The Psi Effect

Empiricism vs Rationality

The placebo effect is really weird.

No one, or virtually no one, bothers to question the existence of the effect. It’s too well attested by medical practitioners, too inescapable a fact of life to be seriously disputed. Even critical reports, written to minimize the hype around placebos, admit that “When administered in a blinded fashion, a placebo will provide a small effect.”

And such a wording is still dubious. In many cases, the effect size is large, or larger even than the independent effect of a specific treatment. For instance, a recent meta-analysis found that the effect of placebos on depression was d = 1.69. For non-statisticians, this is usually described as a “very large” effect. Even more striking is the fact that the actual non-placebo drugs had an effect size of d = 2.50, implying that, if you’re taking drugs to fight depression, two-thirds of the effects are from placebo, and only the remainder is accounted for by the actual chemicals in the pill.¹

So placebos are extremely interesting. Modern medical science is amazing, but when you really internalize the meaning of the placebo effect, it’s not much of a logical leap to conclude that we’ve actually had effective medicine even before we had effective medicine. Consider prehistoric foragers wounded falling down ridges or poisoned by berries, going to visit to medicine men to improve their spirit. Consider ancient egyptians stung by scorpions or suffering from heatstroke, and then seeing a priest to invoke the names of the gods on their behalf. We talk about these treatments being time-consuming and useless, but even if all they manage to accomplish is a plausible placebo effect, they work anyway. Right? For thousands of years people have been subjecting themselves to bizzare, complex, and totally unscientific rituals, and then doing better just because they thought “Yikes this tastes terrible, and just look at the medicine man’s expression! This treatment must be really good.”

Yet the mechanism by which a placebo acts is not exactly well explained. “Something something psychological expectations affecting physiological outcomes something?” Yeah OK, this it seems as though it might just be plausible, as opposed to the alternative explanation, which goes something like “You control your body’s ability to heal and perform well unconsciously, just by believing stuff,” and sounds totally nuts.

But isn’t it interesting that the physiological pathways which would need to be activated for placebos to fight lung cancer,² revive nerve pathways weakened by Parkinson’s disease,³ reduce depression,⁴ or perform well in sports and exercise⁵ are individually complex enough, and collectively diverse enough, that we might well wonder whether there aren’t actually four placebo effects? Maybe more?

So like I said, the placebo effect is really weird. It’s definitely not what we would expect on rational grounds; and I don’t really believe in it because I’ve been able to verify it myself. It’s just that empirical findings—not just from single studies, but numerous meta analyses—are reported so consistently that, I think, we’d be foolish not to believe in it. If we wanted to deny the placebo effect exists, we’d pretty much have to insist that all the scientists publishing on the placebo effect over the past hundred years or so were all lying.

Personally, I don’t much go in for conspiracy theories. They can be true. Sure, I guess if push comes to shove, people do secretly conspire together. But this isn’t my go-to explanation when I encounter something surprising or counterintuitive in science. Long-time readers will know I love quantum mechanics; I’ve never independently verified the placebo effect any more than I’ve verified the findings of quantum mechanics. But when multiple scientists in different countries report a very consistent effect across hundreds of studies, it makes most sense to trust them, at least tentatively, because the alternative makes us look dogmatic or like some sort of mental patient.

This is also the way I look at parapsychology and the psi effect. And not just sort of, or a little bit. This is exactly the way I look at parapsychology and the psi effect.

The Psi Effect

For those who aren’t obsessed with obscure research paradigms, parapsychology is a discipline dedicated to studying alleged paranormal phenomena. Although there’s plenty of work we could discuss on precognition and psychokinesis, I want to focus on what is broadly agreed upon—namely, the parapsychological findings on psi.

