Journal of the Society for Psychical Research 61, 1997
by Rupert Sheldrake and Pamela Smart

Introduction

Pet-owners often say that their dogs and cats are telepathic or have a "sixth sense". However, apart from a few pioneering discussions of anecdotal evidence (e.g. Rhine, 1951, Rhine and Feather,1962; Pratt, 1964). the seemingly psychic powers of pets have been more or less ignored by psychical researchers. They have also been neglected by researchers into animal behaviour, and indeed by all other scientific disciplines.

In a book entitled Seven Experiments that Could Change the World (Sheldrake, 1994) one of us has proposed that there is are great opportunities for simple, inexpensive and repeatable research with domestic animals, in particular in relation to the apparent ability of some pets to know in advance when their owner is coming home.

As part of this research programme, we have investigated the incidence of uncanny behaviour by pets. This survey was carried out by telephone in Greater Manchester, North-West England, and involved 394 households selected at random.

Methods

Data were collected by means of telephone interviews conducted by Pamela Smart (P.S.), following a standard questionnaire format.

The households to be surveyed were in and around Ramsbottom, a market town, population 13,500, near Bury, Greater Manchester, in which P.S. lives. After some preliminary tests, we chose her home town for this survey because we found that when she identified herself at the beginning of her telephone calls as a Ramsbottom resident, respondents tended to be more friendly and more prepared to take time answering questions than people in other places. We also chose her local area for this survey in order to minimize the cost of the telephone calls.

Households were selected from the North-West Manchester Residential Telephone Directory at random. Each household was chosen using two random numbers. The first, a three-digit number from a table of random numbers, was used to select the page of the directory; the second, obtained by throwing a die, was used to select one of the three columns on that page. The first Ramsbottom number in that column was used. (If there was no Ramsbottom number in that column, then the first one in the subsequent column was taken instead.)

P.S. introduced herself as follows:"My name is Pam Smart from Ramsbottom, and I'm doing a local survey on pets and would like to ask you a few questions". Approximately 25% of the people she spoke to were too busy or unwilling to do so. The majority were willing to take part, and she then asked a series of questions and recorded the answers on a standard form, as follows.

Name: ______________________________
Tel: ___________________________
Address: ___________________________________________________________

1. Do you or anyone in your household own a pet?

Yes

No

2. What type of pet?            Species

Age

Sex

3. Have you/they ever noticed the pet getting agitated before a family member has arrived home?

Age

Sex

4. How long before you/they arrive is your pet agitated?
0-5 mins	5-10 mins	10-20 mins	20 mins or more

5. Would you agree or disagree that your pet knows you are going out before you show any physical signs of doing so?
	Agree	Disagree	Don't know

6. Would you agree or disagree that you pet responds to your own thoughts or silent commands?
	Agree	Disagree	Don't know

7. Would you agree or disagree that your pet is sometimes telepathic with you?
	Agree	Disagree	Don't know

8. Would you agree or disagree that any of the pets you have known in the past were telepathic?
	Agree	Disagree	Don't know

9. How frequently have you yourself had a psychical experience?
	Never	Sometimes	Frequently

In cases where respondents had no pets at present, in other words when they answered NO to the first question, they were asked only one more question, Question 8.

Statistical analyses were carried out using 2 x 2 contingency tables and the chi-squared test (Campbell, 1989). Probability values for two-tailed tests were used.

Results

Pet ownership
Out of 394 households surveyed, 202 (51.3%) had pets. This figure is close to the UK national average of 50% (Pet Food Manufacturer's Association, 1995). Dogs were the most common pet, followed by cats. The figures were as follows:

Dogs	122	Gerbils	2
Cats	93	Guinea pigs	2
Rabbits	17	Mice	1
Budgerigars	8	Horses	1
Other birds	8	Pigs	1
Fish	8	Stick insects	1
Hamsters	4

Most of these households had only one kind of pet: 81 had dogs only, and 54 cats only. Thirty-three had both dogs and cats; 8 had dogs and other pets (excluding cats); 6, cats and other pets; and 6, dogs, cats and other pets.

