by Rupert Sheldrake
The Skeptic Vol. 16, No. 4, pp. 8-13 (2004)

Editorial Introduction
Following David Marks' article on the subject of unseen staring and the ESP ability of pets (see The Skeptic 16.3), we now hear Rupert Sheldrake's radical views on these topics. Needless to say, Marks and Sheldrake vehemently disagree. For example, Sheldrake thinks that not only is the claim that people - and animals - can tell when they are being stared at by unseen others true, but that it may have an evolutionary basis, because prey animals that could detect when a predator is looking at them would have a better chance of survival than animals that did not. Do Sheldrake's and Marks' fundamental differences of opinion derive from the different disciplines they trained in? Is it possibly the case that Sheldrake is even more sceptical than the sceptics? Whatever the answers to these questions are, Sheldrake is certainly a stern iconoclast.

David Marks' comments on my research on staring and on return-anticipating dogs are shortened versions of Chapters 8 and 9 of his book The Psychology of the Psychic (2000). They were outdated and misleading when they were first published, and three years later are even more so.

In his book Marks quantified his subjective estimate of the reality of various "paranormal" phenomena as 'infinitesimally low' (p. 306). The chapters of his book resemble a series of show trials in which a guilty verdict is a foregone conclusion; the only question is on what grounds the apparent evidence is to be condemned. He tells us he wrote the book "in the hope that the fantasies, fictions and figments of the paranormal, cleverly disguised as facts, will in future be seen for the impostors that they are" (p. 19).

Marks is a very unreliable guide to research in these areas. He treats the questions he writes about as irritating distractions from what he calls Urgent and Serious Problems, and is impatient with the factual details. He does not seem to have read the relevant papers, published in peer-reviewed journals, and disregards evidence that does fit in with his beliefs.

The sense of being stared at
As Marks pointed out, several surveys have shown that the majority of people claim to have detected when they were being stared at by unseen others. This is not an "exceptional experience", but appears to be quite common; it is not paranormal but normal, in the sense that happens quite commonly to ordinary people, and also to many species of animals (Sheldrake, 2003). I think its evolutionary basis may lie in relations between predators and prey. Prey animals that detect when a predator is staring and escape would tend to survive better than animals that did not.

My experiments on the sense of being stared at have given consistent positive and highly significant results, and have now been widely replicated. Anyone interested can try this simple experiment using the instructions on my web site ( www.sheldrake.org).

There is also a growing body of evidence from a parallel stream of research using closed circuit television (CCTV), that people can unconsciously detect when they are being looked at on a TV monitor by an observer in a remote location, as measured by changes in skin resistance. This work has replicated in several different labs (for a recent review, see Delanoy, 2001). I return to a discussion of some recent CCTV research below.

In my own tests, people work in pairs. One serves as the starer, the other as the subject. In a randomized series of trials, the starer either looks at the back of the neck of a subject, or looks away and thinks of something else. The subject then guesses whether he or she is being looked at. The subject is either right or wrong. By chance, 50 per cent of guesses would be correct.

In the looking trials, the scores are generally positive and highly significant, around 60 per cent correct, while in the not-looking trials they are at chance levels (Figure 1A). These results have been replicated in many trials in schools and universities (Sheldrake, 2002).

Similar results were obtained whether or not the subjects were given feedback (Sheldrake, 1998, 1999a, 2000a, 2001a) and also in trials conducted through closed windows and at distances up to 100m, eliminating auditory or olfactory cues (Sheldrake, 2000a). The same pattern of results was obtained using coin-tossing for randomization (Sheldrake, 1998, 1999a), and with randomized instruction sheets prepared in advance (Sheldrake, 1999a, 2000a, 2001a).

These results are consistent with the reality of a sense of being stared at. Such a sense would be expected to work when people are in fact being stared at. But it would not be expected to work in not-looking trials. We do not have a sense of NOT being stared at; subjects would just be guessing under these conditions; results at chance levels would be expected.

This pattern of results argues against explanations in terms of subtle cues or cheating, which would lead to an elevation of scores in BOTH types of trial. This pattern of results thus provides internal evidence against the subtle cue and cheating hypotheses.

Following methods similar to my own, Marks's colleague John Colwell and his associates (Colwell, Schroeder & Sladen, 2000) obtained the same pattern of results that I and others have found (Figure 1B).

