In Bem’s study, students were placed in front of a monitor whose screen was divided in half. An image would be shown on one half of the screen, and the subjects had to guess whether it would show up on the left or right half. According to Bem, the students correctly guessed where the image would be posted 53% of the time.
This of course, is well within chance. But Bem argues:
“Fifty-three percent would sound very small when chance would be fifty percent, but you look at political campaigns, if you win a campaign in a two-candidate race, such as Obama versus McCain, then fifty-three percent is sometimes be considered a landslide.” Alfonsi continues, “Landslide proof, he says, that ESP exists.”
I was stunned at the conflation that lies at the heart of this fallacy.
First of all, I am not familiar with any circumstance in which a 53% victory in a two-person election is thought of as a “landslide”, since a landslide refers to a victory with an overwhelming margin, at least according to the Random House Dictionary. A 53% victory would only represent an overwhelming margin if the second candidate received a percentage somewhere in the 30’s, or if the total did not add up to 100%.
But this is really beside the point. The real fallacy is not whether 53% is a landslide, but the fact that Dr. Bem is conflating what is needed for victory in an election with what is needed for results to be considered unaccountable by chance. Or perhaps he genuinely doesn’t know anything about chance percentages, thus illustrating once again that even trained doctors and scientists do not necessarily know how to determine whether an idea is scientific or pseudoscientific. To Dr. Bem, if I correctly predict 53 out of 100 coin flips, this constitutes evidence that I’m psychic.
I wondered if there would be any token skeptic to point out confusion between victory in a contenst and mathematical chance, and was a bit heartened when Alfonsi then mentioned that some of Bem’s colleagues disagree, pointing to Professor Robert Park, who was then shown saying:
“Nonsense radar is going off loudly, yes. Yeah, this is pure nonsense.”
That’s it? Nothing more articulate or detailed than that? No mention of how Bem doesn’t seem to know the difference between majority rule and the mathematics of chance? I've read Park's books, Voodoo Science and Superstition, and given how good he is at explaining pseudoscience for the layman, one wonders how elaborate Park’s answer was, and how much was left on the cutting room floor. You’d think those interviewing him would at least allow him to make one specific counterargument.
The segment then continued by mentioning how the U.S. intelligence community studied ESP, and that psychics helped them, as when they aided in the discovery a kidnapped U.S. general in the 1980s, at least according to a 1995 Nightline clip that they then show, in which some unnamed official attests to the accuracy of remote viewing. Sharyn Alfonsi concludes the segment by noting that critics point out that those psychics were used in about 500 cases, and were only right in about a dozen of them.
I suppose we could be grateful that that little bit of info was used to close out that segment. Still, there was nothing new about Bem’s ignorance, and I’m left wondering why ABC would bother with this prosaic entry into woo-woo reporting.
You can currently view the segment at: http://abcnews.go.com/watch/world-news-with-diane-sawyer/SH5585921/VD55105332/world-news-106-drastic-health-cuts-
The segment begins Act II of the episode, at the 14 minute mark.
Luigi Novi is an illustrator, writer and photographer from Union City, New Jersey whose work has been published or recognized by the Syfy Channel, Los Angeles Daily News, the Hudson Reporter, Clearview Cinemas, Toys R Us magazine, Image Comics, the New York Yankees, the Northern New Jersey Boy Scouts and across hundreds of articles on Wikipedia. On Randi.org, his text pieces can be seen in the May 2, 2003 and September 7, 2004 editions of Swift, and his caricatures of Allison DuBois and James Randi can be seen in several editions of Swift from 2005 - 2006.