The percentages of households with dogs and cats in the Ramsbottom survey were higher than the national averages: 30.9% had dogs, compared with 23.0% nationally; and 23.6% had cats, compared with 20.8% nationally (Pet Food Manufacturer's Association, 1995). These differences probably reflect the fact that Ramsbottom has a higher proportion of detached, semi-detached and terraced houses than are found in large cities. People living in such houses tend to have more dogs than those living in flats. The higher ratio of dogs to cats in Ramsbottom than in the UK as a whole is also consistent with regional trends. The lowest ratio of dogs to cats is found in southern England (Pet Food Manufacturer's Association, 1995).

Not all respondents replied to the question about the gender of their pets. Where the gender was given, there were 44 male and 48 female dogs, and 10 male and 18 female cats.

Dogs and cats that seem to know when a family member is coming home
Apart from the single exception of a cockatiel, the only kinds of animals that seemed to know when a member of the family was coming home were dogs and cats. Data for dogs and cats on replies to the Questions 3 and 4 are shown in Table 1.

A far higher proportion of dogs than cats seem to anticipate arrivals. In this survey the figures were 46% and 14% respectively. The difference in the reactions of dogs and cats is very significant statistically (p=1 x10-6). In the 33 households with both dogs and cats, only one cat was reported to respond, compared with 17 dogs.

Many pets became agitated less than 5 minutes before a member of the household arrived. In this short period, it would be hard to distinguish between different possible explanations for their agitation, including the possibility that the animals heard or smelt the person approaching. However, a substantial proportion of the animals that reacted to a person's arrival did so more than 10 minutes in advance:16 per cent of the dogs and 23 per cent of the cats. In such cases it should be possible to explore by means of experiments the possible reasons for their reaction, as discussed below.

Table 1. Replies to Questions 3 and 4

Have you/other members of your household ever noticed a pet getting agitated before a family member has arrived home?
Numbers (and percentages) of dogs or cats
	Total	Yes	No	Don't know
DOGS	122	56 (46%)	60 (49%)	6 (5%)
CATS	93	13 (14%)	78 (84%)	2 (2%)

How long before you/they arrive is your pet agitated?
Numbers (and percentages) of dogs or cats reacting
	Total	5-10 min	10-20 min	20+ min	Don'tknow
DOGS	56	9 (16%)	4 (7%)	5 (9%)	10 (18%)
CATS	13	0	1 (8%)	2 (15%)	5 (38%)

The responses of pets to thoughts and intentions.
The ability of pets such as cats and dogs to respond to their owners' thoughts is taken for granted in popular books on animal training (e.g. Woodhouse, 1980). In preliminary informal surveys, we found that many pet owners claimed that their animals seemed to know when they were going to go away or go out, even before they had shown any obvious signs of doing so. Many also claimed that their pets seemed to be able to read their minds, responding to their thoughts or intentions. Some assumed that this communication depended on telepathy. We asked questions to find out how common these impressions are (Table 2).

The results show that a higher proportion of dogs than cats appear to respond to their owners' intentions and thoughts. A majority of dog owners thought that their dogs knew in advance when they were going to go out, and that their dogs responded to their thoughts and were sometimes telepathic with them. Only a minority of cat owners had these impressions about their cats. These differences between dogs and cats were significant statistically (for Question 5, p = 0.005 ; Question 6, p = 0.01; Question 7, p = 0.02).

Table 2. Replies to Questions 5, 6 and 7.