Figure 1.
A.   Results from staring trials conducted by Sheldrake and others in Europe and the United States (data from Table 5, Sheldrake, 1999a). There were 13,900 trials altogether; for the totals, p < 1 x 10-15.
B.   Results from staring trials conducted by Colwell et al. (data from Table 1, Colwell et al., 2000, for trials with feedback). There were 2,160 trials altogether; for the totals, p < 0.001.

Marks (2000) and Marks & Colwell (2000) were faced with the problem of explaining why these findings replicated my own. They speculated that subjects might have learned implicitly to recognize patterns in the randomized sequences used in their trials. These particular counterbalanced sequences were downloaded from the New Scientist web site, and were the same as some of those I used in some of my own trials. They proposed that because these sequences deviated from "structureless" randomizations, subjects who were given feedback could have learned implicitly to detect patterns in the sequences, thus enabling them to guess at above-chance levels. They announced their hypothesis as if it were a fact in the title of their article in the Skeptical Inquirer : "The psychic staring effect: An artifact of pseudo randomization."

This Marks-Colwell hypothesis is fatally flawed for four reasons:

1. Marks (2000) and Marks and Colwell (2000) were apparently unaware that their implicit learning hypothesis had already been refuted by thousands of trials involving structureless randomizations (Sheldrake, 1998, 1999). where implicit learning would have been impossible. In addition, a computerized staring experiment has been running at the New Metropolis Science Museum in Amsterdam since 1996, and more than 18,500 subjects have taken part. A program in the computer provides structureless randomizations for the sequence of looking and not-looking trials. The results are positive and astronomically significant statistically.
2. The implicit learning hypothesis has been refuted by thousands of trials with no feedback, with the usual pattern of positive and highly significant results (Sheldrake, 2000a). Implicit learning depends on feedback, and hence cannot explain these results.
3. . If implicit learning led to positive scores in looking trials, then it should also have done so in not-looking trials. But it did not (Figure 1). Why not? Marks did not mention this problem; perhaps he hoped his readers would not notice it.
4. In Colwell et al.'s experiment, the same subjects took part in nine successive 20-trial sessions with feedback. There was a statistically significant learning effect in successive sessions, but only in the looking trials, not in the not-looking trials. This is consistent with the subjects learning to detect stares more effectively. But such learning would not have been possible in the trials I conducted. Each subject was tested only once, in a single 20-trial session, and hence the learning hypothesis cannot account for the experimental data shown in Figure 1A.

In an attempt to support the implicit learning hypothesis, Colwell and his colleagues did a second experiment, with a different person doing the staring. This time they used 'structureless' randomizations.

As they expected, the result was non-significant. Marks and Colwell jumped to the conclusion that this was due to the different randomization method. They did not mention the change of starer. In the first experiment the starer was a graduate student with a sympathetic interest in the sense of being stared at; in the second the starer was one of Colwell's colleagues.

Unfortunately, the design of this experiment was confounded. Two variables were changed at once: there was both a different randomization method, and a different starer.

This is no mere quibble. It is already known that starers differ in their effectiveness. In a series of tests on the sense of being stared at using closed circuit television, Richard Wiseman and Marilyn Schlitz (1997), Schlitz obtained statistically significant positive results, in accordance with her previous findings. By contrast, when Wiseman, a sceptic, conducted the experiments and served as starer, the results were not significant.

Such experimenter effects are not symmetrical. The failure of the subjects to detect Wiseman's stares implies only that Wiseman was an ineffective starer, or that his negative expectations, influenced the subjects' responses, or both. By contrast, the detection of Schlitz's stares by the participants implies the existence of an unexplained sensitivity to stares.

Wiseman's negative expectations influenced the way he looked at the subjects. He himself said that he found it "an enormously boring experience" and that in most of the trials he was "pretty passive about it" (Watt, Wiseman & Schlitz, 2002). In addition, he may have influenced the expectations of the subjects at the University of Hertfordshire, where he works. At the time of this experiment, he was already well-known nationally through many TV and radio appearances, and in addition was often featured in the local Hertfordshire media. Most, if not all, of the subjects must have known of his reputation as a leading media sceptic, and been aware of his negative expectations.

