Animal Predictors: Psychic, Sensitive, or Silly?

by Brian Dunning
April 29, 2014

In the wake of a popular 2014 hoax email going around claiming that animals were fleeing Yellowstone National Park in record numbers to escape an impending volcanic eruption, it probably makes sense to have a Skeptoid episode addressing animal predictions in general. Most are not hoaxes. That doesn't necessarily mean they're psychic, though. There are a range of possible explanations for the apparent ability. Perhaps the animals have some special sensitivity, perhaps it's an error made by the people who observe them. Today we're going to take a look at a few popular cases of famous, modern animals believed to have the power of prediction.

Oscar the Cat

In 2007, the media went wild over an article published in the highly respected scientific journal The New England Journal of Medicine claiming that a cat named Oscar was able to predict which patients at the Steere House Nursing and Rehabilitation Center in Providence, Rhode Island were about to die, and would curl up with them until they did. The story proved so popular that its author, Dr. David Dosa, a geriatrician at the Center, was offered a book deal and expanded the story of Oscar's amazing predictive ability into a 240-page book, Making Rounds with Oscar: The Extraordinary Gift of an Ordinary Cat. Oscar's story has since been included in virtually every list of psychic animals in every kind of media, and is often cited as proof that the ability exists, particularly due to its publication in such an esteemed journal.

But please, hold the horses a moment. The opening section of the Journal is called Perspectives, and includes essays, editorials, and opinion pieces. Dosa's article was in this section; it was most certainly not presented as research, but simply as a fun anecdote. Dosa made no representation that it was either scientific or based on serious study of the cat's behavior.

By the time of the book, Dosa said some 50 deaths at the Center had been preceded by visits from Oscar. But as many science journalists have noted, no data was ever collected or analyzed. No mention was made of how often Oscar visited other patients. Since it's a nursing home, most patients are terminally ill and remain there until they die, so it's hardly even possible for Oscar to ever be wrong. No criteria were ever observed for the length of time between Oscar's last visit and the patient's death, the duration of Oscar's visit, or how those numbers compared to his visits to other patients. Moreover, Dosa even states in the book that "for narrative purposes" he "made some changes that depart from actual events".

From what we know of Oscar, there is no need to suggest that he has the power of prediction, either psychic or based on some smelling ability or behavioral sensing. Oscar's story can almost certainly be explained by confirmation bias: the tendency of workers at the center to more strongly notice Oscar's actions when they confirm the belief, in exactly the same way that many hospital workers notice busier nights during a full moon, a notion that's been conclusively disproven. But we can't know for sure since nobody has ever studied the way Oscar divides his time between the living and the dying. Until they do, we have a cute story, but certainly not a psychic cat.

Punxsutawney Phil

One of the oldest and best known animals that made predictions is Punxsutawney Phil, centerpiece of the American tradition of Groundhog Day. Each year on February 2 at sunrise, a groundhog named Phil is brought out in Punxsutawney, PA to see whether he casts a shadow. If he does, the winter will last six more weeks; if he does not, then an early spring is expected.

Punxsy Phil is, of course, just for fun; nobody really takes his predictions seriously. And that's good, because he's wrong about 2/3 of the time, according to the records kept by the Punxsutawney Area Chamber of Commerce. At 50% accuracy, a coin toss would be better than Phil.

But Phil is a great example of how statistics could be used to support whatever conclusion you want; something of which we must be vigilant for any of these stories of animal predictions. Take the date six weeks after February 2, and list the historical temperatures for each. Or the cloud cover, or the rainfall, or whatever indicator you want; but let's use temperature for this example. Find the mean temperature, go as far above or below that as it takes to give Phil the desired right/wrong record. Presto; just by choosing the criteria that support our preferred conclusion, it would be easy to prove that Punxsutawney Phil has a statistically significant ability to predict either a long winter or an early spring.

Using similar after-the-fact analysis, we might also come up with a claim such as:

Rupert Sheldrake's Psychic Dogs

In 1994, English biochemist and paranormal researcher Rupert Sheldrake surprised the worldwide press with his announcement that dogs used psychic powers to determine when their owners were about to return home. He did extensive observation on one dog named Jaytee, and determined that Jaytee would walk to the porch shortly before his owner would return home. Sheldrake's work was criticized by Professor Richard Wiseman at the University of Hertfordshire, in large part due to this same kind of after-the-fact analysis. Sheldrake had, according to Wiseman, collected data first, and then afterward looked for patterns. Just as it would be to establish winter/spring criteria after comparing Punxsutawney Phil's predictions to actual temperature records, it would be similarly easy to establish criteria for how long before the owner's return home Jaytee would have to go to the patio to constitute a successful prediction.