5. Would you agree or disagree that your pet knows you are going out before you show any physical signs of doing so?
Numbers (and percentages) of dogs or cats
	Total	Agree	Disagree	Don't know
DOGS	122	84 (69%)	29 (24%)	9 (7%)
CATS	63	30 (48%)	28 (44%)	5 (8%)
6. Would you agree or disagree that you pet responds to your own thoughts or silent commands?
Numbers (and percentages) of dogs or cats
	Total	Agree	Disagree	Don't know
DOGS	122	65 (53%)	40 (33%)	17 (14%)
CATS	63	21 (33%)	29 (46%)	13 (21%)
7. Would you agree or disagree that your pet is sometimes telepathic with you?
Numbers (and percentages) of dogs or cats
	Total	Agree	Disagree	Don't know
DOGS	122	66 (54%)	38 (31%)	18 (15%)
CATS	63	23 (37%)	29 (46%)	11 (17%)

This minority of cat owners is probably even smaller than the data in Table 2 indicate, and the statistical significance of the diferences between dogs and cats even greater. The data in Table 2 refer only to cases where people specifically referred to their cats. Of the 33 respondents with both dogs and cats, only 3 commented on the cats' behaviour; the other 30 talked only about their dogs, implying that they had not noticed their cats responding to their thoughts and intentions. If these 30 cases are included, the total number of cats rises to 93, and the proportion that appear to know when their owners are going out consequently falls from 48% to 32%. Likewise, the percentage responding to thoughts or silent commands falls from 33% to 23%, and those that seem to be telepathic from 37% to 25%. The statistical significance of the differences between dogs and cats then becomes very high indeed (for Question 5, p = 1 x 10-7 ; Question 6, p = 2 x 10 -6; Question 7, p = 1 x 10-5).

The similarity of the responses to questions 6 and 7 indicates that most respondents considered that their pets' responses to their thoughts or silent commands were telepathic, at least sometimes.

Telepathic connections with pets in the past
Both pet-owners and people who have no pets at present were asked about telepathy in pets they knew in the past (Table 3). Ninety-two per cent of current pet owners and 85 per cent of those without pets at present said they had kept or known pets in the past. In both groups just over 50 per cent agreed that pets they had known were telepathic.

Table 3 Answers to Question 8: Would you agree or disagree that any of the pets you have known in the past were telepathic?

	Numbers (and percentages) of people
	Total	Agree	Disagree	Don't know
Present pet owners	186	99 (53%)	62 (33%)	25 (13%)
No pets at present	163	90 (55%)	65 (40%)	8 (5%)

Pet owners' own psychic experience
Not all the pet owners answered Question 9, about their own psychic experience, but most did so (178 out of 202).

A majority (54%) of these respondents said that they had themselves had psychical experiences (Table 4). This figure is in general agreement with the results of other recent surveys. For example, in a survey of young people in South-East England, Gaynard (1992) found that 54.4% claimed personal experience of at least one paranormal event. In a review of various national surveys, Haraldsson (1985) gave figures of 64% for both Britain and Iceland and 60% for the USA. In a survey of American adults, Gallup and Newport (1991) found that 75% claimed experience of at least one paranormal occurrence and 50% claimed experience of more than three.

Table 4. Answers to Question 9: How frequently have you yourself had a psychical experience?

	Numbers (and percentages) of people
	Total	Frequently	Sometimes	Never
All pet owners	178	22 (12%)	74 (42%)	82 (46%)
Dog owners	109	14 (13%)	51 (47%)	44 (40%)
Cat owners	52	6 (11%)	16 (31%)	30 (58%)

In our survey, more dog owners than cat owners claimed to have had psychic experiences themselves, 60% as opposed to 42%. This difference is statistically significant (p= 0.04).

We compared the answers about their pets given by pet owners who had had psychic experiences frequently or sometimes ("psychic owners") with the answers given by owners who said they had never had psychic experiences ("non-psychic owners"). The data in Table 5A show that for all questions, and for both dogs and cats, psychic owners reported more positive responses than non-psychic owners. These differences were not statistically significant when dogs and cats were considered separately, except in the case of Questions 6 and 7 for dogs and 7 for cats. But when the data for dogs and cats were combined (Table 5B), psychic owners gave significantly more positive answers than non-psychic owners to all questions except Question 5.