Finally, Marks nor Colwell has shown any interest in testing their implicit learning hypothesis, only in publicising it. They could have examined their own data to see if subjects really did score more positively after particular patterns occurred in the sequences of looking and not-looking trials. Apparently they have not done so. I asked Colwell for his data so that I could test their hypothesis myself. He refused to let me see them. I then asked Marks but he has not replied to my repeated requests.

I pointed out the fundamental problems with Marks's argument in the Skeptical Inquirer (Sheldrake, 2001b). He and Colwell responded by saying that the change of starer was "a complete red herring", and resorted to rhetorical blustering instead of reasoned argument (Marks and Colwell, 2001).

Now, two years later, Marks is recycling his original claims unchanged, in his usual polemical style. This is lazy propaganda, not science. A paper as sloppy as this would be rejected by any self-respecting peer-reviewed journal. I assume Marks works to higher standards in his own professional field.

Return-anticipating dogs
Marks' account of research on return-anticipating dogs is equally misleading, and shows the same disregard for data that do not fit in with his beliefs.

Many pet owners claim that their dogs or cats know when a member of the family is returning and go to a door or window as if to wait for the person coming home. Sometimes the animals start waiting half an hour or more before the person arrives. These responses are said to occur even when people come home at unusual times, when no one at home knows when to expect them, and even when people travel in unfamiliar vehicles such as taxis (Sheldrake, 1999c).

I have carried out many trials with return-anticipating dogs, especially with Jaytee, a dog belonging to Pam Smart, in Ramsbottom, Lancashire. To start with, we recorded Jaytee's anticipatory behaviour on 100 occasions when Pam was absent for a wide range of times, some as short as an hour, others as long as long 12 hours. Jaytee anticipated her returns on 82 percent of these occasions, both with short and long absences (Sheldrake & Smart, 1998). He also anticipated her returns at least 10 minutes in advance when she was travelling in unfamiliar vehicles, such as taxis.

Subsequently, in a series of 100 videotaped trials, the place at which Jaytee waited by the window was filmed continuously on timecoded videotape throughout Pam's absences. These films were evaluated "blind" by independent scorers, who recorded all the times at which Jaytee was by the window. The data showed that he was waiting by the window very significantly more when Pam was on her way home from destinations at least 5 miles away than in the main period of her absence (Sheldrake & Smart, 2000a).

Marks suggested that the anticipatory behaviour of the Pam Smart's dog Jaytee could be explained by the dog learning when Pam could be expected home, and signalling accordingly. But if he had read our published papers he would have know that this hypothesis had already been refuted. Jaytee responded to Pam's homecomings after absences of very different durations. We tested for the possibility of learning effects by comparing Jaytee's behaviour after short, medium and long absences. His anticipatory behaviour was similar in all cases, ruling out the learning hypothesis (Sheldrake & Smart, 2000a, Figure 4).

Figure 4.
The time courses of Jaytee's visits to the window during P.S.'s long, medium and short absences. The horizontal axis shows the series of 10-minute periods (p1, p2, etc.). The vertical axis shows the average number of seconds that Jaytee spent at the window in each 10-minute period. Data for all 30 experiments are shown, as well as data for normal experiments after the exclusion of the seven noisy experiments. The last period shown on the graph represents the first 10 minutes of P.S.'s return journey (ret), the point for this is indicated by a filled circle of square. The bars show standard errors. When PS was returning in the short experiments in period 8, Jaytee was at the window a significantly higher proportion of the time than in period 8 of the medium and long duration experiments (by a factorial analysis of variance, p=0.004). Likewise, Jaytee spent a significantly higher proportion of the time at the window when Pam was on the way home in the medium experiments in period 11 than in period 11 of the long absences, when she would not be returning for more than another hour (p=0.003).

Marks commented, "Why Sheldrake chose to use a pre-arranged bleep period that started between 80 and 170 minutes after PS had left is unclear." The reason was explained in the Methods section of the paper that Marks criticized but apparently did not read. There was an upper limit to how long these videotaped trial could run, owing to the duration of the videotapes themselves, which lasted for a maximum of 4 hours, using the long play setting on the camera. Nevertheless, within this limit, there was a wide variation in the times of Pam's absences, from 80 to 220 minutes.

