A Physicist Investigates

This is the second part on my series on the parapsychological double-slit experiment.

The original double-slit experiment was performed in 1803 by Thomas Young.
The parapsychological version was thought up by one of his intellectual heirs, almost 200 years later.

In the Beginning

Stanley Jeffers, a professor of physics at York University in Toronto, Canada, came across the results from PEAR lab during a sabbatical in 1992. The PEAR project claimed to have shown that people could influence Random Number Generators (RNGs) by their intention alone. There were also earlier experiments which claimed that people could influence radioactive decay.
Jeffers became interested enough to follow up on that. Jeffers came up with a more direct way to test this idea by having people try to influence the diffraction of photons (unsuccessfully). (Jeffers and Sloan 1992)

Afterwards, he refined that experiment by using a double-slit set-up. This makes for an exquisite test of whether humans have some unknown means of affecting the world.
If there is some way of making information available, about through which slit the particles went, then this should affect the interference pattern. Moreover the change in the pattern would allow one to estimate just how much information was made available.
It’s not necessary for anyone to become consciously aware of this information, it’s enough that it exists. For example, if people can somehow influence what happens at one slit but not the other, then that could “tag” the particle in such a way that it becomes possible to distinguish through which slit it went. Such a thing would be sufficient to affect the interference pattern.
More directly, if people can clairvoyantly observe one slit (aka remote view it), then this too should make which slit information available.
There’s a catch, though. That’s only true as far as we understand the laws of nature. However, the paranormal, by definition, is not bound by that.

There’s another thing that could lead to a positive result. The experimental apparatus needs to be carefully calibrated. If the right parts are bent out of shape by just a few micrometers or if the laser is somehow affected then this could also lead to a change in the pattern. It would be quite tricky to mimic the change expected from available information but it could still lead to a positive result under the right (or wrong) circumstances. Either if the subconscious psychokinetic ability is also ingenious or if the pattern is not scrutinized properly.
Jeffers seems not to have considered this possibility but, of course, outside of parapsychology such ideas, IE psychokinesis, tend to be regarded as rather philosophically than actually possible. Nevertheless, parapsychologists have little hesitation to jump to any conclusion and it would be quite in line with what is often claimed.

All in all, Jeffers’ double-slit set-up promises to be a very sensitive means of detecting even the smallest influence, whatever it’s nature.

A Double-Slit Diffraction Experiment to Investigate Claims of Consciousness-Related Anomalies

Stanley Jeffers ran 74 sessions, each consisting of many attempts to influence the pattern. Then he gave up and presented his findings at a conference. After that, the people of the PEAR lab borrowed the device and used it to run 20 sessions.
I do not have the original conference report by Jeffers available to me but was able to find first hand information by Jeffers in the book Psi Wars: Getting to Grips with the Paranormal.
The two experiments were reported together in The Journal of Scientific Exploration, which is dedicated to publishing stuff that is, put bluntly, too stupid or crazy for normal peer-reviewed journals. (direct link to paper)
The article is authored by Ibison and Jeffers but appears written by Ibison.

While Jeffers completely failed to find anything in his 74 sessions, Ibison of the PEAR lab reported a “significant” result with only 20. Some will say that the latter experiment had a positive result.

I fear, I have to go into a little detail on that. This is an issue that will be relevant for all the following experiments. The interesting thing here is the contrast of the interference pattern. But you can’t measure it with arbitrary precision.
The pattern itself will be influenced by tiny disturbances of the apparatus caused by changes in ambient heat or vibrations. The sensor which picks up the brightness, basically a predecessor to the CCD chip in current digital camera, is not perfect either. And then there’s more esoteric things, too.
The upshot is that every measurement will have a slightly different result.
According to Jeffers in Psi Wars, repeated calibrations yield typical values for the contrast of 0.991 with a standard deviation of 0.001, while Ibison talks of “around 5% of the peak value”. I don’t quite know what Ibison means by that since it’s not how measurement uncertainty is usually reported. However, it does seem to imply more uncertainty than Jeffers’ figure. Maybe there’s a typo in one of the sources or maybe Ibison at the PEAR lab was not able to calibrate the device as well as Jeffers who was, after all, a pro at this.

What is clear is that any influence by the subjects must have been so small that it never rose above the noise, that is, the fluctuations due to imperfect measurement. Unfortunately, none of the sources I have at hand provides information on what kind of an upper bound this puts on the subjects’ maximal influence.

It is possible to measure something even if it hides in the noise. To do so, one must simply repeat the measurement over and over again and form an average. In the long run the errors in either direction will balance out. The more often one repeats the measurement, the more reliable the average. Using statistics, one can compute how reliable a result is, which is then expressed with error bars.
Basically, Ibison found that the contrast was slightly altered and computed that there was only a 1 in 20 chance for that to happen, merely from noise. That means, if you repeated the experiment many times, you’d get such a result or a better one 1 in 20 times, if you’re only gazing at noise. That’s, of course, not very impressive, even not taking into account that such figures tend to be exaggerated. There’s a couple of things that can make a result appear more unlikely than it really is. Physicists tend to ask for much more clear-cut results.

Conclusion

Fundamentally, this should be a discouraging result for parapsychology. Even though a device was constructed able to detect influences far smaller than that what most parapsychological experiments require, the supposed effect was still hidden in the noise.

Someone who believes that there really was an effect there has two basic options. The preferred one should be to build a better apparatus, that has less noise and so will let the true effect stand out.
The other one is, as mentioned, to run this experiment a lot of times but that will not be nearly as useful, nor as convincing.

Further Thoughts

I mentioned that one way of achieving a positive result in this test is by exerting some physical force on the apparatus. A physicist or engineer would test this directly by having people try to exert that force on a force sensor of some kind.
Obviously, parapsychologists don’t do it that way. It fails to show what they know to be true.
One answer to that is abandoning careful experiments and simply argue that some feat can’t possibly have been a trick because they surely would have noticed. Parapsychologists apparently do not suffer from the same limitations to perceptions as the rest of us. Or perhaps of a few more inabilities to see something.
The other answer is to conduct experiments involving randomness, like the RNGs of PEAR or radioactive decay or something as simple as die. These experiments seem much less silly than any argument for the genuineness of something that looks like a magic trick.
But are they really? Why not measure the effect directly?
Why is there always randomness involved?
An obvious answer is that the “effect” is simply the result of data mining. My take on that here.

But, since I aim to deliver the whole truth, I must also tell you that parapsychologists also have ideas on the issue. One idea that Ibison mentions in the mentioned article is that maybe experimenters can see into the future and somehow know, subconsciously, when to start an experimental run so that the purely random processes will deliver a favorable result. Ibison is quite open with the fact that the apparatus was basically just generating random numbers and not necessarily measuring anything.

Footnote

Stanley Jeffers has published his conclusion on the results on the PEAR lab results Skeptical Inquirer in 2006 and 2007, and also written an essay which was published in the book Psi Wars: Getting to Grips with the Paranormal. (Review by Caroline Watts)

Advertisements