The Princeton Engineering Anomalies Research (PEAR)
The brainchild of Robert G. Jahn, who, in 1979, when he was Dean of the School of Engineering and Applied Science at Princeton University, claimed he wanted “to pursue rigorous scientific study of the interaction of human consciousness with sensitive physical devices, systems, and processes common to contemporary engineering practice.” In short, he wanted to be a parapsychologist and test psychokinesis (PK) and remote viewing (called remote perception by the PEAR folks). The PEAR folks are best known for their study of the mind influencing the behavior of machines, so this entry will focus on that work.
Scientists have been unable to find any clear and decisive evidence for psychokinesis. Those who claimed to move objects with only the power of their mind use tricks such as blowing on objects, moving them with thin threads, and using static charges to move objects. Some parapsychologists have not given up the chase, however. They began searching for micro-psychokinesis (MPK), minds affecting machines in ways that can't be detected except by statistics. Just as significant variance from chance in an ESP experiment is taken as evidence of ESP (the psi assumption), so a statistically significant deviation from chance in an MPK experiment is taken as evidence of MPK.
The PEAR lab shut down in February 2007 to a yawning scientific community.
In the 1960s, physicist and parapsychologist Helmut Schmidt started using random event generators to do MPK experiments. According to Dean Radin (1997), over the years Schmidt provided solid scientific support for the MPK hypothesis (or precognition, since there does not seem to be any way to tell the difference between MPK and precognition. Is the mind affecting the outcome of a random event generator or is anticipating what the outcome will be?)
In 1986, Jahn, Brenda Dunne, and Roger Nelson reported on millions of trials with 33 subjects over seven years trying to use their minds to override random number generators (RNG). Think of the RNG as producing zeros and ones. Over the long haul, the laws of probability predict that in a truly random sequence, there should be 50% of each produced. The subjects in the PEAR experiments tried to use their minds to produce more zeros (or ones, depending on the assignment). In short, the PEAR people did what many drivers do when they try to use their thoughts to make a red light turn green.
Radin thinks the PEAR group replicated Schmidt's work in 258 experimental studies and 127 control studies. C. E. M. Hansel examined the studies done after 1969 and before 1987 that attempted to replicate Schmidt’s work. He notes: “The main fact that emerges from this data is that 71 experiments gave a result supporting Schmidt’s findings and 261 experiments failed to do so” (Hansel 1989: 185). That is the beauty of meta-analysis: you can transform a failure rate of nearly 4 to 1 into a grand success.
In 1987, Dean Radin and Nelson did a meta-analysis of all RNG experiments done between 1959 and 1987 and found that they produced odds against chance beyond a trillion to one (Radin 1997: 140). This sounds impressive, but as Radin says “in terms of a 50% hit rate, the overall experimental effect, calculated per study, was about 51 percent, where 50 percent would be expected by chance” [emphasis added] (141). A couple of sentences later, Radin gives a more precise rendering of "about 51 percent" by noting that the overall effect was "just under 51 percent." Similar results were found with experiments where people tried to use their minds to affect the outcome of rolls of the dice, according to Radin. And, when Nelson did his own analysis of all the PEAR data (1,262 experiments involving 108 people), he found similar results to the earlier RNG studies but "with odds against chance of four thousand to one" (Radin 1997: 143). Nelson also claimed that there were no "star" performers.
However, according to Ray Hyman, “the percentage of hits in the intended direction was only 50.02%" in the PEAR studies (Hyman 1989: 152). And one ‘operator’ (the term used to describe the subjects in these studies) was responsible for 23% of the total data base. Her hit rate was 50.05%. Take out this operator and the hit rate becomes 50.01%. According to John McCrone, "Operator 10," believed to be a PEAR staff member, "has been involved in 15% of the 14 million trials, yet contributed to a full half of the total excess hits" (McCrone 1994). According to Dean Radin, the criticism that there "was any one person responsible for the overall results of the experiment...was tested and found to be groundless" (Radin 1997: 221). His source for this claim is a 1991 article by Jahn et al. in the Journal of Scientific Exploration, "Count population profiles in engineering anomalies experiments" (5:205-32). However, Jahn gives the data for his experiments in Margins of Reality: The Role of Consciousness in the Physical World (Harcourt Brace, 1988, p. 352-353). McCrone has done the calculations and found that 'If [operator 10's] figures are taken out of the data pool, scoring in the "low intention" condition falls to chance while "high intention" scoring drops close to the .05 boundary considered weakly significant in scientific results."