The data Marks referred to (examined for him by John Colwell) were taken from a subset of 12 videotaped trials. In these data, there was no evidence for a learning effect. Readers can judge for themselves by looking at the results for all 12 trials (Figure 2, Sheldrake & Smart, 2000a).

Figure 2.
The time courses form all 12 experiments in which P.S. came home at randomly selected times in response to being beeped. The ordinate shows the total number of seconds that Jaytee spent at the window in each 10-minute period, the abcissa the series of 10-minute periods defined in relation to the time at which P.S was beeped to come home. Data for all Jaytee's visits to the window, including irrelevant visits, are indicated by circles, and data from which irrelevant visits have been excluded are indicated by squares. The beep window is indicated by a line with two arrowheads, and this represents the period during which P.S could have received the signal to come home. Experiments with beeps in the first half of the beep window (early beeps) are on the left, and those with beeps in the second half of the beep window (late beeps) are on the right. The points for the 10-minute periods immediately following the beep during which P.S was returning are indicated by filled circles or squares. The graphs show the duration of all Jaytee's visits to the window in each 10-minute period, both with and without the exclusion of irrelevant visits. In one of these experiments, Jaytee did not go to the window at all, but in all the others he was at the window for the highest proportion of the time when PS was on her way home.

We also tested the learning hypothesis in another way, by running a series of control trials in which Pam did not come home at all. According to Marks' hypothesis, Jaytee should have gone to wait at the window after a routine time had elapsed. He did not (Figure 5, Sheldrake & Smart, 2000a). Marks simply ignored this evidence, or did not take the trouble to read it.

Figure 5.
Time sent by Jaytee by the window on evenings when PS was not coming home. The first of the 30-10 minute periods was from 6:30 top 6:40p.m., the last form 9:50 to 10:00p.m. The figures shown are averages from 10 evenings. The bars show standard errors. Observations on Jaytee at PS's sister's house.

Marks also based his case on a claim made by Richard Wiseman and his colleagues. At my invitation, Wiseman, Smith & Milton (1998) carried out four tests with Jaytee in 1995, using similar videotaped methods to my own.

In my own randomized tests, the dog was at the door an average of 4 per cent of the main period when his owner was absent, and 55 per cent during the first 10 minutes of her return journey (n = 12; p = 0.0001). In Wiseman et al.'s tests in the same location, which Marks thinks were better designed than ours, the respective figures were 4 per cent and 78 per cent (n = 3; p = 0.03) (Sheldrake, 1999b). Far from refuting the pattern of results that Smart and I observed, Wiseman et al. replicated it. But they wanted to debunk Jaytee's abilities. They ignored our data, and discarded most of their own to arrive at the conclusion they expected: Jaytee had failed their tests. They invented an arbitrary criterion by which to judge Jaytee. If he went briefly to the window before Pam set off for no apparent reason (as judged from the videotape), he had failed. These "failures" were part of the 4 per cent of the time Jaytee was at the window when Pam was absent. After these "failures", his waiting at the window when Pam was on the way home could be ignored, even though he was there for 78 per cent of the time.

Wiseman publicized this sceptical claim very widely through a press release, and in newspaper and TV interviews (Sheldrake, 1999b, 2000c; see also Wiseman, Smith & Milton, 2000). Marks repeated his claim uncritically, ignoring the actual data.

Jaytee is by no means unique. We have found similar patterns of return anticipation in other dogs. In a series of videotaped trials with a Rhodesian ridgeback, called Kane, the dog was at the window an average of 1 percent of the time during his owner's absence, and 26 percent of the time during the first 10 minutes of her homeward journey (n = 10, p = 0.0004; Sheldrake & Smart, 2000b).

Open-minded research
I am a biologist, rather than a parapsychologist. I am convinced there is much we do not understand about living organisms. That is why I believe it is important to investigate phenomena such as the sense of being stared at and apparent telepathy in animals: we could learn more about animal nature and human nature by doing so. If forms of the sixth sense really exist, they are likely to have evolved in relation to biological needs, and to be widespread in the animal kingdom. To accept their existence would not involve the abandonment of science and reason, and the collapse of civilization as we know it; rather it would extend the scope of science and of evolutionary understanding.

