AN AUTOMATED TELEPHONE TELEPATHY TEST

Description: Telephone telepathy is the phenomenon whereby people can tell who is calling on the phone before looking at the caller ID or answering the call even when they could not have known by normal means who the caller was. This is very commonly experienced in real-life situations, but often dismissed as mere coincidence combined with selective memory, forgetting all the times that the person identifies the person wrongly. The phenomenon has now been investigated in a series of experiments where the subject has 2, 3, or 4 potential callers. The callers are selected at random by the experimenter or by an automated procedure, and the subject then has to guess which of the callers is on the line under conditions in which there is no clue from the caller ID system or any other normal means.

This project involves a new kind of automated telephone telepathy test. One of the participants registers online through Rupert Sheldrake’s website, www.sheldrake.org putting in his or her name, phone number, age and sex, and also the names and phone numbers of two friends or family members. As soon as this registration process is completed, the test begins. This test is programmed using a cloud computing system called Twilio. The participants can use either landlines or mobile telephones. The computer rings up all 3 participants who when they answer the phone and talk to each other, as in a conference call. They then hear instructions about the test and are given a 5 second warning before the first trial begins. There are six trials all together and all participants are reconnected with each other and can talk to each other between trials. The entire test takes about 15minutes.

When the first trial begins, one of the participants, say A, selected at random is told that he or she will be thinking about one of the other participants, selected at random, say B. B is told that he is being called and is asked to guess which of the other 2 participants is calling him. He has a minute in which to register his guess and does so by pressing number 1 for participant A and 2 for participant C. He is then connected to the person who has been thinking about him and they talk for about 10 seconds before the other person is brought back online and all three can chat for about 45 seconds until the next trial begins. Again the roles are assigned at random, and the next trial begins.

There are two variant versions of this test. In one of them, the person who is neither the caller nor receiver, in the example above, C, is told that she is neither the caller nor the receiver but asked to guess who the caller is, registering the guess by pressing 1 for A or 2 for B. In the second version, the person who is neither the caller nor the receiver is told she is neither the caller nor the receiver and listens to music. She is not asked to make any guess.

Participants for this are recruited through recruitment websites, or directly through Rupert Sheldrake’s website and are assigned version 1 or version 2 of this test at random, if recruited through Rupert Sheldrake’s website, or applicants to do the test recruited through recruitment websites are assigned 1 or the other version alternately. Thus the first applicant is assigned version 1, the second applicant version 2, the third applicant version 1, the fourth applicant version 2 etc.

The hypothesis being tested is that version 2 will be more effective than version 1, because in version 1 the person who is neither the caller nor the receiver, in this example C, is thinking about the other two people and may interfere with the telepathic detection abilities of the receiver, because the receiver might feel an influence not only from the caller, A, but also from the other person, C, who is trying to guess who the caller is. In version 2 of this test, where the person who is neither caller nor receiver is hearing music, this possible telepathic interference should be removed or reduced, and hence the accuracy of the receivers guesses should increase.

This experiment is being carried out under unsupervised conditions and it is therefore for participants to cheat. However none of the participants are told that there are two versions of the test or that the hypothesis being tested is to see if scores are higher with version 2 than version 1. Hence a consistent pattern of better performance in version 2 than version 1 would be extremely hard to explain by cheating.

NUMBER OF TRIALS
We plan in advance to conduct 500 trials with each version of the test.

STATISTICAL ANALYSIS
Overall hit rates on both versions of the test will be compared using 2x2 chi-squared test and also the Fisher Exact test.