Discussion

How reliable are people's reports of their pets' behaviour?
With any questionnaire there is the problem of knowing how reliable people's answers are. In the present case, one possible source of bias might be a tendency for some pet owners to exaggerate the powers of their pets, owing to their fondness for them. Another source of bias might have the opposite effect: some people pay relatively little attention to their animals and do not observe them closely enough to notice many of their reponses.

People who claimed to have had psychic experiences themselves were more likely to report that their animals were telepathic with them (Table 5). This could mean that they tended to impute psychic powers to their pets in conformity with their own experience and beliefs. Or it could mean that they were more observant about these aspects of their animals' behaviour than non-psychic pet owners. And just as belief in psychic powers could bias people to imagine abilities that that their pets do not have, so scepticism about psychic powers could bias people to deny abilities that their animals in fact possess.

It is impossible to know how much these and other forms of bias may have influenced our data. But what the data clearly reveal is that many pet owners claim their animals show psychic powers.

Table 5. Relation of owners' psychic experience to the observed behaviour of their dogs or cats.

A. Number (and percentage) of psychic and non-psychic owners who answered yes to Question 3 and agreed with Questions 5,6,7 and 8.
	Dogs		Cats
	Psychic	Non-psychic	Psychic	Non-psychic
Totals	65	44	22	30
Question
3.	33 (51%)	16 (36%)	6 (27%)	5 (17%)
5.	46 (71%)	32 (73%)	11 (50%)	11 (37%)
6.	42 (65%)	18 (41%)	10 (45%)	7 (23%)
7.	42 (65%)	18 (41%)	13 (59%)	5 (17%)
8.	39 (60%)	20 (45%)	16 (73%)	15 (50%)
B. Number (and percentage) of psychic and non-psychic owners of dogs and of cats who answered yes to Question 3 and agreed with Questions 5,6,7 and 8, together with the statistical significances of the differences in responses between psychic and non-psychic owners.
	Dogs + Cats
	Psychic	Non-psychic	Significance(p)
Totals	87	74
Question
3.	39 (45%)	21 (28%)	0.03
5.	57 (66%)	43 (58%)	NS
6.	52 (60%)	25 (34%)	0.001
7.	55 (63%)	23 (31%)	0.0005
8.	55 (63%)	35 (47%)	0.04

How typical is Ramsbottom?
All these pets owners were in and around Ramsbottom. Their responses may not have been typical of Britain as a whole if there are strong local and regional variations in beliefs in the psychic powers of animals; or if the inhabitants of small towns are very different from city dwellers or country dwellers; or if there is any special feature of Ramsbottom that could bias the results. We have no reason to suppose that there are major regional differences in belief in a "sixth sense" of animals, nor that inhabitants of towns are more or less prone to this belief than those who live in cities or villages. And Ramsbottom is an ordinary kind of town, showing a typical pattern of pet ownership.

From the point of view of this survey, Ramsbottom is exceptional only in that Pamela Smart lives there. Apart from saving money on telephone calls, we chose to carry out this survey in her home town because we thought people might be more likely to take the time to answer our questions if they were talking to a local person, rather than a complete stranger.

Pamela Smart has been the subject of several articles in local newspapers about research with her dog Jaytee, who seems to know when she is coming home. But she and Jaytee have also been featured in national newspapers and TV programmes, so any possible bias introduced by her identifying herself on the telephone may not be specific to Ramsbottom, but could have influenced respondents in other parts of Britain too. Perhaps talking to Pamela Smart did indeed make some people more willing to take part in the survey; perhaps some people may have been more likely to claim their pets were psychic when talking to her; perhaps some felt more free to speak frankly to her than to someone they knew nothing about.

Only further surveys carried out in different places by other people will enable us to estimate how important these possible sources of bias might have been.

