological and sociological causes for such experiences. Those outside of our world are free to observe and judge for themselves. But my point is that psi isn't just about the extreme and the strange. It can be an ordinary and consistently present part of one's life.

Further, as the comments on Michael's recent and excellent posts on Leslie Flint demonstrate, we who believe don't believe everything (pace the Skeptics who enjoy portraying as credulous idiots. Crazy, too.). Yet, on the whole, I find reports of exceptional individual cases of psi to jibe with my own experience, and I think they are strong evidence for psi.

In conclusion

Based on my reading of psi laboratory research and exceptional cases of psi, as well as my own experience, I am 100% certain that psi is real and materialism is completely disproved and an obsolete worldview. Skeptics, however, will never be convinced.

If I am correct, that puts the unstoppable force of the truth of psi against the immovable object of the Skeptics' belief system. What will be the end result of such an interaction?

I am going to go with a prediction that I have heard elsewhere and found convincing: Individual people and society as a whole will be convinced of the existence of the paranormal once someone is using it to make money. I don't mean for readings but in the form of a process, product, or service that consistently works and that people want. At that point, the Skeptics will be forced to move, since money talks, and you-know-what walks. My guess is that 100 years will not pass without this happening, and it will probably happen much sooner than that.

Comment: Unless, of course, there's something about the nature of psi that intrinsically works against such a goal and strategy. In which case: the skeptics will never accept it, ever.

(End note from Matt: Eric Newhill, who is a frequent commenter here and an actuary and analyst at a large insurance firm, was kind enough to review this post for accuracy on statistical matters. Also, Michael always provides significant guidance on my guest posts. Much thanks to you both!)