The claim that dogs wait by the window before their owners set off for home, at randomly chosen times, in unfamiliar vehicles. Parapsychology's most beloved experiment, and its most carefully filmed dog.
Documented
One hundred videotaped trials
Rupert Sheldrake and Pamela Smart ran the protocol on her dog Jaytee: an unmanned camera on the porch he used as a waiting post, the owner out of the house, her return time chosen at random and signalled by pager, her vehicle an unfamiliar taxi. Across the trials, Jaytee's porch time stayed low all afternoon and climbed sharply in the minutes after she set off for home. The work ran in the Journal of Scientific Exploration and Anthrozoos in 2000, and Sheldrake's owner surveys in England and California found about half of dog owners reporting the same anticipation.
Assessment: The trials, the pager randomization and the published graphs are documented. The honest caveats are also documented: the decisive comparisons rest on small numbers of trials, and the owner knew the window of time in which she could be paged.
Stories
Every owner already knew
Long before the cameras, the phenomenon was the world's most common paranormal anecdote: the dog at the door before the engine sound, the cat at the window before the footsteps. Sheldrake extended the frame to other species, most famously N'kisi, an African grey parrot whose owner, Aimee Morgana, tested whether his running commentary matched photographs she was viewing in another room. The parrot scored above chance, and the case entered the literature in 2004.
Assessment: The anecdotal base is enormous, worthless as proof, and enormous again as the reason the research exists at all. N'kisi's result is a documented one-off with a sample size of one parrot.
Interesting claims
The matching-graphs dispute
Richard Wiseman, Matthew Smith and Julie Milton tested Jaytee four times and announced failure in the British Journal of Psychology in 1998. Sheldrake read their paper and pointed out that their own graphs show Jaytee at the window most when his owner was returning; the quarrel, run through the Journal of the Society for Psychical Research in 1999 and 2000, was over criteria, first-signal versus proportion-of-time. Wiseman's later summary offers two normal explanations: sensory leakage from the household, and an anxious dog visiting the window more as the afternoon wears on.
Assessment: Documented as a live dispute about scoring, not about film. The most telling fact in the file is the empty one: no independent laboratory has run the full protocol since, so the dispute remains the data.
Sources and further reading
Sheldrake and Smart, A dog that seems to know when his owner is coming home, Journal of Scientific Exploration 14 (2000), 233-255
Sheldrake and Smart, Testing a return-anticipating dog, Anthrozoos 13 (2000), 203-212
Wiseman, Smith and Milton, Can animals detect when their owners are returning home?, British Journal of Psychology 89 (1998), 453-462
Sheldrake, commentary, JSPR 63 (1999), 306-311, and Wiseman, Smith and Milton, reply, JSPR 64 (2000), 46-50
Wiseman, How much is that doggy in the window, the skeptic's own account of the matching graphs
Morgana and Sheldrake, Testing a language-using parrot for telepathy, JSPR 68 (2004)
Sheldrake, Dogs That Know When Their Owners Are Coming Home (1999), the primary text
Worth asking
If both scientists agree on the footage and disagree on the conclusion, what is actually being measured: the dog, or the definition of a signal?