A PDF of the following article can be downloaded here.
How much is that doggy in the window?
A brief evaluation of the Jaytee experiments
Richard Wiseman
Background
In the early 1990s, Mathew Smith, Julie Milton, and I investigated
‘Jaytee’, a dog who could allegedly psychically predict
when his owner was returning home (Wiseman and Smith, 1996; Wiseman,
Smith & Milton, 1998). We believed that the results of our study
did not support the dog’s alleged ability. At roughly the same
time, Rupert Sheldrake (RS) conducted additional work with Jaytee,
and argued that his findings suggested that Jaytee was able to signal
when his owner started to return home from a distant location (see,
e.g., Sheldrake, 1999b; Sheldrake and Smart, 2000). RS also criticised
the study conducted by Mathew Smith, Julie Milton and myself (Sheldrake,
1999a, 1999b). We subsequently replied to the points raised in this
critique (Wiseman, Smith & Milton, 2000), and RS replied to our
reply (Sheldrake, 2000).
I am frequently asked about RS’s experiments with Jaytee. This
short note briefly outlines some of my thoughts.
RS carried out two main experiments with Jaytee. Each experiment employed
the same general procedure, involved both Jaytee and his owner (PS),
and took place at PS’s parents house. Each experiment was composed
of several trials. During each trial:
- An unmanned static camera was set up to continuously film a porch
area that Jaytee visited to allegedly signal PS’s return.
- PS then left the house, and the camera recorded Jaytee’s behaviour
whenever he visited the porch.
- The resulting videotapes were later analysed by measuring the amount
of time that Jaytee was present in the porch during three periods:
The first ten minutes after PS set off to return to the house (the
‘return period’).
The ten minutes prior to the return period (the ‘pre-return
period’).
The time when PS was out prior to the ‘pre-return period’
(the ‘main period’).
When averaged across the trials, the data from both experiments showed
that Jaytee was at the porch a relatively small percentage of time
during the ‘main period’, longer during the ‘pre-return
period’, and longest during the ‘return period’.
One experiment consisted of 30 trials involving ‘ordinary’
homecomings (in which PS’s return time was not randomly determined),
and the other consisted of 12 trials involving ‘random’
homecomings (in which her return time was randomly determined).
When evaluating any empirical research, it is important to be able
to assess the degree to which the results might be skewed by data
selection, post hoc analyses, and other possible problems. The data
from the Jaytee experiments has been described in both a popular book
(Sheldrake, 1999b) and academic journal (Sheldrake and Smart, 2000).
Both sources present graphs showing the amount of time that Jaytee
visited the window throughout the 12 ‘random’ homecoming
trials. In both the book and the paper, 6 of the 12 trials are classified
as ‘early’ returns (in which PS returned relatively early
in the trial) and 6 as ‘late’ returns (in which she returned
relatively late). There are some important discrepancies between the
book and paper. For example, in the paper, the trial which took place
on the 11/2/97 is classified as an ‘early’ trial, whilst
in the book it is classified as ‘late’. In the paper,
the trial on the 1/7/97 is classified as ‘late’, whereas
in the book it is classified as ‘early’. In addition,
some of the data patterns appear different in the two sources. In
the paper, the data from the trial on the 19/3/97 shows Jaytee spending
very little time at the porch in the early part of the trial, whereas
in the book he spends a considerable amount of time there. Likewise,
in the paper, the trial on the 21/9/97 shows a spike in Jaytee’s
activity that appears to be missing from the corresponding graph in
the book.
However, assuming that the data is sound, there are two main normal
explanations that could potentially account for Jaytee appearing to
psychically know when his owner is returning home. I have labelled
these the ‘sensory leakage’ hypothesis, and the ‘anxiety’
hypothesis. The following two sections each discuss whether such explanations
might account for the results of the two experiments.
‘Ordinary’ homecomings
The ‘sensory leakage’ hypothesis. If
Jaytee knew roughly when PS was due to return, his alleged psychic
behaviour could simply reflect genuine anticipation. During the 30
trials, PS’s return time was not randomly determined, and so
Jaytee might have gleaned a rough return time from several sources,
including, for example, PS’s normal routine, interactions with
Jaytee prior to leaving the house, the behaviour of PS’s parents,
the way PS was dressed, etc.. RS has acknowledged that this is a potential
problem, but argues that it was excluded the in the ‘random’
homecomings studies discussed below.
The ‘anxiety’ hypothesis. If Jaytee became
more and more anxious over time, he might visit the porch more frequently,
thus producing the observed pattern. RS assessed this notion by examining
the trials during which PS returned at different times. His argument
can be illustrated as follows. Imagine taking all of the trials in
which PS set off to return home after half an hour (‘early’
trials), and comparing them with the trials in which she set off to
return home after an hour (‘late’ trials). According to
the ‘anxiety’ hypothesis, Jaytee would be expected to
spend roughly the same amount of time at the porch 25 minutes into
both the ‘early’ and ‘late’ trials. However,
if Jaytee possessed psychic ability, he would be expected to spend
a larger amount of time at the porch 25 minutes into the ‘early’
than ‘late’ trials. Using this approach, RS compared three
sets of trials (those that were 80-100 minutes long, those lasting
between 110-170 minutes, and those 180 minutes or longer), and argued
that the resulting patterning did not support the ‘anxiety’
hypothesis. However, it is important to realise that although this
hypothesis might not account for the data in this study, there is
no reason to believe that Jaytee would never engage in such behaviour.
As noted above, in these trials, Jaytee may have been aware of when
PS was likely to return