The prevalence of seemingly psychic pets
All reports of telepathic responses by pets were confined to cats and dogs. None of the other animals, such as rabbits, were said to have shown such behaviour.

This survey shows that that over half the dog owners and over a third of the cat owners believe they have a psychic bond with their pets, at least sometimes. Moreover, a majority who had had pets in the past thought some of them were telepathic (Table 3). Even among the 192 people currently without pets, 163 had kept or known pets in the past, and 90 had found them to be telepathic. In other words, 90 out of 192 or 47% of the non-pet-owners had known telepathic pets in the past. Assuming that the population of Ramsbottom is fairly typical of the nation as a whole, we can conclude that roughly half the adult population have had personal experience of seemingly telepathic animals.

Again assuming that the present survey is roughly representative of the country as a whole, the approximate numbers of dogs and cats in the UK that seemingly respond telepathically to their owners can be calculated. There are dogs in about 5.3 million households in the UK, and cats in about 4.7 million (Pet Food Manufacturer's Association, 1995). Taking the percentage of telepathic dogs as 54 % and cats as 25 %, this gives 2.9 million telepathic dogs and 1.2 million telepathic cats in the UK. Even though these figures are very approximate, they serve to show that there are many households with the potential for research on the seemingly psychic behaviour of pets.

In a similar way, the numbers of dogs and cats that anticipate their owner's return by 10 minutes or more can be calculated on the basis of the percentages shown in Table 1. The calculated number of dogs is about 400,000, and of cats about 150,000. There is thus a large pool of households in the UK in which this anticipatory behaviour could potentially be investigated experimentally.

Comparison of cats and dogs
The smaller proportion of cats than dogs to respond to their owners' returns, departures, thoughts and intentions could mean that they are less sensitive than dogs. But it could also mean that they are less interested, less motivated or less connected with their owners; cats tend to be less sociable and more independent than dogs. More dog owners have close relationships with their animals than cat owners, and tend to be more closely attached to them (Hart, 1995). Moreover, kittens have a shorter period of sensitivity to socialization than puppies, and unless they are handled frequently before they are weaned, they are less likely to form strong attachments to human beings (Karsch and Turner, 1988).

There may also be differences between the kinds of people that choose cats or dogs as pets. Our survey revealed one such difference (Table 4) in that a significantly higher proportion of dog owners than cat owners said they had had psychic experiences.

Both dog and cat owners who had had psychic experiences were more likely to claim that their pets were telepathic (Table 5). The lower proportion of people with psychic experience among cat owners could have led to a tendency for cat owners not to notice or report such behaviour in their animals. However, this possible source of bias cannot account for more than a small part of the reported differences between dogs and cats, because the differences still showed up when the responses of non-psychic pet owners are compared. In reply to Question 7 about their pet being telepathic, 41% of non-psychic dog owners agreed, compared with 17% of non-psychic cat owners (Table 5).

The need for experimental investigations of the seemingly psychic powers of pets
Most of the phenomena we asked about in this questionnaire do not necessarily imply the existence of psychic powers, and some could perhaps be explained in terms of the sharp senses of dogs and cats together with subtle clues of which their owners were unaware. To test these possibilties, experimental investigations are necessary.

In the case of pets that know when a member of the household is returning, possible explanantions fall into five main categories:

1. Pet owners make exaggerated or untrue claims about their pets' abilities. Pets do not really anticipate when their owners are coming home; this is an illusion in the minds of dog and cat lovers.
2. The pets are responding to a routine on the part of the returning person, and simply respond at the same time each day. They do so either on the basis of an internal clock, or as a result of clues from the environment such as the signature tune of a daily TV show.
3. The pets are responding to cues from people at home, whose behaviour or emotional state changes when they know a member of the family is about to arrive.
4. The pets hear, see or smell the person coming; for example they may be responding to the sound of a familiar car engine, or to footsteps in the street.
5. The pets are responding on the basis of abilities currently unknown to science, such as a "sixth sense" or psychic or telepathic powers.

