From Jim Holt's review of
by Georges Charpak and Henri Broch (Johns Hopkins Univ. Press) in today's
Wall Street Journal
Have you ever had a premonition? Did you once have, say, a passing thought about an uncle, only to receive a phone call five minutes later informing you that the beloved relative has dropped dead? If so, this probably struck you as eerie. You might have vaguely believed it was ESP.
Was it? Let's see. Suppose you know of 10 people who die each year. Furthermore, suppose you think of each of them once annually. There are 105,120 five-minute intervals in a year. A simple probability calculation shows that there is a 10 in 105,120 likelihood that you will, as a matter of chance, have a thought about one of these people in the five minutes before you hear of their death. Multiply this likelihood by the population of the U.S. (about a quarter of a billion people) and you find that roughly 25,000 people each year -- about 70 a day -- will have a "psychic" experience of this sort. In fact, it's pure coincidence.
One can quibble with Holt's back-of-the-envelope calculation, but the underlying point is a good one.
It seems to me that this is an incorrect interpretation of the phenomenon since the feeling in question is not based on how often it happens to other people, but to yourself. Going through a similar, and more generous time allowances, we can estimate that there is a 10 in 365 chance for a person to think of the deceased within 24 hours preceeding their death. That comes to roughly once every 36 years - thus at least once in a lifetime for most people. If one were to go into the 5-minute intervals as the original author suggested, a chance per individual is about one in 10,000 per year. If we give a person 50 years in which deaths do come as a surprise (figuring childhood does not count and after a certain age you think about it too much to really count). Subsequently, if my lifetime chance is about one in 2,000, then I would need to discuss this occurrance with more than 2,000 people before finding another one who had a similar experience. I think this works for most people to consider as a pretty unique occurance.
Obviously, one can build more elaborate models pretty easily, but I think the point remains - even if one is far from unique in having had the experience, it *is* a rare one for an individual involved.
ps. much like lightening striking the earth 100 times every second * does not mean that some places get very little lightning and some a great deal. Same goes for rain.