Download this as a Word DOC

The Online Staring Experiment has now been running since 2003. Here is a summary of the results from October 14th, 2003 to December 4, 2014.

Altogether 1751 pairs of people have taken part, giving a total of 35,020 trials. By chance people would be right 50% of the time. In fact the overall score was 56.4% correct. The statistical significance of this result is astronomical, with odds against chance of quadrillions to one.

Of the 1751 subjects who took part, 1031 scored 11 or more correct out of 20, in other words they were more right than wrong. 484 people were more wrong than right, scoring 9 or less out of 20, and 
236 people got 10 out of 20, exactly at the chance level. Again, the statistical odds against this result occurring by chance are astronomical.

There was a difference in the results in looking and not-looking trials. People were more successful at guessing correctly when they were being looked at than when they were not being looked at. In the looking trials the average hit rate was 60.1%, and in the not looking trials 52.7%. This agrees with the pattern shown in previous research on the sense of being stared at, summarized here.

I think this difference arises the sense of being stared at is likely to work best when we are actually being stared at, in other words in the looking trials. In the not looking trials, people are being asked to detect the absence of a signal, which is harder to do, and under those conditions many people are just guessing.

Some people did the trials without feedback (1086 pairs) while most people did them with feedback (664 pairs). With feedback the hit rate of 57.6% was very significantly higher than without feedback, 54.4%. This may be because the test is more engaging when feedback is given, and feedback may also help people to learn to be more sensitive.

Some subjects were blindfolded (932 people) and others were not (838). The blindfolded subjects had a significantly lower hit rate (55.5%) than those without blindfolds (57.3%) perhaps because they felt more relaxed, and perhaps because they were picking up subtle visual clues, even though they were asked to keep their eyes closed. However the fact that blindfolded subjects scored very significantly above chance shows that the ability to detect when they were being stared at cannot be explained in terms of visual clues

Since July 2005, this test has incorporated a sound signal given by the computer at the beginning of each trial. However, some people’s computers could not produce this signal, and so in some tests the lookers had to signal to the subjects by a click, bleep, or tap when the trials began. Such signals could convey subtle clues as to whether the looker was looking or not, which is why the computer gives a standardized sound signal to avoid this problem. In fact the 532 tests conducted without the standardized sound signal had a significantly higher hit rate (58.9%) than the 906 trials with the standardized signal (54.6%).

The effects of the relationship between the participants on the hit rates are shown in Table 1. Family members and colleagues scored lower than friends, and the highest hit rates were with teacher-student pairs, acquaintances and strangers. At first sight this is surprising. The pattern of results in telepathy tests is the opposite, in that family members tend to score much higher than acquaintances or strangers. But as I discuss in my book The Sense of Being Stared At, the ability to detect stares is different from telepathy in that it often occurs between strangers, and even between animals of different species. It probably evolved in the context of predator-prey relationships, whereas telepathy evolved in the context of communication between bonded members of social groups.

Table 1. Effects of the relationship between participants in tests for the sense of being stared at.

Relationship Trials Hits Hits %
Friends 13,620 7829 57.5
Spouse/partner 7260 4039 55.6
Parent/child 4460 2501 56.1
Siblings 2940 1646 56.0
Other family 4580 2421 52.9
Teacher/student 320 188 58.8
Colleagues 940 508 54.0
Acquaintances 320 201 62.8
Strangers 560 391 69.8

These experiments took place under unsupervised conditions, so it is impossible to eliminate the possibility that some people were cheating, or that some starers inadvertently gave subtle sensory clues to the subjects. However, these positive results are very similar to previous results under supervised conditions, and are a useful addition to the growing body of evidence for the reality of the sense of being stared at.

I am grateful to all those who took part. If you have not tried the test, do have a go!
Online Staring Experiment

Rupert Sheldrake

London, December 2, 2014.


I am grateful to Charles Overby and Ashwin Beharee for developing this test.