Home page Back
bobnotes Sheldrake-Shermer Dialogs SHELDRAKE-Shermer 2nd The Jaytee Issues

There is a timeline for the Jaytee studies and disputes.


1991 — Pam Smart’s parents notice Jaytee’s waitingforher return actions. PS has regular work hours.

1993 — PS laid off; hours of return vary; Jaytee’s waiting actions still track her accurately.

1994, April — PS reads about Rupert Sheldrake’s interest in dogs like Jaytee, contacts him, and PS’s parents begin keeping a log. PS also keeps a log. Between May 1994 and February 1995, 100 logs are kept.

1994, November — Austrian State TV, science division, records a videotaped experiment on Jaytee at home.

1995 — Between June and December, Wiseman, Smith and Milton are permitted to conduct their own experiments (4) on Jaytee at home by RS and PS et al. They also use videotape.

1996, August — Wiseman and Smith present their results at a conference.

1996, prior to WSM submitting their article published in 1998, RS gives them his graph data on Jaytee.

1998 — RS and PS publish a paper describing those 100 records (from 1994-1995) and a large number of other experiments they have conducted since 1995 that entail varying appropriate parameters, such as means of return, randomized times of return and the like.

1998 — WSM publish the paper that Wiseman and Smith presented at the 1996 conference.

1999 — Sheldrake publishes a critique of the WSM paper.

1999 — Sheldrake publishes Dogs That Know When Their Owners Are Coming Home with full treatment of Jaytee.

2000, January — WSM respond to the 1999 critique.

2000, later — RS replies to that.

2000 — Sheldrake and Smart publish their second paper on Jaytee. (Also here in journal format.)

2011 — 2ed of RS’s Dogs That Know book includes an appendix summarizing his many encounters with skeptics. This includes the Jaytee experiments.

2015 — an Orban Wallace video is referenced by RS that summarizes the Jaytee situation. Features Pam Smart as well as RS.

2015 — Sheldrake summarizes his refutations of Wiseman.

############

Shermer speaks directly to the Jaytee experiments in two paragraphs. I quote them here in italics, inserting [square-bracketed numbers] for reference purposes in the analysis I give following the two paragraphs. The linked paper he begins by referencing is the one listed above at 2000, January.


In this this paper by Richard Wiseman, Mathew Smith, and Julie Milton, for example, the authors reveal what happens when you operationally define [1] what constitutes a “hit” in psychical research, in this case whether or not Jaytee the dog knew when his owner Pam was coming home. In initial observations by the owner it seemed like Jaytee had foreknowledge based on his moving from the house to the porch. Sometimes Jaytee went to the porch when Pam was coming home, but there were plenty of times when Jaytee went to the porch with no connection to Pam at all. Wiseman et al. insisted on testing the claim [2] by actually counting the number of such behaviors and especially the length of time Jaytee would stay on the porch waiting for Pam to return. Under these conditions Jaytee went to the porch 12 times without correlation with Pam’s return. One explanation for this nonsignificant finding is that Jaytee was distracted by the neighbor dog in heat and thus went to the porch with something else in mind. Wiseman et al. returned months later and carried out two more experiments but failed to find any pattern [3] between Jaytee’s behaviors and that of his owner.

Your attempt—after the fact—to find a pattern in the video data by changing the criteria of a two-minute stay on the porch to ten-minute chunks of time during which Jaytee allegedly spent more time on the porch during those periods when the owner was returning home than not, was gainsaid by the authors when they noted that such patterns should arise naturally by the fact that a dog is likely to do little after its owner departs, but then as the day goes on he is more likely to start anticipating the owner’s return (just by normal time elapse and the dog’s memory of the owner’s usual time away) and make more trips to the porch. As well, searching the video record post hoc for patterns is a form of data snooping that is subject to the confirmation bias, and allowing Pam to determine when she would come home means her behavioral patterns might not be random at all but subject to her own unconscious preferences that Jaytee may have learned over time.


[1] I am unaware of any significance Shermer may intend for this expression in contrast, perhaps, to some non-operational definition he wants to lay at Sheldrake’s doorstep. But I am aware that everything about these experiments certainly turns on how one defines a “hit” for the dog. We must, then, discuss this carefully.

