Now, we don’t need anything close to 100% accuracy to have had a successful experiment. Generally, anything above 30% is very significant statistically (given sufficient trials,) with 25% the baseline for chance. If we do find statistically significant results, are we ready to publish a paper? (The parapsychologists in the room are not allowed to answer.) The answer is no, because we haven’t finished eliminating the other possibilities.

For example, is our randomizer really random? In one experiment of this kind the random number generator automatically did not choose the same number twice. This increases the baseline for chance from 25% to 33% and suddenly 30% is just about right.

Is there any way for subject A to signal subject B other than by psychic means? If they’re in the same building, subject A could stomp on the floor and theoretically give the number to subject B rather easily.

Did we do enough trials? Thirty percent is not at all significant if we only did 100 trials. And the reverse of this is, how many times did we do this experiment? If we set the bar at 2,000 tries, and then did it fifty times on fifty different occasions, we may very well end up with a single test that shows statistically significant results. But the forty-nine times we ended up with chance results may have had something to do with it; eventually we were going to get lucky. We might even get something statistically BELOW chance, which we could report as “psi missing.” (I’m not making this up.) We could publish a paper announcing on two trials we had statistically significant results above chance on one trial, and statistically significant misses below chance on another. Parapsychologists will toast us. We just have to bury the data on the other 48 trials.

Finally, if we pass all the other questions, we still have to answer what might seem obvious: what did we actually test for? Was it ESP, with subject B reading the mind of subject A? Or did subject A send the information to subject B? It could have been precognition; subject B could have been looking into the future at a printout of the results from subject A’s computer. It could even have been PK; subject B might have influenced the randomizer in the computer to get it to come up with specific numbers. This sort of confusion is what makes Pratt’s defense of Soal perfectly reasonable to other parapsychologists.

There IS a much simpler test of psi, or more specifically, PK. Take a sewing needle and put it between two magnets in such a way that is rests upright on its tip. Put the entire thing in an enclosed glass container. Sit the subject at the other end of the room and tell him to try and get the needle to spin.

It’s simple, it’s direct, and it would be very difficult to do without psychic powers, which is probably why no parapsychologist has tried it, even though it was suggested years ago.

But none of this actually gets to the root of what’s wrong with parapsychology and psi because, experimental problems notwithstanding, there is a huge gap where there should be theoretical psi research.

Before psi is going to be accepted by the scientific community at large (and by skeptics like me) someone is going to have to come up with a positive theory. By this I mean what’s missing is a theoretical explanation for how psi COULD exist. This is psi’s achilles’ heel, and it’s ultimately why I don’t believe in it.

The problem is, psi experiments are all essentially designed to prove psi by process of elimination. This is, at its core, an argument from ignorance. Take the needle experiment again. If someone sits at the other end of that room and manages to make the needle spin, well, we’ve eliminated every physical possibility from our list of explanations, so would we be right to conclude that it MUST be psychic power at work? Not necessarily, no.
Just because we can’t think of another explanation doesn’t mean there isn’t one. It may very well be that our subject psychically willed the needle to spin, but it is no less likely that a mouse hiding under his chair psychically willed it to spin, nor is it any less likely that some other unknown phenomenon is at work causing the needle to spin. There is no logical distinction between one unknown phenomenon and another, and because we don’t have a positive theory that makes definite, testable predictions, we cannot conclude with any degree of certainty whatsoever that psi was involved.

There are many reasons to doubt that any practical theory of psi will be forthcoming, because just about every attempt to do so will violate at least one well-established law or prove to be a biological impossibility. Take the issues of instantaneous communication and remote viewing. If a psychic can instantly communicate with someone else, or “see” something that is a great distance away, by exactly what mechanism is this instant information being received? The psychic can’t just “know” it without it being communicated to him and so, if real communication is taking place, there HAS to be a delay between the sending and the receiving. But despite the fact that nothing can travel faster than light, remote viewers claim they can go to other planets immediately, and even to other galaxies that are many light years away. The lack of a time gap creates a huge credibility gap.

You might say-- as many proponents of psi do-- that an undiscovered aspect of physics is the answer to this question. Consider what you’re doing when you argue this point. You’re invoking an undiscovered phenomenon to explain another undiscovered phenomenon. You’ve now gone from science to fantasy, perhaps without even realizing it.

One field of legitimate science that has recently been co-opted to serve psi proponents is quantum mechanics. In quantum theory, there is something equivalent to instantaneous information transfer, best illustrated by the EPR paradox.  The problem is, applying anything in quantum mechanics to the macroscopic universe is a mistake. This is like saying that because an ant can survive a fall from a tabletop, you can live through a fall off of the Empire State building. We do not use quantum mechanics-- which is largely a non-observational series of mathematical constructs already-- to make predictions about the world on our scale. It’s not designed to do so.

So why don’t I believe in psi? There’s no working model to explain how it CAN work. There’s no experimental proof that it works, or even exists, and every experiment conducted on it to this point has been proven to be either severely flawed or possessed of results that have much more prosaic explanations.

Most people I know who profess belief in psi utilize a much simpler argument. “There MUST be something there” they argue. Occasionally anecdotal evidence is provided, evidence that in my eyes is either non-spectacular or simply a matter of coincidence. Or luck, even. Cases of “I was just thinking of so-and-so and then he called” which are much more readily ascribed to remembering hits and forgetting misses abound. I don’t begrudge anyone’s belief, even those of the parapsychologists. (I do, however, wish more of them were as honest as John Taylor, who eventually disavowed his own research when he realized he was fooling himself.) What I find surprising is that in the first part of the twenty-first century, my opinion that psi does not exist puts me in a significant minority.

Which is why, when the question “why don’t you believe in psi” is posed, I usually answer with a question of my own: “why do you?”

Why does anybody?