described above, does not serve any useful purpose and drags down the credibility of everything else published by that journal.
As with the creationists described in chapter 2, proponents of ESP claim that their results are unfairly rejected because of conventional science’s dogmatic attachment to old ideas. My reaction is the same as it was in the case of intelligent design: what possible reason would scientists have to object if convincing evidence for psychic phenomena was reported? As with intelligent design, the discovery of special powers of the mind would open up wonderful new avenues of research that would surely be generously funded by taxpayers. Mainstream scientists have not accepted the claims of parapsychology for exactly the same reason they have not accepted the claims of intelligent design.
The data do not warrant it.
From the first experiments in the mid-nineteenth century to the present, the claim of evidence for ESP simply does not stand up under the same scrutiny scientists apply when considering any extraordinary claim.
The Significance of Experiments
Let me expand on the issue of statistical significance of experiments, which is the basis on which many reported extraordinary claims can be quickly discarded. Parapsychologists argue that they should be held to the same standard of statistical significance as medical science, where claimed positive effects of, say, a new drug, are published when the statistical significance (”
P
value”) is 5 percent (
P
= 0.05) or lower. That is, if the experiment were repeated many times in exactly the same fashion, on average one in twenty would produce the same effect, or a greater one, as an artifact of the normal statistical fluctuations that occur in any measurement dealing with finite data.
But think of what that means. In every twenty claims that are reported in medical journals, on the average one such report is false—a statistical artifact!
Contrast this with the standard in the field of research where I spent my career, elementary particle physics. There the standard of
P
value for publication of an important new discovery is one-hundredth of one percent (
P
< 0.0001). This guarantees that, on average, only one in ten thousand such reports is a statistical artifact.
A possible justification for the low standard in medicine may be that medical journals are not venues for extraordinary new discoveries but places where promising new therapies are disseminated to the healthcare community as rapidly as possible. If one in twenty are spurious, that may be regarded by some as a small price to pay if a life might be saved by a therapy that works. Nevertheless, I think the medical standard should be higher, given the large number of false reports that are later withdrawn. Think of all the wasted money, effort, and lives that probably go into useless therapies under the current arrangement.
Indeed, medical researchers are beginning to recognize the inadequacy of their journal standards. Epidemiologist John Ionnidas has gone so far as to write, “Most published research findings in Medicine are false 27 .” A recent paper in
British Medical Journal
recommends the
P
value threshold be changed to
P
< 0.001, not as tight as in physics but probably suitable for medical science, given all its added complications 28 .
Parapsychologists, on the other hand, are not in the business of saving lives. They are more like particle physicists or astronomers, seeking to uncover facts about the fundamental structure of nature, where no one will die if the report of an important discovery is held off for a few months or years.
Almost without exception, claims of evidence for psychic phenomena come nowhere close to having the statistical errors small enough to rule out more mundane explanations for the results 29 . The handful that claim reasonable statistical significance all have methodological flaws that render their results unconvincing. And none are independently
As with the creationists described in chapter 2, proponents of ESP claim that their results are unfairly rejected because of conventional science’s dogmatic attachment to old ideas. My reaction is the same as it was in the case of intelligent design: what possible reason would scientists have to object if convincing evidence for psychic phenomena was reported? As with intelligent design, the discovery of special powers of the mind would open up wonderful new avenues of research that would surely be generously funded by taxpayers. Mainstream scientists have not accepted the claims of parapsychology for exactly the same reason they have not accepted the claims of intelligent design.
The data do not warrant it.
From the first experiments in the mid-nineteenth century to the present, the claim of evidence for ESP simply does not stand up under the same scrutiny scientists apply when considering any extraordinary claim.
The Significance of Experiments
Let me expand on the issue of statistical significance of experiments, which is the basis on which many reported extraordinary claims can be quickly discarded. Parapsychologists argue that they should be held to the same standard of statistical significance as medical science, where claimed positive effects of, say, a new drug, are published when the statistical significance (”
P
value”) is 5 percent (
P
= 0.05) or lower. That is, if the experiment were repeated many times in exactly the same fashion, on average one in twenty would produce the same effect, or a greater one, as an artifact of the normal statistical fluctuations that occur in any measurement dealing with finite data.
But think of what that means. In every twenty claims that are reported in medical journals, on the average one such report is false—a statistical artifact!
Contrast this with the standard in the field of research where I spent my career, elementary particle physics. There the standard of
P
value for publication of an important new discovery is one-hundredth of one percent (
P
< 0.0001). This guarantees that, on average, only one in ten thousand such reports is a statistical artifact.
A possible justification for the low standard in medicine may be that medical journals are not venues for extraordinary new discoveries but places where promising new therapies are disseminated to the healthcare community as rapidly as possible. If one in twenty are spurious, that may be regarded by some as a small price to pay if a life might be saved by a therapy that works. Nevertheless, I think the medical standard should be higher, given the large number of false reports that are later withdrawn. Think of all the wasted money, effort, and lives that probably go into useless therapies under the current arrangement.
Indeed, medical researchers are beginning to recognize the inadequacy of their journal standards. Epidemiologist John Ionnidas has gone so far as to write, “Most published research findings in Medicine are false 27 .” A recent paper in
British Medical Journal
recommends the
P
value threshold be changed to
P
< 0.001, not as tight as in physics but probably suitable for medical science, given all its added complications 28 .
Parapsychologists, on the other hand, are not in the business of saving lives. They are more like particle physicists or astronomers, seeking to uncover facts about the fundamental structure of nature, where no one will die if the report of an important discovery is held off for a few months or years.
Almost without exception, claims of evidence for psychic phenomena come nowhere close to having the statistical errors small enough to rule out more mundane explanations for the results 29 . The handful that claim reasonable statistical significance all have methodological flaws that render their results unconvincing. And none are independently
Similar Books