In this survey, half the number of dogs and about a third of the cats that responded did so 5 minutes or less before a person arrived home. With such short reaction times, sounds and smells may well play a part in the animals' responses. However, when pets respond 10 minutes or more before a person's return, sounds and smells are less plausible explanations. A person may be many miles away when the reaction occurs, and may be travelling by car, bus or train.

With pets that respond 10 minutes or more in advance, simple experiments can help decide between the possible explanations (Sheldrake, 1994). The behaviour of the animal should, if possible, be recorded on video, with the camera set up to film the place where the pet usually waits for the returning person. To test explanations 2 and 3 the person should come home at an unusual time, randomly selected, and the people at home should not know when they are returning. To test explanation 4, they should return by an unusual method, for example by being driven home by a friend in an unfamiliar car, or in a taxi, or on a borrowed bicycle. Any of these experiments would at the same time test explanation 1, the illusion theory, by revealing whether pets really do show objectively observable anticipatory behaviour.

We have been carrying out such experiments ourselves. In ongoing studies we are assessing the possible role of psychic links between pets and owners when the owners are still miles away from home.

When pet owners are in their homes, the investigation of possible telepathic influences requires the elimination of sensory cues that might reveal the owners' thoughts and intentions, including cues of which the owners may be unconscious. The best way to do this is to film the pet in one room while the owner is behind closed doors in another room, or preferably in another building, and to see if the pets respond when the owner forms an intention at a randomly selected time. For example, a person may form the intention to take their dog for a walk. If normal sensory cues can be eliminated, and the dog still gets excited, then this would imply the existence of an unexplained link between the owner and the dog.

We are convinced that there is a great potential for these kinds of investigations. Not only do many dogs and cats seem to have better-developed psychic powers that most people, but they provide good opportunities for repeatable research. Human subjects often get bored by doing repetitive tests for parapsychological experiments, but fortunately dogs do not get bored with being taken for walks, nor cease to be excited by their owners' coming home. The results of this survey show that pets with seemingly psychic powers are common.

Acknowledgments

We gratefully acknowledge a research grant from the Lifebridge Foundation, New York. We would like to thank all the people who took part in this survey, and Dr Tom Merriom and Dr Richard Wiseman for helpful advice and discussion.

References

Campbell, R.C. (1989) Statistics for Biologists . Cambridge: Cambridge University Press.

Gaynard, T.J. (1992) Young people and the paranormal. JSPR 58, 165-180.

Gallup, G.H. and Newport, F. (1991) Belief in paranormal phenomena among American adults. Skeptical Inquirer 15, 137-146.

Haraldsson, E. (1985) Representative national surveys of psychic phenomena. JSPR 53, 137-146.

Hart, L.A. (1995) Dogs as human companions: a review of the relationship. In: The Domestic Dog: Its Evolution, Behaviour and Interactions with People (ed. Serpell, J.). Cambridge: Cambridge University Press.

Karsh, E.B. and Turner, D.C. (1988) The human-cat relationship. In: The Domestic Cat: The Biology of its Behaviour (ed. Turner, D.C. and Bateson, P.). Cambridge: Cambridge University Press.

Pet Food Manufacturer's Association (1995) PFMA Profile 1995 . London: Pet Food Manufacturer's Association.

Pratt, J.G. (1964) Parapsychology: An Insider's View of ESP , Chapter 8. London: W.H.Allen.

Rhine, J.B. (1951) The present outlook on the question of psi in animals. JP 15, 230-251.

Rhine, J.B. and Feather, S.R. (1962) The study of cases of "psi-trailing" in animals. JP 26, 1-22.

Sheldrake, R. (1994) Seven Experiments that Could Change the World . London: Fourth Estate.

Woodhouse, B. (1980) Talking to Animals . London: Allen Lane.