According to McCrone, the "size of the effect is about .1 percent, meaning that for every thousand electronic tosses, the random event generator is producing about one more head or tail than it should by chance alone" (McCrone 1994). Jahn says that the measured effect of MPK was "not large enough that you're going to notice it over a brief experiment, but over very long periods of study, we see a systematic departure of the behavior of the machine in correlation with what the operator wants it to do" (Park 2000: 198). Most experiments in medicine or psychology use fewer than 100 trials, or perhaps a few hundred at most. Big trials will have 25,000 or more subjects. Massive prospective studies might survey 250,000 people. The most commonly used P-value in the social sciences and medical studies is P<0.05, where there is a one in twenty chance that the result is a statistical fluke. The 95% confidence interval, used as a standard in most of these studies, is arbitrary, however. It can be traced back to the 1930s and R. A. Fisher. There is nothing sacred about the standard, but it was not introduced to be used with studies having millions of data points. The RNG studies go into the millions of trials, allowing a very small effect to generate a very large statistic. When we’re dealing with small effects and millions of trials "even the slightest departure from the assumptions might suffice to produce artificially significant outcomes" (Hyman 1989: 151). The main assumption that Jahn and his colleagues made may not be warranted. "It is not clear that any of these machines is truly random. Indeed, it is generally believed that there are no truly random machines. It may be that lack of randomness only begins to show up after many trials" (Park 2000: 199).
These data should remind us that statistical significance does not imply importance. Science that claims to have identified barely detectable causal agents observed near the threshold of sensation, which are nevertheless asserted to have been detected with great accuracy and be of great significance, is one of the signs of what Irving Langmuir called pathological science and Bob Park calls voodoo science.
Furthermore, Stanley Jeffers, a physicist at York University, Ontario, has repeated the Jahn experiments but with chance results (Alcock 2003: 135-152). (See "Physics and Claims for Anomalous Effects Related to Consciousness" in Alcock et al. 2003. Abstract.) And Jahn et al. failed to replicate the PEAR results in experiments done in Germany (See "Mind/Machine Interaction Consortium: PortREG Replication Experiments," Journal of Scientific Exploration, Vol. 14, No. 4, pp. 499–555, 2000).
Based on the results of these experiments, Radin claims that “researchers have produced persuasive, consistent, replicated evidence that mental intention is associated with the behavior of …physical systems” (Radin 1997: 144). That sounds like a hasty conclusion to me. He also claims that “the experimental results are not likely due to chance, selective reporting, poor experimental design, only a few individuals, or only a few experimenters” (Radin 1997: 144). He's probably right except for the bit about it being unlikely that the experimental results are due to chance or to only a few individuals.
Jahn, six of his associates, and PEAR even have a patent (US5830064) on an “Apparatus and method for distinguishing events which collectively exceed chance expectations and thereby controlling an output.” The PEAR people are so convinced of the breakthrough nature of their work that they have incorporated as Mindsong Inc. They claim their corporation "is developing a range of breakthrough products and research tools based on a provocative new technology—proprietary microelectronics which are responsive to the inner states of living systems." One of their breakthrough products is some software "that allows you to influence, with your mind, which of two images will be displayed on your computer screen." They also sell a device for several hundred dollars that lets you do your own testing of mental influence on randomized outputs.
On their website, PEAR states that after more than twenty-five years they are shutting down and moving on. "Over the next few years, PEAR will be concluding its experimental operations at Princeton University," says the notice on their "Future" tab.* It seems that Bob Jahn and Brenda Dunne are not quitting, however, for they are looking for a number of like-minded folks who want to spend their time or money on the study of minds interacting with machines. They've set up a new outfit called International Consciousness Research Laboratories. As their first fundraiser, they are selling "a multi-DVD/CD set entitled The PEAR Proposition" for a mere $62, including psychic shipping and handling. The replication studies are available on a blank DVD disc, available at no cost to true believers.
Perhaps the most disconcerting thing about PEAR is the fact that suggestions by critics that should have been considered were routinely ignored. Physicist Bob Park reports, for example, that he suggested to Jahn two types of experiments that would have bypassed the main criticisms aimed at PEAR. Why not do a double-blind experiment? asked Park. Have a second RNG determine the task of the operator and do not let this determination be known to the one recording the results. This could have eliminated the charge of experimenter bias. Another experiment, however, could have eliminated most criticism. Park suggested that PEAR have operators try to use their minds to move a "state-of-the-art microbalance" (Park 2008, 138-139). A microbalance can make precise measurements on the order of a millionth of a gram. One doesn't need to be clairvoyant to figure out why this suggestion was never heededhttp://www.skepdic.com/pear.html