¿ìè¶ÌÊÓƵ

Why do so many people believe in the paranormal? The answer to this
question, and the recent research exploring it, tell us little about the
paranormal itself but much about the way our minds work.

There have been many surveys of belief in the paranormal. The proportion
claiming belief varies with the sample and the question asked but is usually
well over half. More interesting is the main reason given: that people have
had psychic experiences themselves.

There are three obvious explanations for this. First, they might really
have experienced the paranormal. If true, we need to rewrite much of science,
and soon. Secondly, they might be making it up. For anyone who has had these
experiences this does not seem a plausible explanation. Thirdly, they might
be misinterpreting perfectly normal events – suffering from what we might
call a paranormal illusion.

We often see things that are not there. Stunning visual illusions happen
because of the way our visual systems try to make sense of the world. Even
when we know how the illusion works, it does not go away. The illusions
are the price we have to pay for a perceptual system that does very well
in a confusing world. Paranormal experiences may be analogous; the price
we have to pay for the way our brains look for connections in chance and
probability.

In describing psychic experiences, people typically claim that they
dreamed of something that came true or they ‘knew something was wrong’ with
their mother, sister or great-grandfather, and it was. In other words, something
happened that was far too unlikely to have been ‘just chance’. So it must
have been psychic.

Put like that, it is easy to see how a mistake can happen. The conclusion
that the event was paranormal depends on a probability judgment and people
are notoriously bad at making them. If you think you are an exception try
some of the questions in the Box.

Psychologists have long known that people use a variety of heuristics,
not mathematics, to guess probabilities. These heuristics include ‘representativeness’
and ‘availability’ – the ease with which examples can be brought to mind.
Such errors might not be so bad except that people also tend to have great
confidence in their erroneous judgments, even in the face of contrary evidence.
Could it be that all belief in the paranormal comes from misjudgments of
probability? Recent evidence suggests that it may be so.

Paranormal phenomena fall into two classes. The first is ESP, or extrasensory
perception – the pseudo-sensory aspect of the paranormal. It includes telepathy,
clairvoyance and pre-cognition. If this does not seem to be probability
in action, consider an example.

The most common form of spontaneous ESP is the dream that comes true.
Imagine you dream of your friend Maximillian. In the dream you are at his
funeral and his friends and relatives are weeping at his death. The next
day you hear that Maximillian died that very night. For most people this
seems too much of a coincidence to be just chance.

The problem with chance is that it is very hard to understand. In the
case of the dream, could we work out the chances? Trying to do so will show
up some of the problems.

How often do you dream of your friend dying? You would probably say
you never had that dream before. So let us take a conservative estimate
and assume that you have such a dream once in a lifetime. Next we need to
know how likely it is that man died the same day as your dream. Or do we?

Thinking this way we are already committing the most obvious but easily
overlooked fallacy. It all depends on how much information you have before
you start. If you had woken from the dream, immediately written it down,
sent it off to the Society for Psychical Research, and if you had never
done that before, then you would be justified in asking ‘how likely is it
that Maximillian will die now?’ Naturally the answer would be ‘very unlikely’
and you might need to seek some alternative explanation to ‘just chance’.

But it is far more likely that you noticed the coincidence only when
you heard about his death. So this kind of coincidence might have happened
to anyone on any day. Now the calculation is different. How often would
you expect anyone to dream of a friend dying the night that it happened?

Statistician Christopher Scott has analysed it this way. There are about
55 million people in Britain and they live about 70 years each. If each
has one such dream in a lifetime there should be 2000 every night. Also
about 2000 people die in each 24 hours. So there will be 4 million coincidences
among 55 million people. In other words such an ‘amazing’ coincidence will
be expected about once every two weeks. No further explanation is required
but try to believe that if it happens to you. As Scott puts it: ‘You, if
you’re human at all, will conclude that there’s cause and effect there.
You can’t regard yourself as just one among 55 million.’

So most people reject the chance hypothesis and if they can find no
‘normal’ alternative they will turn to the idea of ESP.

But what kind of explanation is that? I would say none at all. Psychical
research has been under way for more than 100 years, and parapsychology
(its more laboratory based counterpart) for more than 50. Yet they have
made virtually no progress in understanding what, if anything, ESP is. There
are certainly some impressive experiments. Recent fully automated ganzfeld
tests at Princeton are an example.

