Wait, ESP Is Real? 01.12.11  
Search only in the 5WJ archives

Average people can predict which side erotic images will appear on.

John Cloud

Statistically significant findings support the existence of precognition. And erotic pictures make a difference.

The idea that some people can see the future is one of those peculiar notions that is at once prehistoric and contemporary. You can find references to seers at least as far back as writing from ancient Greece, and you can also find numerous modern-day psychics on craigslist. So it's not surprising that academics would want to study precognition. But it is surprising that a major research psychologist has completed a study suggesting that extrasensory perception (ESP) is real.

The psychologist is Daryl Bem, an emeritus professor at Cornell whose work on sexuality and personality has won him awards and postings at Stanford and Harvard. So when a Psychology Today blogger wrote in October that Bem had completed nine thorough experiments on ESP, the psychology community started buzzing. When The New York Times picked up the story last week, it exploded.

The Times focused on the controversial decision by the Journal of Personality and Social Psychology to print Bem's study later this year. The journal is published by the American Psychological Association, and while it is not the most prestigious psychology publication, it is peer-reviewed and well-respected. The Times quoted skeptics calling Bem's ideas gimcrack science, and in the ensuing controversy, he was invited to appear on CNN and The Colbert Report.

But what, exactly, did he find? His experiments involved more than 1,000 subjects who took nine different tests, most of them simple computer trials that asked the participants to guess whether certain images would pop onto the screen. For instance, in the first experiment, 100 Cornell students looked at digital photos that mostly depicted non-sexual scenes but occasionally and randomly showed consensual adult sex acts. Across 100 sessions, Bem writes in his paper, participants correctly predicted when the erotic photos would show up 53.1% of the time, which crosses the bar for statistical significance (meaning there is at least a 95% probability that the findings could not be due to chance). By contrast, their hit rate on predicting non-sexual photos was about what you would expect: 50-50.

Similarly, in Bem's fifth experiment, 100 Cornell undergrads were exposed to two side-by-side photos and asked to pick which one they liked more. Then the computer randomly picked one of the two photos and showed it again for a few milliseconds. Before the photo was shown, the subjects were asked to predict which picture they thought the computer would choose. Once again, the students made the correct prediction 53.1% of the time, but only when the computer selected the picture they had already designated as their preferred one. If the computer chose the other one, the subjects' accuracy did not cross the statistically significant barrier.

In both this experiment and the one that used erotic photos, the findings suggest that precognition might have something to do with the salience of the image. We do better predicting a satisfying experience a sexy picture or a familiar and favored one than we do an unsatisfying or neutral one.

Still, like most serious research psychologists, Bem is careful about describing why his results turned out to support ESP. One significant finding in his data: subjects who score high on measures of stimulus-seeking in other words, people who like to take risks were far better at precognition than ordinary people. Stimulus seekers could predict the erotic computer images an astonishing 57.6% of the time. Statistically speaking, there's a 99% chance that this finding could not be due to chance.

Bem isn't sure why risk takers are so much better at ESP than others, and he is careful to note that he's not sure what kind of ESP, if any, is occurring. Maybe it is precognition: the participants are predicting the future. Maybe it is clairvoyance: the participants are somehow accessing information that already exists in the computer but hasn't yet been shown. Maybe it is psychokinesis: the participants are influencing the computer's random photo generator to show erotic images when they want to see one.

Or maybe it is all an accident. Maybe the computer programs he used were flawed (although he used several different gold-standard methods to generate randomness in the images). One flaw I identified in Bem's experiments is that, for the most part, the students were aware of the purpose of the experiments. They were explicitly told that the tests were designed to examine ESP. Generally, psychologists blind study participants to the reason for tests so the subjects won't be influenced by preconceived notions about them. I don't know how that foreknowledge could explain the results, but replication studies should blind participants.

Bem has invited me to Cornell next month, after exams are over and his lab re-opens, to take these tests myself. So look for a Lab Rat column in the coming weeks on whether I have ESP or whether the whole notion is more sci-fi than science.


5th WorldReturn to 5W