Jaytee: The Dog Who Knew When His Owner Was Coming Home

Pam Smart adopted Jaytee from a rescue centre in 1989. Her parents noticed that the dog would move to the bay window and sit there shortly before Pam returned — even when she came home at irregular hours, by different transport, and without telephoning ahead. In 1994 the biologist Rupert Sheldrake began filming the window continuously. What followed is one of the most heavily documented, and most bitterly disputed, datasets in parapsychology.

Why this case is unusual

Most animal-anticipation claims die at the anecdote stage. Jaytee did not, for one reason: there is continuous videotape, time-coded, of the window area during the owner's entire absence. That converts a story into a measurable signal.

Between May 1994 and February 1995 Pam Smart's parents logged Jaytee's behaviour on 100 occasions. Sheldrake then ran 30 videotaped experiments where Pam's return time was chosen at random by the experimenter and communicated to her by pager, after she had already left the house.

The measure

The dependent variable was the proportion of each period that Jaytee spent at the window, excluding periods explainable by ordinary causes (a passing cat, a delivery van). Sheldrake divided each absence into:

1. the main period (Pam away, no return underway)
2. the pre-return period (roughly the ten minutes before Pam set off)
3. the return journey itself

In the randomised-return trials, Jaytee spent an average of 4 percent of the main period at the window, 55 percent of the pre-return period, and 23 percent of the return journey.

The pattern was consistent across seasons, distances from three to fifty kilometres, and different vehicles.

The counter-analysis

Richard Wiseman, Matthew Smith and Julie Milton ran four of their own trials in 1995 and reported a negative result. Their criterion was different: they scored the first signal — the first time Jaytee went to the window for more than two minutes — and found it occurred early in the absence, long before Pam set off.

Sheldrake replied that Jaytee visited the window frequently and briefly for ordinary reasons, so a "first signal" rule guarantees an early hit and discards the shape of the whole curve. When Wiseman's own four tapes were re-plotted using the proportion-of-time measure, they produced the same rising curve as Sheldrake's thirty.

::: callout Both teams agree on the raw footage. The dispute is entirely about which statistic is the legitimate measure of anticipation — a methodological argument, not an argument about whether the dog moved. :::

Prosaic explanations that have been tested

• Routine. Ruled out by randomised, experimenter-chosen return times.
• Hearing the car. Jaytee reacted while Pam was still kilometres away and travelling in unfamiliar vehicles.
• Smell. No plausible mechanism at the distances and wind conditions recorded.
• Cueing by the people at home. Pam's parents did not know the return time; they were also filmed.
• Selective reporting. The continuous tape means non-events are on the record too.

What remains open

No one has replicated the effect at scale with other animals under the same filming protocol. A single dog, however well documented, is a single dog.

The Jaytee tapes are a rare thing: a paranormal claim that survived contact with a control condition, and then stalled — not because it was debunked, but because the field never funded the replication that would settle it. Twenty-five years later the correct verdict is still "unresolved and under-tested", which is a far more uncomfortable answer than either side wants.