The idea is to isolate the subject from ordinary sensory input by covering
the eyes with halved ping-pong balls and playing white noise through headphones
while they relax on a comfortable bed and describe their imagery. Meanwhile
a ‘sender’ is looking at a target picture chosen randomly from a set of
possible targets. After the half hour or so is over the subject has to try
to match up the imagery experienced with each of the pictures and choose
which was the target.

Earlier experiments were critised for all sorts of shortcomings but
the recent ones have been automated at every stage of the procedure – and
still the subjects seem to pick the right picture more often than expected
by chance. (Here we are back to chance again.) There are some who still
argue that the results could be spurious and this argument is not likely
to be easily resolved. But either way this still does not give us an explanation.
If there is an ‘extrachance’ effect here, what is it?

We get nowhere by saying that something happened ‘by ESP’. We have no
acceptable theory of what ESP is, what constraints apply to it or when and
how we can expect it to operate. After a century of research on ESP, it
is still negatively defined. To say that the dream was due to ESP is only
to admit our ignorance.

So people who experience amazing coincidences are either subject to
some totally mysterious new phenomenon, or they are suffering from an illusion
of probability. How could we find out?

Tom Troscianko and I, at the University of Bristol, hypothesised that
if the origin of belief in ESP lies in misjudgments of probability, then
we would expect believers (usually referred to as ‘sheep’) to be less accurate
in their probability judgments than goats (disbelievers). This we tested
by giving schoolchildren, university students, medical workers and others,
a set of computerised tests .

In general the goats did better at these tasks, as we predicted, and
as is consistent with the idea (but does not prove) that the sheeps’ errors
are responsible for their belief in ESP. Interestingly, the university students
did no better than the schoolchildren which implies that these judgments
are not improved by education. Another well-known error lies in ‘subjective
random generation.’ Put simply, most people have no idea of how random numbers
behave. When they are asked to generate a string of random numbers many
people avoid repeating the same digit twice – it is as though they think
that this would not be random.

At the University of Zurich in Switzerland, Peter Brugger and his colleagues
have been exploring the relationship between this error and belief in the
paranormal. In keeping with our hypothesis they found that sheep avoided
pairs more than goats did – in both real ESP experiments and in tests of
random string generation. They suggest that most, if not all, of the major
findings in parapsychology can be attributed to errors in random numbers
generation or response bias.

For example, it is claimed that young children do better at ESP tests
– the children are more biased. Extroverts apparently do better at ESP –
they are more biased than introverts. Brain injury and psychiatric disorder
are associated with the paranormal – and also with increased bias. And finally
there is the famous decline effect: that ESP scores are found to decrease
with longer trials – so does bias.

Of course this could explain the parapsychological findings only if
there were inadequacies in the randomisation procedures used. In a perfect
ESP experiment the subjects’ bias should make no difference. So to refute
this idea parapsychologists will have to show that their findings appear
just as strongly in the best controlled ESP tests.

The other major kind of paranormal effect is psychokinesis (PK) or ‘mind
over matter.’ This may seem even further from being a matter of probability,
but oddly it is not.

Imagine you are driving along the road in a hurry towards some red traffic
lights. You ought to slow down. If they stay red you will be going too fast.
‘Change, change’ you are tempted to mutter, as though this would make a
difference. If they do change at that precise moment it is tempting to feel
that you have made it happen.

This feeling of being in control can easily overwhelm logic. The most
bizarre example I ever experienced was taking part in a ritual to make the
sun rise at dawn on Midsummer Day. After hours of chanting and processing,
when the sun popped up, dead on schedule, we really felt as though we had
made it happen!

This temptation to attribute random events to one’s own actions is called
the illusion of control. First described by Ellen Langer, it appears in
all sorts of behaviours. In a way it is the equivalent of the tendency to
want to make sense of coincidences. We learn how to control the world around
us by observing coincidences between our own actions and the things that
happen. Just as with visual illusions, the processes that help us to learn
carry a cost – we sometimes associate things that are actually unconnected.

