In ganzfeld studies, a receiver who's unaware of participating in a psi experiment describes his or her thoughts for 30 minutes. In a separate room, a sender who can hear the receiver views a video or still photo. The sender tries to telepathically coach the receiver into imagining the same scene. Afterwards, the receiver chooses from a set of four images the one that most closely resembles his or her earlier thoughts.

In 1994, psychologist Daryl J. Bem of Cornell University and his colleague, the late Charles Honorton of the University of Edinburgh in Scotland, described surprising results from Honorton's series of 11 ganzfeld experiments. They found that more often than could be explained by chance, receivers chose the image that matched the one seen by senders (SN: 1/29/94, p. 68).

Now, psychologists Julie Milton of the University of Edinburgh and Richard Wiseman of the University of Hertfordshire in Hatfield, England, have pooled the results of 30 ganzfeld experiments using a statistical method called metanalysis. In the July Psychological Bulletin, the same journal in which Bem and Honorton presented their results, the researchers report that their analysis shows no consistent evidence for psi.

"Although the new studies failed to replicate the effects of earlier studies, it is not clear why they did so," Milton says.

"This has been the history of parapsychology for about 150 years," remarks psychologist Ray Hyman of the University of Oregon in Eugene. "Procedures look good at the beginning, and then they fizzle out. Whether or not this is different is hard to say, but this metanalysis suggests that on the average it doesn't look like there's much going on."

The metanalysis has generated heated discussion among psychologists. Some argue that Milton and Wiseman were unjustified in lumping all 30 studies together because their results were so disparate. Milton contends that a standard statistical test of variation among the results showed that they could treat the studies as a uniform set.

Bem says, however, "The reason the effect isn't significant is that there are three studies that are pulling down the average, and those studies are very nonstandard." Further, 6 of the 30 studies showed significant psi effects—more than would be expected by chance, he adds.

Since the metanalysis was completed, nine more ganzfeld studies have been published. Milton acknowledges that the psi effect would be statistically significant if the analysis were updated to include these studies. However, she observes, a single study had an especially strong result, but no clear-cut effect spans the broad range of investigations. "That will be a crucial thing to demonstrate in order to make a strong claim that the studies show a genuine anomaly," she says.

"My hope is that some of the proponents of the ganzfeld procedure will try to use [the new report] as a springboard to develop something replicable," says psychologist Scott O. Lilienfeld of Emory University in Atlanta. "The ball is now in their court."

References & sources for this article