What’s the story on "ganzfeld" experiments?

SDStaff DavidB replies:

The website for the Committee for Scientific Investigation of Claims of the Paranormal (CSICOP) has several articles on the topic, so I’m not sure which one you’re talking about (though your description makes me think it’s the one written by Matt Nisbet — see below). But let’s take a look at the ganzfeld phenomenon from the top.

Ganzfeld experiments (ganzfeld is German for “whole field”) have been around in some form since the 1930s. The studies that drew the most recent attention were done by Daryl Bem, a Cornell University psychologist, and Charles Honorton, a University of Edinburgh parapsychologist. Ganzfeld experiments involve covering a subject’s eyes (generally with halves of ping-pong balls), bathing them in a red floodlight, and feeding white noise through headphones into their ears. This supposedly makes the subjects unable to receive sensory information. Then another person in a room shielded from the first subject tries to mentally send a specific picture. The first subject reports whatever imagery comes to mind and then tries to identify which of a group of pictures the second person was trying to send via ESP.

The idea here is that ESP is such a weak force, it is normally drowned out by our other senses. So with all other sensory input suppressed, the subject should be able to better use ESP.

Bem and Honorton used meta-analysis to combine the results of several studies, and found a hit rate of about 35%–enough above the expected chance rate of 25% to be significant, if true.

“If true” is, of course, the key. Psychologist and skeptic Ray Hyman found statistical anomalies in the original ganzfeld experiments. His analysis of various experiments “showed that this database did not justify concluding that ESP was demonstrated” (Skeptical Inquirer, March 1996). He went on to note that “both Honorton and I agreed that there were sufficient problems with this original database that nothing could be concluded until further replications, conducted according to specified criteria, appeared.” In later experiments, he found that the experimenter interacted with the subject during the process of judging whether or not a hit was made. He notes: “This means the judgments from trial to trial were not strictly independent.” He discovered other flaws in the experimental procedure as well.

More recently, Julie Milton (University of Edinburgh) and Richard Wiseman (University of Hertfordshire) did a follow-up meta-analysis on 30 new studies not included in the Bem and Honorton analysis. Scott Lilienfeld reports in one of those CSICOP articles (Skeptical Inquirer, November 1999l that those findings “stand in stark contrast to those of Bem and Honorton and raise serious questions concerning the replicability of the ganzfeld findings.” In science, replicability is essential. If other scientists can’t reproduce your results, chances are they may have been due to some flaw in the experiment rather than a real effect. Milton and Wiseman found a result “which corresponds to essentially chance performance and can most charitably be described as negligible.”

The Skeptic’s Dictionary entry on ganzfeld experiments discusses claims made by one of Honorton’s co-authors, Rick E. Berger, Ph.D. Berger said the odds were “a million billion to one” that the hit rates they got weren’t due to chance. However, since the experiment involved an interpretation of the subject’s verbal description, it’s possible to get a hit without really picking an image specifically, just because it happens to be the one the experimenter thinks is closest. (This relates to Hyman’s comments about the experimenter interacting with the subjects.)

One example the Dictionary gives is a verbal description from Berger’s website:

The Dictionary notes: “There are quite a few images that would ‘match’ this description, since the description itself contains at least eight distinct images (the Lincoln memorial, Lincoln, 4th of July, fireworks, Valley Forge, bombs, Francis Scott Key, Charleston) to which one could easily add a couple more, such as the American flag, the star spangled banner, and, oh yes, George Washington, which was the image selected as most closely resembling the verbal description. We’re not told what the other three choices were.”

The Dictionary continues: “One wonders why, if this 8.2%, million billion to one difference is evidence of telepathy, the verbal descriptions are not more precise. For example, why didn’t the psychic ‘see’ George Washington, since that was what the image was? Why did he see the Lincoln memorial and a bunch of other things? How can they be sure of what they are measuring?”

Finally, the Dictionary brings up another important point — the experiments weren’t done with a “none of the above” option, nor were there control experiments in which the “sender” sent nothing. The article notes: “If Berger and Honorton would do a ganzfeld where the sender sends no messages at all throughout the entire experiment, my guess is that the receiver would still ‘receive’ and give a verbal description of his vision.” Indeed, it would be interesting to see the results of such an experiment.

With regard to your question about the paper submitted with a prominent author’s name on it, I’m guessing you’re referring to Matt Nisbet’s Generation sXeptic article, where Nisbet discusses his interviews with Bem and Hyman. He noted that Bem considered his own reputation in psychology as having contributed to a journal’s acceptance of the article. Nisbet’s point seems to be that in other scientific arenas, studies are reviewed blindly; but in psychology they’re not, which can lead to papers by well-known people being accepted when perhaps they shouldn’t have been. I don’t see that Nisbet is biased other than being biased towards good science — and I think he did a good job of letting each man speak for himself.

The point is, whatever you may have seen on TV, there is no “amazing” ESP proof. At best there are some statistical suggestions that something may be afoot. Right now, however, the evidence points more at experimental error than at ESP.

This surely won’t be the final test of ganzfeld — parapsychologists undoubtedly will continue to use it in hopes of proving the existence of ESP. It’s good that they’re using scientific methodology, but they must be continuously aware of their own biases and make sure those don’t enter into the analysis. Unfortunately, history tells us that in the field of parapsychology that happens all too often.

SDStaff DavidB, Straight Dope Science Advisory Board

Send questions to Cecil via cecil@straightdope.com.

STAFF REPORTS ARE WRITTEN BY THE STRAIGHT DOPE SCIENCE ADVISORY BOARD, CECIL’S ONLINE AUXILIARY. THOUGH THE SDSAB DOES ITS BEST, THESE COLUMNS ARE EDITED BY ED ZOTTI, NOT CECIL, SO ACCURACYWISE YOU’D BETTER KEEP YOUR FINGERS CROSSED.