P.E.A.R. R.I.P.
Unceremoniously tucked into the AP Science News wire last week was a short, photoless item that served as an obituary for a most unusual scientific endeavor: the infamous Princeton Engineering Anomalies Research (PEAR) laboratory at Princeton University. With its sterile tech-speak title and stated focus upon the vague phenomenon of “human/machine interactions,” PEAR may not seem particularly controversial on the surface, yet in its 28 year existence it was treated as a joke by hysterical media reports and heaped with the scorn of the the scientific community. Confused? What if I told you those human-computer interactions were supposed to involve ESP, telekinesis, and “quantum consciousness?”
Yes, PEAR was actually an Ivy League lab focused on the study of ESP, or extra sensory perception, the stuff of late-night Sci-Fi channel documentaries and the occult section of the used bookstore. Founded in 1979 by Roger Jahn, an expert on jet propulsion and former dean of Princeton’s engineering school, the group spent nearly three decades in a basement laboratory space attempting to conclusively prove that humans can control the function of machines with their mind. While this enterprise may summon up images of Bill Murray giving electric shocks to undergrads in Ghostbusters, PEAR set out to study this phenomenon using the strictest scientific standards, generating a pile of publications and putting out a book in 1988.
And the response of the scientific community to this body of work? Let’s just say it was…underwhelmed. Scientists at Princeton bemoaned the loss of credibility that PEAR’s existence brought to the university, peer-reviewed journals scoffed at the lab’s submissions (one editor quipped “when you are able to transmit this text to us telepathically, we shall consider it seriously.”), and traditional sources of funding zipped their wallets up tight. Fortunately, the media was there to treat PEAR’s investigations with the proper decorum and dignity…or not. Most outlets lumped the laboratory researchers in with Sasquatch-hunters and Loch Ness Monster-freaks, and even the AP obit couldn’t resist a snide lead-in, commenting that PEAR “will be shuttered at the end of the month. Maybe you already knew that.” LOLocaust!
With the exception of a few more even-handed reports, PEAR has been consistently dismissed out of hand (by media and scientists alike) for their fantastical premise, and very few critics seem to have actually deigned to investigate their data and specific claims. It’s not that the research was boxed away in the dusty eves of some gothic mansion. In fact, the PEAR website contains direct links to many of their publications, most printed in a journal the lab directors themselves helped start in response to stonewalling by the establishment. The most fascinating of these articles, a 50-page manifesto entitled The PEAR Proposition (beware: large PDF file), is a sometimes torturous, sometimes shockingly confessional, tour de force of scientific acrimony; an account of the lab’s accomplishments and battles that is all the more teeth-grindingly bitter due to the solidly intellectual style it adheres to.
The article is partly a review of PEAR’s conclusions and models, and partly a history of the lab’s existence, trials, and accomplishments. If you can get past the occasional puddle of linguistic quicksand like “one possible format for visualization of the subliminal seed space that undergirds the tangible reality and conscious experience regimes utilizes an array of complex vectors whose components embody the pre-objective and pre-subjective aspects of their interactions” (whew), you get flashes of genuine mad-scientist charisma: infectious enthusiasm and belief in the work, gracious indebtedness to volunteers and patrons (including a McDonnell and a Rockefeller), wry counterattacks to career enemies. While the lab’s actual results are not entirely convincing, shrouded as they are in convolutions of statistics and impenetrable terminology, it’s easy to come away with the impression that the researchers have been treated poorly by stubborn, arrogant colleagues and dismissive reporters.
PEAR’s actual findings don’t suggest that there are Carries and Gellers walking among us. Their research focused instead on the much more banal task of observing subtle manipulations in the function of machines called Random Event Generators (REGs). At their most basic, a REG is just a random number generator, producing random series of binary ones and zeroes or choosing random numbers between 1 and 100. “Operators” would sit in PEAR’s laboratory and try to influence the machine’s random function; for example, they would concentrate on willing the binary number generator to produce more 1s than 0s or vice versa. After sometimes running literally millions of trials in this fashion, PEAR researchers analyzed the REG’s data stream and found significant deviations from a truly random series of numbers that corresponded with the direction the operator was trying to induce. In other words, they concluded, people had the power to alter a machine’s functional output merely by thinking about it.
Experiments of this sort were repeated again and again over PEAR’s nearly three decades of existence, with scores of different operators, different REG machines (including robots and a device that looks to me like a gigantic Plinko game), and occasionally involved more esoteric protocols, like placing the operator thousands of miles away from the REG, or having them try to influence the machine several hours before it was actually turned on. According to The PEAR Proposition, most (but not all) of these experiments confirmed the lab’s initial observation of “anomalous effects.” Sounds great, right? Any day now we should be driving cars with our mind and starting the coffee machine from our shower, not to mention playing a whole multitude of clever practical jokes on co-workers and friends.
Ah, but as always, science is a cruel tease. The effects that PEAR researchers reportedly observed were extremely subtle – roughly one deviation from absolute random for every 1000 data points. To put that number in perspective, let’s imagine the binary number generator spits out a 0 or 1 every second, and you sit there and concentrate on generating more zeroes. For every 16 minutes’ worth of numbers, you will have generated one more zero than the machine would have otherwise spit out. And that’s on average; in order to come up with a rate of deviation that’s more than mere chance, the PEAR researchers had to conduct millions upon millions of trials.
