SkepticblogSkepticblog logo banner

top navigation:

Likelihood and the Paranormal

by Daniel Loxton, Oct 01 2013

Kyle Hill has a great cryptozoology post up on his But Not Simpler Scientific American blog today, titled “Why Bigfoot is Unlikely Only If You Know What ‘Unlikely’ Means.” The piece includes a shout-out for my new book with Don Prothero, Abominable Science!, and also includes some quotes drawn from an email exchange with yours truly. I thought I might share some more of those comments here for your interest:

Hill: The theme of the piece is “The Unlikely” and I want to look at how the un-likelihood of something is treated differently in science and pseudoscience.

Loxton: In some ways pseudoscientists and paranormal enthusiasts such as cryptozoologists assess probability very differently from scientists. A scientist generally starts with the conservative working assumption that proposed new ideas are not true or that hypothetical new entities do not exist, and then revises her probability estimate upwards only when the evidence forces her to do so. A pseudoscientist typically starts with the assumption that a novel proposal seems to be true, and then revises her probability downward as the evidence leaves her no choice—if she is willing to surrender the possibility to any degree at all. (Some paranormalists are very reluctant to do so. Despite all the exposed hoaxes and bad ideas in cryptozoology, for example, I’m not sure that any cryptid or class of evidence has ever been abandoned or ruled out by the community of cryptozoologists overall.)

Assuming we’re all willing to follow the evidence, the more conservative approach to assessing likelihood is not necessarily the only game in town. It depends if you are more worried about accepting falsehoods (a “Type I error”) or rejecting truths (a “Type II error”) . Psychologist Ray Hyman made this point in regard to a parapsychology enthusiast with whom he had engaged in dialogue:

Dave sees Type II errors as much more serious than Type I errors. For me, the skeptic, it is just the opposite. And this is not a matter of rational judgment as such. But it is more of a matter of how we weigh the costs of each type of mistake. Dave, like all believers in the paranormal, believes that the phenomena of reincarnation, psychic communication, mystical experiences, and the like are just too important and meaningful to overlook because of excessive demands of scientific rigor.1

Cryptozoologists do much the same.

Another systematic difference between how skeptical scientists and paranormal proponents assess likelihoods is in the weight that each gives to eyewitness testimony regarding alleged personal experiences with the paranormal. Proponents tend to see every eyewitness account as increasing the probability at least a little,2 while skeptics emphasize a proverbial warning: “the plural of ‘anecdote’ is not ‘data.'” A mountain of anecdotes is by itself just so much uncorroborated noise. Worse, new anecdotes can actually decrease the likelihood that a disputed phenomenon actually does exist. If eyewitness accounts implicitly entail predictions for evidence that consistently fails to emerge (if hunters see sasquatches, for example, then hunters can also shoot them—but they don’t) then new accounts diminish the chances rather than increasing them. This point was wisely argued by Bigfoot proponent Grover Krantz:

Many sasquatch enthusiasts seem to think that by finding more widespread [anecdotal and trace] evidence of the species, they are in effect strengthening the argument that the species is real. Up to a certain point this reasoning is valid … But when it is suggested that a wild primate is found native to all continents, including Australia, then credibility drops sharply. … Beyond a certain point, it can be argued that the more widespread a cryptozoological species is reported to be, the less likely it is that the creature exists at all.3

Krantz was one of the rare trained scientists to promote Bigfoot, so it is not surprising that he saw this point so clearly. Most cryptozoologists cheerfully ignore his warning. For most, the mountain of evidence only grows, forever. It never topples under its own weight.

In other ways, however, people all tend to assess probability the same, no matter what background we come from: when something seems overwhelmingly well attested, overwhelmingly consistent with previously admitted evidence, we stop thinking in probabilistic terms and start thinking of things as facts. To many Bigfoot enthusiasts, the existence of sasquatches seems like a given—a rock solid established fact, like the existence of France. To some, the accumulated mass of evidence seems just too large to explain away; to others, a visceral personal experience makes probabilistic arguments seem moot. “I know what I saw,” trumps abstract reasoning almost every time. And honestly, how could it not? That’s human. It’s the way we’re built.

References

  1. Ray Hyman. The Elusive Quarry: A Scientific Appraisal of Psychical Research. (New York: Prometheus Books, 1989.) p. 423
  2. Consider the advice from pioneering cryptozoologist Bernard Heuvelmans: “Needless to say, the more witnesses who can be produced for each particular sighting—and, in the long run, for each kind of hidden animal—the stronger the case. Essentially, this is because an abundance and unanimity of testimonies has perforce greater weight. … Accordingly, the inquiring cryptozoologist should endeavor, as his first important task in the field, to collect as many eyewitness testimonies as possible….” Bernard Heuvelmans. “The Sources and Method of Cryptozoological Research.” Cryptozoology: Interdisciplinary Journal of the International Society of Cryptozoology. Vol. 7, 1988. p. 5
  3. Grover Krantz. Bigfoot Sasquatch Evidence. (Surrey: Hancock House, 1999) p. 197

Note to Commenters: I invite and encourage civil discussion, scholarly debate, and open exchanges of ideas on this thread. At the same time, I expect all commenters to keep these useful principles firmly in mind. As on my other posts, I will delete posts that seem to me to be abusive. It’s not that kind of blog.