Psi is defined as an anomalous transfer of information. Better-than-chance accuracy on attempts at guessing what happenned to your friend on the way to your apartment isn’t psi; you know them well, and cues like texts you received (or didn’t receive) could give things away. Better-than-chance accuracy on attempts to guess the suit of a card someone just drew in a game of poker isn’t psi, since you could maybe have seen a reflection of the card in their glasses, or maybe you could draw an inference from the cards in your own hand, or maybe you could have read some information from their expression or the way they bet. Psi has to be totally and completely anomalous—unexplainable by normal means—to qualify as psi.

A common test of psi is to take a subject, sit them down, cover their ears with headphones playing white noise, put half-ping-pong-balls on their eyes, shine a light on their face, leave them in this state while their friend watches a video in another building. Then you bring them out show them the video along with three other random clips, and ask them to choose the correct video. This is called a ganzfeld experiment, and better-than-chance accuracy for something like this qualifies as psi.

From wikipedia. Ganzfeld experiments look like fun, if your idea of fun is a vaguely dissociative experience under sensory deprivation

Traditionally in ganzfeld experiments, there are four videos to choose from. Obviously then, the chance hit rate is 25%. According to a vast array of ganzfeld studies, the average hit rate is sigificiantly above 25%.

When I was getting ready to write this article, I had to restrain myself from giving a thorough review of research establishing that psi really does seem to exist, detailing the limitations on the effect and our ability to trust the solidity of the results, and showing that the ganzfeld studies are about as rigorous as anything we might ever hope to see in science. I’ll give that article as a follow up for anyone who wants to see it. (Edit: It’s right here.)

But frankly, I’m not writing this post for people who don’t know what the results are. I’m writing for those who do know, because they’re broadly members of the rationalist community—people like like Eliezer Yudkowsky and Scott Alexander, the latter of whom wrote, in The Control Group is out of Control:

Parapsychologists are able to produce experimental evidence for psychic phenomena about as easily as normal scientists are able to produce such evidence for normal, non-psychic phenomena…

You might remember Bem as the prestigious establishment psychologist who decided to try his hand at parapsychology and to his and everyone else’s surprise got positive results. Everyone had a lot of criticisms, some of which were very very good, and the study failed replication several times. Case closed, right?

Earlier this month Bem came back with a meta-analysis of ninety replications from tens of thousands of participants in thirty three laboratories in fourteen countries confirming his original finding, p < 1.2 * -10¹⁰, Bayes factor 7.4 * 10⁹, funnel plot beautifully symmetrical, p-hacking curve nice and right-skewed, Orwin fail-safe n of 559, et cetera, et cetera, et cetera…

This is far better than the average meta-analysis. Bem has always been pretty careful and this is no exception. Yet its conclusion is that psychic powers exist.

OK, so what is the takeaway for any sensible person, then? How should we interpret the findings of empirical methods?

Science says subatomic particles can’t be measured in both position and velocity to arbitrary precision, but can tunnel through insurmountable barriers, are shown statistically to become entangled over macroscopic scales, et cetera, et cetera, et cetera. Should we stubbornly refuse to believe those things because they don’t make sense? When I found out about these things as a physics undergrad, I was pretty suspicious at first, because they obviously contradict what seems to make sense from intuitions built up over years of living on Earth and working on regular Newtonian physics. But the studies were all there; it was settled before my time.

Science says a sugar pill labelled like a prescription drug acts like the real drug so long as the person who takes it thinks it does. Should we smile and congratulate ourselves on not being taken in by such obvious nonsense? When I found out about the placebo effect as a kid, I privately suspected that eventually it would turn out not to be true, or to be true only for a few things like pain relief which are mostly subjective anyway. But then, I found out about the giant pile of research showing the placebo effect is real, so OK, the placebo effect is real.

Science also says it’s possible to transfer information in some way that we don’t currently understand, but popular opinion manifested on Wikipedia says we can’t trust the studies, so never mind, forget the scientists and just believe what you’re told by everyone else?

And then Scott Alexander does just that, and he does it so hard that he literally tells us all that what we can do is just let personal opinion trump everything else, and creates a helpful meme to demonstrate, concluding:

Science! YOU WERE THE CHOSEN ONE! It was said that you would destroy reliance on biased experts, not join them! Bring balance to epistemology, not leave it in darkness!