By contrast, Marks claims that "a normal or 'N'-theory interpretation (NIE) has proved to be a perfectly adequate explanation making any form of paranormal or 'P'-theory interpretation (PIE) redundant or superfluous". But it is not enough merely to suggest an "NIE interpretation": such hypotheses need testing, and the ones Marks proposed have been refuted by the data. He is himself a good example of the "powerful effect of belief and selective attention".

Like Marks, I am a sceptic, but of a different kind. His scepticism is directed towards anything he regards as "paranormal", taking as normal that which lies within the limits of current scientific understanding. My scepticism is directed towards the assumption that we know enough to proclaim what is possible and what is not.

References

Colwell, J, Schroeder, S, & Sladen, D. (2000) The ability to detect unseen staring: A literature review and empirical tests. British Journal of Psychology, 91, 71-85.
Delanoy. D. (2001) Anomalous psychophysiological responses to remote cognition: the DMILS studies. European Journal of Parapsychology, 16, 30-41.
Marks, D. (2000) The Psychology of the Psychic (2nd ed.) Amherst, NY: Prometheus Books.
Marks, D. & Colwell, J. (2000) The psychic staring effect: An artifact of pseudo randomization. Skeptical Inquirer September/October, 41-9.
Marks, D. & Colwell, J. (2001) Fooling and falling into the feeling of being stared at. Skeptical Inquirer, March/April, 62-63.
Sheldrake, R. (1998) The sense of being stared at: Experiments in schools. Journal of the Society of Psychical Research, 62, 311-323.
Sheldrake, R. (1999a) The 'sense of being stared at' confirmed by simple experiments. Biology Forum, 92, 53-76.
Sheldrake, R. (1999b) Commentary on a paper by Wiseman, Smith and Milton on the 'psychic pet' phenomenon. Journal of the Society for Psychical Research, 63, 306-311.
Sheldrake, R. (1999c) Dogs that Know When Their Owners Are Coming Home, And Other Unexplained Powers of Animals. London: Hutchinson.
Sheldrake, R. (2000a) The 'sense of being stared at' does not depend on known sensory clues. Biology Forum, 93, 209-224.
Sheldrake, R. (2000b) The 'psychic pet' phenomenon. Journal of the Society for Psychical Research, 64, 126-128.
Sheldrake, R. (2001a) Experiments on the sense of being stared at: The elimination of possible artefacts. Journal of the Society for Psychical Research, 65, 122-137.
Sheldrake, R. (2001b) Research on the sense of being stared at Skeptical Inquirer, March/April, 58-61.
Sheldrake, R. (2002) Seven Experiments That Could Change the World (second edition). Rochester, VT: Park Street Press.
Sheldrake, R. (2003) The Sense of Being Stared At, And Other Aspects of the Extended Mind. London: Hutchinson.
Sheldrake, R. & Smart, P. (1998) A dog that seems to know when his owner is returning: Preliminary investigations. Journal of the Society for Psychical Research, 62, 220-232.
Sheldrake, R. & Smart, P. (2000a). A dog that seems to know when his owner is coming home: Videotaped experiments and observations. Journal of Scientific Exploration , 14, 233-55.
Sheldrake, R. & Smart, P ( 2000b) Testing a return-anticipating dog, Kane. Anthrozoos, 13, 203-212.
Watt, C., Wiseman, R. & Schlitz, M. (2002) Tacit information in remote staring research: the Wiseman-Schlitz interviews. The Paranormal Review, 24, 18-25.
Wiseman, R. & Schlitz, M. (1997) Experimenter effects and the remote detection of staring. Journal of Parapsychology, 61, 197-207.
Wiseman, R., Smith, M. & Milton, J. (1998) Can animals detect when their owners are returning home? An experimental test of the 'psychic pet' phenomenon British Journal of Psychology, 89, 453-462.
Wiseman, R., Smith, M. & Milton, J. (2000) The 'psychic pet' phenomenon: A reply to Rupert Sheldrake. Journal of the Society for Psychical Research, 64, 46-49.

Rupert Sheldrake, Ph.D. is a biologist and author of Dogs That Know When Their Owners Are Coming Home, And Other Unexplained Powers of Animals. He was a Fellow of Clare College, Cambridge and a Research Fellow of the Royal Society. His web site is www.sheldrake.org