As the onlookers will see, noone here actually read the papers in question.
Here they are:
A Dog That Seems to Know When His Owner Is Coming Home: Videotaped Experiments and Observations
: Rupert Sheldrake & Pamela Smart, Journal of Scientific Exploration, Vol. 14, No. 2, pp. 233–255
CAN ANIMALS DETECT WHEN THEIR OWNERS ARE RETURNING HOME?
AN EXPERIMENTAL TEST OF THE ‘PSYCHIC PET’ PHENOMENON
: Richard Wiseman, Matthew Smith, Julie Milton, British Journal of Psychology Vol. 89, 1998, p.453-462
The claim is that Jaytee, the terrier of Pam Smart, is able to detect when owner Pamela Smart is on her return journey by waiting in his porch.
- "It happens on a schedule": Both experimenters started on an random timepoint unknown to the ower
Pam Smart. Ruled out.
- "Dogs have excellent hearing": Both experimentersstarted the journey at least 13 minutes apart and
used different, non-identifiable means of transportation (bike, taxi, experimenter car)
- "Confirmation bias": Both experimenters videotaped the dog and tabulated the occurences of the porch
with an observer who did not know the return time.
- "Separation anxiety": Both authors took precautions to record how Jaytee behaves and
if it is normal. Was taken care of and is therefore irrelevant.
What Sheldrake did:
Sheldrake recorded ca. 100 experiments, 12 under controlled conditions above, 10 control experiments and the rest under more natural conditions. Before the videotaped sessions, they analyzed written data recorded by Pam Smart over longer timeperiods.
Sheldrake tabulated Jaytees position and used two statistical methods. Sheldrake's main method divided the data in three categories, return period (Pam is on the way home), prereturn period (10 minutes before departure) and main
period (all else). Jaytee was according to Sheldrake 55% the time at the window at the return period, 23% at the prereturn period and 4% average for the same 10 minute period at the main period. (The main period is much longer than return/prereturn so the missing 22% indicates that the main period was in average nearly 5 times longer ~ 50 min). Sheldrake used the repeated-measures analysis of variance (ANOVA) and a paired-sample t-test to test if the differences were significant and they were according to Sheldrake highly significant (p < 0.0001, p. 239). The increased percentage of time was statistically significant for the prereturn period (p = 0.04) and highly significant for the difference between prereturn and return (p = 0.0009).
What Wiseman did:
Wiseman did 4 experiments. They divided the experiment time into 10 minute time blocks, Jaytee is only counted as success if
a) Jaytee goes to his porch only 10 minutes after the return signal.
b) Jaytee goes to his porch the first time for no apparent reason (must be decided by a judge).
c) In three cases, Jaytee must remain for at least 2 minutes.
- Leaving time: 21:00. Jaytee remains 53 seconds on 19:57, 134 seconds on 20:09 and
over 600 seconds between 20:58 and 21:04 because a car pulls up. Fail.
c) is introduced.
- Leaving time: 14:18.Jaytee remains 140 seconds on 13:59, and 169 seconds because
the fish delivery van comes up. Fail.
- Leaving time: 21:39. Jaytee remains 1233 seconds after 21:31, but as this is too soon,
it counts as fail.
- Leaving time: 10:45. Jaytee goes at 10:44, but as the terrier needs to vomit, he must
leave the porch. Fail.
I leave it to the reader to decide it for themselves because I think it is such an extremely charged issue.