There’s a basic statistical quandary here, one that PEAR graciously owns up to in their Proposition. Every scientist hopes their experiments will achieve statistical significance, that magical state of being 95% certain that your results are not due to the effects of chance (less than 5% probability of chance effects being the industry standard in most fields for real vs. imaginary data). Thanks to the advent of computers and statistical programs, there are thousands of ways to analyze one’s raw data and hope for the best, and the slightly sleazy practice of mixing and matching these analyses is an essential skill for any successful investigator. There is one basic law common to all this statistical finagling: significance can be reached by either increasing the magnitude of your effect (duh), or by drastically increasing the number of trials, samples, or subjects.
The latter approach was favored by the PEAR group due to the extremely subtle nature of the observed effect; hence, the millions of trials. The PEAR studies also do some borderline fudging by combining datasets from different operators and different experiments, because, as they put it, “only over very large individual or collective datasets can convincing anomalous trends emerge.” Sure enough, the data does reach significance after all this massaging, but the actual difference between the operator-“influenced” results and random fluctuations is a mere .01% - an almost infinitesimal shift. As Robert Todd Carroll at The Skeptic’s Dictionary somewhat condescendingly notes about PEAR’s results, “these data should remind us that statistical significance does not imply importance.”
An even bigger problem with PEAR’s data is that the sheer volume of work required to achieve significance makes these experiments incredibly difficult to replicate, replication being another integral brick in the road to widespread acceptance. Again, the authors face up to this objection; there’s even a section in The PEAR Proposition entitled “Replicability Issues.” It turns out that attempts to reproduce the anomalous effects outside of the PEAR lab failed, with results falling well short of significance. Jahn and Dunne attribute this discrepancy to “hide-and-seek characteristics” and a “decline effect,” and proceed to throw more stats at the data until it fits a model, but the damage has already been done, resulting in what the authors themselves describe as “consequent guffawing by the skeptics.”
Personally, despite hearing about these flaws and shortcomings, I’m not so quick to throw PEAR’s entire body of work in the trash. The old saying: “There are three kinds of falsehoods: lies, damn lies, and statistics,” applies as much to the scientific mainstream as it does to labs like PEAR on its fringes. From the most technical physics, to the massive clinical studies that accompany a new drug’s release, science is riddled with small or abstract effects that must be tested in enormous numbers to be proven genuine.
Yet it’s not methodology that’s the most problematic aspect of their work What gave the PEAR group the most difficulty was their apparent inability to propose a clear mechanism or model for the effects they observed. A clear and testable mechanism for the observed anomalous effects would do wonders for PEAR’s acceptance within and without the scientific community. But instead, each new complication - the above-described “irregular replicability,” differences in the magnitude of effect based on sex differences, effects of “empathy,” effects of group size, etc. - required them to add an additional layer to an already baroque theoretical model. Three interconnected mechanisms are proposed in The PEAR Proposition, and I can only begin to understand one of them: an extension of quantum mechanics terminology to describe the function of human consciousness. Schrödinger and Heisenberg make guest appearances, the physics community’s uproarious response is described (they call Jahn and Dunne’s use of quantum mechanics “prostitution”), and the terminology gets even denser and disturbingly New-Agey.
Granted, science does not always demand an immediate explanation for a novel, observed effect, but it certainly frowns on the introduction of new and unexplained forces – even Newton had difficulty pleading his case for gravity because he refused to frame a hypothesis about how it worked. Almost any sort of plausible theory explaining PEAR’s effects would at least open new avenues by which the phenomenon could be tested, allowing the experiments to progress beyond merely repeating observations with various types of REG machines. By going the abstract route of borrowing quantum physics terminology (i.e. wave functions, eigenvalues, and uncertainty), the researchers in a sense throw up a brick wall to such progress, putting forth a model that’s more philosophical than scientific, and incredibly difficult to test in any direct fashion.
In addition to confronting their critics about these fundamental methodological and theoretical issues, PEAR was forced to wage an all-out battle on another front: making the case for studying this phenomenon at all. The PEAR Proposition contains several amusingly proper descriptions of these aggravations (seriously, read pages 203-208, “Ancillary Strands,” if you read nothing else), such as the “recalcitrant university administration,” “considerable bleating from the canonical physical science community,” “despicable” and “sanctimonious attempts by self-styled critics” to discredit the work, and media-derived “transitory embarrassments.” Feelings of persecution run throughout the paper – the authors compare their struggles to the heroes of The Barber of Seville and Field of Dreams at various points – and their exhausted frustration is palpable from Jahn’s farewell quote to the New York Times: “If people don’t believe us after the results we’ve produced, then they never will.”
At its heart, the controversy over PEAR was a controversy over what science is and isn’t. Science is often defined more by what it excludes rather than what it incorporates, and even within an Ivy League institution there are struggles over where the lines fall that demarcate “real” science from the unreliable and unrespectable fringe. These lines are far from static, and depending on how things work out, history ends up praising the staunch objectivity of the individuals who defended these borders, or the creative open-mindedness of the individuals who redrew them. In PEAR’s view, the fact that they conducted their research with the proper experimental design and analysis and submitted their results to the peer review process was enough to qualify their findings as true science. Their opponents, on the other hand, appear to have viewed themselves as, the gatekeepers of scientific integrity, seeking any technical misstep or ambiguity to pop PEAR’s theoretical bubble.
Ironically, science writers are currently fascinated with studies that suggest a reversal of PEAR’s findings - the idea that computers will soon be able to “read our mind.” Perhaps if those dangerous three letters of ESP had never been associated with PEAR’s work, it’s possible the group could’ve better dodged their colleague’s dismissal and the media’s News of the Weird pigeonhole. But saddled as they have been with superimposed occult-studies residue, media mockery, and impenetrable terminology, PEAR’s anomalous effects will probably be forever segregated to the ghetto of Fringe Science, and be mere fodder for AP-wire stand-up humor.