Could it be that this underlies a belief in having PK powers? The idea
is simple. A coincidence happens between an action we make and some external
event. We make a connection between them and look for an explanation. If
there is no obvious normal one (I wasn’t touching the thing, whistling at
it, predicting its actions by inference and so on) then it must be paranormal.
So I have experienced PK.

We may be disappointed if we look to parapsychology for an explanation
of the coincidence. The most recent hope for evidence of PK lies in experiments
on random number generators. A computer is typically connected to a source
of random numbers. A subject then attempt to bias the output in one direction
or another. Dean Radin, at Princeton University, recently analysed all 597
such experiments in the literature. Many of the experiments included a control
condition in which no one was trying to influence anything. He found that,
‘results showed effects conforming to chance expectation in control conditions
and unequivocal non-chance effects in experimental conditions’ and he took
this as evidence of a ‘consciousness-related anomaly in random physical
²õ²â²õ³Ù±ð³¾²õ.’

Does this point to a possible explanation for people’s PK experiences?
There are two issues here. One is the size of the effect which, in this
database, was so tiny that it is extremely doubtful anyone would notice
it in everyday life. In this sense parapsychology fails to provide an explanation.
Also this research provides no theory as to how or why the random numbers
might behave differently when someone is trying to influence them. It is
not even clear why consciousness should be involved.

If the alternative is that people are misjudging probabilities, how
could this be tested? We hypothesised that sheep should be more prone to
an illusion of control than goats. This would then lead them more often
to think they had exerted PK.

Previous experiments had shown that sheep thought they had more control
in a PK task. However, if sheep believe they actually have PK (and we cannot
rule out the possibility that they do) then it is quite reasonable for them
to think they are in control. So this result is only to be expected. We
wanted to know whether the greater illusion of control would also apppear
in tasks which did not appear to be PK tasks.

At the University of Bristol we asked people to play a computer game
where they had to stop a flipping coin on the desired face. They pressed
a button and then the coin stopped flipping either a fixed number of times
after the press, or a random number of times, although the subjects were
not aware of this difference. We predicted that in the random conditions
(which amounts to PK) sheep would still think they had more control than
goats. This is indeed what we found.

We also found something else much more interesting. We asked the subjects,
if they imagined doing the task with their eyes closed, how many hits they
would get. The average should be 10 and indeed most of the goats said 10
but less than half the sheep did, with an average of 7.9. We called this
the ‘chance baseline shift.’

The implications of this are quite interesting. If someone thinks that
by chance they will get 7 hits and they actually get 10, they are going
to think they have done very well and look for an explanation. Obviously
no reasonable explanation will be found since only chance was operating.
So the obvious leap is to the paranormal.

I think all belief in the paranormal may come about this way: as an
illusion of casuality. It is not stupid to have apparently psychic experiences,
any more than it is to see visual illusions. It just reflects the way our
minds work. So does the paranormal ‘really’ exist? I have no idea. I can
only say that the fact that so many people have psychic experiences is not
evidence that it does.

Test your mental powers on these probability puzzles

1) A hat contains a large number of pieces of paper with the numbers
1, 2, 3, 4 and 5 on them in equal proportions. Write a list of 30 numbers
in the order you think they might be drawn from the hat.

2) A hat contains 10 red and 10 blue Smarties. I pull out 10 and 8 are
red. Which am I more likely to get next?

3) A box contains green and yellow buttons in unknown proportions. I
take out 10 and 8 are yellow. Which am I more likely to get next?

4) In the last four matches between Mytholmroyd Athletics and Giggleswick
United, Mytholmroyd have kicked off first every time, on the toss of a coin.
Which is more likely to kick off next time?

5) How many people would you need to have at a party to have an even
chance that two of them will have the same birthday (not counting the year
of birth)?

These questions are based on those used in the Bristol experiments.
The first question asks for a string of random numbers. A truly random string
would, on average, contain just under 6 repeats of the same digit. Most
people produce far fewer repeats. The answer t otheothers are blue, yellow,
either and 22. Most people give much higher answers for question 5. The
reason may be that they’re thinking of it in terms of the question: ‘How
many people would you need to have an even chance that one has the same
birthday as me?’ This egocentric view underlines much of our confusion over
probabilities.

Dr Susan Blackmore is in the department of psychology of the University
of Bristol and the University of Bath.