Scott clarified in the comment thread that he was pretty willing to accept “they got a significant result by coincidence,” even though the chance for this was, by his admission, a few orders of magnitude less than 1%. Eliezer Yudkowsky even made an appearance to share his considered opinion that meta-analyses in general are “bullshit”, saying “they can deliver whatever answer the analyst likes.”

This is a point well taken! I grew up with fundamentalist Christians, after all—people who very regularly dismissed scientists like Charles Darwin as biased frauds—so I’m very familiar with criticisms of scientific endeavors that deliver whatever the analyst likes. And what I take away from this is that being a rationalist and a part of the rationalist movement really just means being dogmatic (but instead of doing it in between Bible verses in church, you mostly do it on the Internet).

Psi Says More About Reality Than About Things Like Telepathy

Part of the problem, I suspect, is that skeptics don’t think that parapsychology is really the same kind of thing as relativistic or quantum physics. When I talk about parapsychology and the psi effect, I’m not trying to say “I have psychic powers.” I’m trying to say exactly the same kind of thing that I mean when I talk about cosmology or morality—I’m trying to say something about metaphysics, about the fundamental nature of reality, about kind of universe we live in.

I’m trying to say, basically, “It looks like we live in a universe that doesn’t admit to simplistic descriptions.”

Many believers in psi complain that skeptics are wedded to materialism. But talking to skeptics my sense is not that, (or not only that) they are materialists. It’s that they are really, really wedded to a specific interpretation of materialism that expressly forbids certain things like psi from being possible. For what it’s worth, their model of reality works pretty well for them. And be fair—even people like me who come down in favor of the parapsychologists know that lots of people who talk about anomalous experiences are over-credulous, mushy-headed, or desperate for attention. We already know that kind of person exists. So, claims regarding the existence of parapsychological phenomena have a neat and tidy explanation.

The trouble comes in when the neat and tidy explanation, well-suited to explaining the way your aunt Beatrice keeps changing the discussion at Thanksgiving to her psychic premonitions, is applied to a vast body of data gathered by large numbers of educated scientists carrying out meticulous replications of one anothers’ experiments over the course of a few decades. At that point it really becomes obvious that many skeptics have to apply that explanation, even when it starts to seem a poor fit, because they literally cannot tolerate the alternative—their basic understanding of reality can’t allow it.

But there are many ways of thinking about reality, and not all of them are incompatible with psi. Just a few I can think of include (with my best attempt at labeling them as impartially as I can):

• Straightforward Materialism: Stuff is made of matter. Matter is made of atoms, which can be measured to arbitrary precision, have clear properties like size and location, and behave in concrete, deterministic ways. Discredited by Modern Physics circa 100 years ago

• Conservative Materialism: The universe is governed by some strange physics, but these effects are now well known, and no new phenomena not seen before about the turn of the 20th century will be discovered. Compatible with Modern Physics, but incompatible with frontiers. Unified Field Theory does not exist. Dark Matter & Dark Energy remain unexplained

• Open-Ended Materialism: There are no spirits, but it’s a strange universe out there.

• Monotheism: Reality is material, but it is governed by God, who created everything. Humans have a non-material soul, and rare miracles are attributable to God.

• Simulationism: Reality is a simulation created by some being or beings in the real world (or at least a higher-level of reality). Miracles are possible, attributable to those in charge playing with the simulation. Note the overwhelming similarity with conventional monotheism.

• Animism: Reality is richly infused with mystical, magical, spiritual, or other supernatural properties. Trees and rocks have spirits. Common to prehistoric societies, and some modern Wiccans.

• Spiritualism: The material world is merely a thin outer layer floating atop a spiritual ocean. This world is not really even real—the spirit world is real.

• Solipsism: I’m the only thing that exists. Everything and everyone else in what I experience as reality is merely a part of my dream.