Wiseman carried out four sets of his own experiments with Jaytee at the owner's house, having first established parameters for what kind of behavior would count as a successful prediction, using Sheldrake's conclusions: Jaytee had to go to the porch and remain there for at least two minutes. There proved to be no significant correlation between Jaytee's behavior and the owner's arrival home. The dog would come and go all day long.

Sheldrake's methodology was akin to what we term the Texas Sharpshooter fallacy: shoot at the side of a barn, and then draw a target around your bullet hole. Wiseman drew his target first, and Jaytee's behavior missed the marked completely.

Paul the Octopus

Paul the octopus lived in a tank at Sea Life Centre in Oberhausen, Germany. In 2008, during the European Football Championship, they put two boxes of food in his tank, one labeled with Germany's flag, and the other with that of Poland, whom Germany was about to play. Paul went to the Germany box first, and sure enough, Germany won that game. Throughout the tournament, Paul correctly predicted 2/3 of the matches he was given. But during the 2010 World Cup, Paul was correct 100% of the time, a feat that seems impossible unless he was truly inspired.

Many news outlets consulted mathematicians, who generally used the coin toss analogy. Paul's chances of correctly picking the 2010 World Cup were 1/64, and with hundreds of animals being used as oracles around the world, it was a virtual certainty that at least one of them would guess all eight matches correctly. And that one is the one we all remember.

But the coin toss analogy is not correct. Paul was given only matches in which Germany played, and since Germany was a top team (eventually winning second place), it was more likely they'd win any given match. Octopus are extremely intelligent, and though they're colorblind, they do recognize shapes. In his six 2008 trials, Paul simply swam to the Germany box every single time, and was wrong two out of the six. Once Paul had learned that the Germany box contained food, it only makes sense that he'd go straight to the one marked with the bold stripes of the German flag each time.

In the 2010 World Cup, Paul continued to choose Germany five out of seven times. In the eighth and final match, Germany did not participate, so Paul had a 50/50 chance on that one, and he guessed it correctly. Paul was correct all eight times. Most of these may be attributable to his trust of the German flag and Germany's winning tendencies, but that doesn't explain his misses. Octopus are imperfect, apparently. However it is noteworthy that in all three instances of Paul's career that he did not pick a box marked with a German flag, the box that he picked also had a national flag with three bold horizontal stripes. Three of the five times he picked Germany in his perfect 2010 season, it was against a country whose flag was very different.

Was Paul truly psychic, or was it all random chance? Probably neither. He simply had an octopus' excellent eyesight and intelligence, and knew that he'd find food wherever he saw three bold horizontal stripes. Are other animals similarly able to sense their surroundings? How about:

Snakes and Earthquakes

We've all heard stories that dogs bark, herds stampede, lowland animals rush to higher ground, and all sorts of other animals exhibit some behavior preceding an earthquake. Do they have a psychic ability, or is it perhaps something to do with heightened senses detecting real physical phenomena? Of all these cases, snakes and their ability to predict earthquakes is probably the most common human belief.

Most research on this subject comes from China and Japan, two earthquake-prone countries with deeply rooted cultural beliefs of special spiritual powers in animals. Correspondingly, whenever you find a news report of such research, it generally comes from one of those two countries.

There are potential physical explanations for snakes to sense earthquakes. Snakes are often mentioned because their entire body is in contact with the ground all the time, and they're known for their sensitivity to nearby footsteps or other disturbances. The heavy earthquake waves that we all feel are called S waves, but they're preceded by faster waves called P waves. P waves and S waves are triggered at the same time when the quake happens, but the P waves travel faster; so if you're some distance from the epicenter, the P waves will reach you earlier. Humans can't generally detect these, but it's possible more sensitive animals could. In any event, the delay is a matter of seconds; not hours, days, or weeks. But large earthquakes are sometimes preceded by smaller foreshocks, which again humans may not feel, but more sensitive animals might.

There are two problems with this idea though. First, the animal behavior we attribute to earthquake foreknowledge is rarely reported when the big earthquake hits that we can feel. If the animals are reacting to small foreshocks, shouldn't they also react to a larger shock? Second, virtually all of the evidence for this is anecdotal. It's based on reports from people and comes after the big earthquake. Such anecdotal evidence is untestable, unverifiable, and subject to all manner of perceptual errors like confirmation bias and post-hoc rationalization. "I saw my dog run into the barn the other day, therefore he must have known about today's earthquake."

To summarize, we don't have any good evidence yet that any animals consistently show any kind of predictive ability. We have good explanations for why they do what we think they do, with no need for the introduction of a special power of intuition or sensitivity. Of course there absolutely might be effects out there we don't know about yet, but until we have a reliable observation in search of an explanation, you should always regard these stories of psychic animals with a touch of skepticism.

By Brian Dunning
Follow @BrianDunning