Like Daniel Loxton’s work? Read more in the pages of Skeptic magazine. Subscribe today in print or digitally!

5 Responses to “Likelihood and the Paranormal”

  1. Max says:

    When screening for cancer or explosives, a Type II error is worse than a Type I error because it’ll kill you.

    “A mountain of anecdotes is by itself just so much uncorroborated noise. Worse, new anecdotes can actually decrease the likelihood that a disputed phenomenon actually does exist.”

    I’ll repost my reply to a similar statement by Dunning.
    http://www.skepticblog.org/2011/09/01/sasquatch-the-quest/#comment-61592

    “mounds of bad evidence aggregate into a pretty strong indicator that the null hypothesis is true”

    Time for Bayesian Inference 101:
    H0 = null hypothesis
    H1 = alternative hypothesis
    P(H1)/P(H0) = prior odds
    P(evidence given H1)/P(evidence given H0) = Bayes factor
    prior odds * Bayes factor = posterior odds

    If P(evidence given H1) > P(evidence given H0), then Bayes factor > 1, so the evidence shifts the odds in favor of H1.

    I think it’s safe to say that a bigfoot sighting is slightly more likely if bigfoot exists than if bigfoot doesn’t exist, so it shifts the odds slightly in favor of bigfoot’s existence, and mounds of sightings shift the odds even more.
    On the other hand, mounds of failures to find good evidence such as bigfoot bones shift the odds in favor of the null hypothesis, because failure to find bigfoot bones is more likely if bigfoot doesn’t exist than if it exists.

  2. Max says:

    “A scientist generally starts with the conservative working assumption that proposed new ideas are not true or that hypothetical new entities do not exist, and then revises her probability estimate upwards only when the evidence forces her to do so.”

    What’s the initial probability estimate? If it’s zero or 100%, then no amount of evidence can change it.

  3. Luara says:

    The Oxford geneticist Bryan Sykes is studying some claimed Sasquatch DNA, and Adrian Erickson’s group released some new video from their Sasquatch habituation project recently.
    Either they come up with definitive evidence for Sasquatch existence or they don’t. Without the definitive evidence, both pro and con arguments are only so much verbiage.
    However, the “con” arguments have the disadvantage that they encourage prejudice and mockery in what should be a scientific question, and discourage serious attempts at investigating it. Creatures can remain undiscovered for a long time. I often come across news stories of new animals discovered, even in densely populated areas.
    The question of whether cryptids exist is a scientific question. It isn’t “paranormal”. The people who investigate such questions seem more scientific than the “skeptics” who remain at the sidelines and say the same things over and over again. Some of the investigators are professional scientists, and even the non-scientists who get out there with videocams etc. trying to gather evidence, are taking a more scientific approach than the skeptics.
    There’s a video on Shermer’s youtube channel, http://www.youtube.com/watch?v=gjuJbvQEgXc
    which is a cryptozoology discussion that is much more sophisticated and interesting than the usual verbiage that both skeptics and believers come out with.
    They discuss lake monsters like the Loch Ness monster, and the relative skeptic suggests that groups of otters playing follow the leader are a good explanation for many sightings.
    The relative believer says he doubts Bigfoot could exist in North America, but finds the Siberian yeti much more credible, the most likely of the cryptids to actually exist.
    But people like Sykes and the Erickson group are bringing the sasquatch subject out of the fog, and in time it will become clear whether there really is a new great ape out there, or only bunches of people (such as at the Sasquatch habituation sites) collaborating in a shared delusion.

  4. Max says:

    I don’t really care about Bigfoot, but when the police turn to the public for leads on a suspect, they get a lot of bogus tips. Yet you can’t argue that the more bogus tips they get, the more likely the suspect disappeared. It makes the TIPS less credible, but doesn’t say much, if anything, about the suspect’s whereabouts.

  5. Randy Grein says:

    Laura, it is possible to study cryptids scientifically, but the existence of most have been relegated to the realm of ‘paranormal’ because of evidence, both positive and negative over quite a long time. The negative evidence includes the sheer number of failed attempts to find significant evidence for what are among the largest creatures in an area long settled; the positive evidence includes a substantial amount of faked evidence purporting to support existence. Yes, there have been many attempts to forge evidence for bigfoot – footprints, footprint casts, hair samples and even video. (The most famous video taken in the 1970’s, if memory serves had me quite excited, until it was admitted to be a forgery.) After a reasonable period of time (in the case of bigfoot, over 100 years) the lack of credible evidence places the burden of proof squarely on proponents. Serious biologists have better things to do than chase down yet another swatch of fur to prove it’s bear rather than a hominid.