• Mathematicianism: What we call reality is, at base, a complex mathematical structure.

• Agnosticism: I’m not even sure what reality is, at base. This is my own position (but I can at least rule out a couple options on this list)

Parapsychological discoveries are compatible with all of these views except the first two, which I’m calling straightforward materialism and conservative materialism. Assuming the key studies in parapsychology weren’t fraudulent—and I’ll stress that yes, they may be fraudulent, though that’s not my go-to explanation—those first two worldviews are the only conceptions of reality that make a person frown and object “but psi is impossible.” The first worldview, straightforward materialism, was already incompatible with everything physicists have been doing for the last 100 years. And given the many unresolved problems remaining in modern physics, even the second worldview starts to seem pretty unlikely, even before we start discussing anything about parapsychology at all.

So there’s really no need to presume a conspiracy among parapsychologists, or act as though science is broken. Just shrug and say it’s a strange universe out there, or this is a simulation, or the supernatural is real, or I’m the only thing that exists, or the universe is all just math, or sheesh, who knows. You can still be atheist, or Christian, or gay, or straight, or woke as a biracial D.I.E. human resources manager, or whatever you want.

The only thing you can’t do is lord it over all the semiliterate horoscope lovers, Bible thumpers, and woo addicts by rubbing it in that every single thing they think is wrong and crazy and dumb. But why are you doing that anyway? Why were you ever doing that? Just argue with them on the basis of studies showing horoscopes are meaningless, numerous contradictions in the Bible, or the persistent failure of their crystals and herbs to help them get into that university they really love. You’re not going to make these people suddenly morph into a skeptic like you are; just let them be. They aren’t what’s interesting about the discussion.

What’s interesting is what the data seem very clearly to be saying about reality.

I know rationalists like to think of themselves as open-minded, flexible, scientific thinkers. But I experience the insistence that “parapsychology is not a science,” and “the psi effect is made up,” and “the control group is out of control” as dogmatic, unempirical, anti-science, and very much in the way. Such an attitude doesn’t even help to uncover a hypothetical conspiracy of fraud in the field; if we’re going to dispel the myth of an unreal psi effect, then establishing some evidence of fraud or misconduct would seem to be the way forward. If that’s too difficult, theoretically it should be possible for unbiased researchers to establish that the true Ganzfeld hit rate is truly 25%, exactly as would be predicted by chance, in the absence of a psi effect. The methods of parapsychology are open to anybody with the resources and dedication to try.

So whether the psi effect is real or fraudulent, empiricism continues to provide the way forward. There are fascinating doors to open, but we’ll never find out what lies on the other side by walling them off and saying “there’s nothing to see here.”

1. Rief, W., Nestoriuc, Y., Weiss, S., Welzel, E., Barsky, A. J., & Hofmann, S. G. (2009). Meta-analysis of the placebo response in antidepressant trials. Journal of affective disorders, 118(1-3), 1-8.

2. Ren, S., Ma, M., He, C., Deng, Y., Chen, X., Liu, Y., ... & He, L. (2022). Placebo effect in the treatment of non-small cell lung cancer: a meta-analysis. Journal of Bio-X Research, 5(03), 132-140.

3. Quattrone, Aldo; Barbagallo, Gaetano; Cerasa, Antonio; Stoessl, A. Jon (August 2018). "Neurobiology of placebo effect in Parkinson's disease: What we have learned and where we are going". Movement Disorders. 33 (8): 1213–1227.

4. Rief, W., Nestoriuc, Y., Weiss, S., Welzel, E., Barsky, A. J., & Hofmann, S. G. (2009). Meta-analysis of the placebo response in antidepressant trials. Journal of affective disorders, 118(1-3), 1-8.

5. Bérdi, M., Köteles, F., Szabó, A., & Bárdos, G. (2011). Placebo effects in sport and exercise: a meta-analysis. European Journal of Mental Health, 6(2), 196.