If we refer to [2] above, we see that Shermer appears to think Pam and her parents define a hit in terms of Jaytee’s moving from the house to the porch. Shermer shows he is aware of the pretty obvious fact that the dog might go to the window for reasons having nothing to do with his owner’s return and states that Wiseman et al. insisted on testing the claim [2] by actually counting the number of such behaviors and especially the length of time Jaytee would stay on the porch waiting for Pam to return.

But what claim is he referring to? One presumes a claim implicit in the sentence just before — In initial observations by the owner it seemed like Jaytee had foreknowledge based on his moving from the house to the porch. To me, it seems like a person would need some way of distinguishing the visits to the window that have nothing to do with Pam’s return from those that do have something to do with it. The parents’ claim, then, must entail their thinking they have some way of telling the coming home visits to the porch from the non-coming home visits.

I take it that Shermer’s “by actually counting” expression is the key to what the parents and any thoughtful observer must think – that the visits to the window must become much more, obviously more frequent as Pam gets closer to home. Likely the visits would turn into more extended stays as well. There would be a correlation of some sort between visits to the window and Pam’s approaching home. There should be a pattern between Jaytee’s behaviors and that of his owner.

But the Wiseman experiments failed to find any pattern. Indeed Jaytee went to the window in one experiment 12 times without correlation with Pam’s return.

Case closed? No psychic dog after all?

Apparently, from what Shermer says in his other paragraph on Jaytee, Sheldrake thinks there is some significant pattern. It looks like — “after the fact” — he decided to switch criteria for a hit claim from a two-minute stay on the porch (when Pam was coming home) to spending more time on the porch during 10-minute intervals when she was coming home than when she was not. “Allegedly.” We get the impression that this is some sort of trick used to game the experiment after it didn’t work out the way Sheldrake had hoped. Those 12 times Jaytee went to the porch when Pam was not coming home get split up into 10 minute periods now and don’t seem as significant compared to maybe the three times in one last 10-minute period that he went to the porch just as Pam was coming home.

Furthermore, Shermer points out that Wiseman has a credible explanation for the increasingly frequent visits in 10-minute units as the day wears on and the owner has yet to return — anyone, certainly a dog, would start getting more concerned with time and go to the porch to check on things more frequently as time wore on. And two more things: one, this searching the data for patterns after the facts have disappointed you sets you up for confirmation bias, and two, allowing Pam to determine her own times of return permits her sympathetic dog to align himself with those by simple virtue of knowing his owner well.

We might wonder why Shermer/Wiseman attempt to discount more frequent visits in 10-minute units as Pam is on the way home if they “failed to find any pattern.” I think the clear implication is that that pattern was found but was discounted as any sort of “hit” for the reasons given. They failed to find any significant pattern, we might say. This is a minor omission.

So let’s summarize the case against Sheldrake’s take on Jaytee:


  1. Wiseman actually counted times and durations of visits by Jaytee to the porch. He used a “hit” definition of a two-minute duration stay on the porch as Pam was coming home. One time Jaytee visited the porch 12 times when she was not coming home. In four different experiments Wiseman failed to find any pattern between Jaytee’s behaviors and that of his owner.
  2. After the Wiseman experiments failed to show enough hits, Sheldrake changed his definition of a hit to significantly higher frequency of visits counted in ten-minute intervals as Pam is on the way home. Making post hoc changes like this suggests a person has confirmation bias.
  3. Although counting this way there may have been (“allegedly”) some “pattern between Jaytee’s behaviours and that of his owner,” a pattern like that should be discounted because it is easily explained as a result of increasing anxiety about his owner’s return on the part of a caring dog as the time wore on.
  4. And, in any case, the fact that Pam got to choose her own times of return in Sheldrake’s experiments opens up the possibility that the dog may have learned Pam’s decision patterns and responded to them in the experiments.

A disinterested reader wanting to decide this issue on a sound basis should begin by confirming or modifying my paraphrase of Shermer’s argument and then reading the Wiseman reference that Shermer uses. Eventually, of course, said disinterested decider should read Sheldrake to see how he answers this.


#############


OK, now for the WSM paper itself. This is not the Wiseman response to Sheldrake’s objections to his original paper, but the paper itself from 1998. The one to which Sheldrake originally objected.

The first thing I notice is that I am wrong about supposing Wiseman noticed the pattern of significantly greater frequency of visits to the porch in the 10-minute units while Pam was on the way home. The writers (WSM) don’t even mention that.

Instead Wiseman defines a hit as when Jaytee would signal that Pam was coming home during the ten-minute period in which she did in fact decide to come home and start to come home. (The experimenters arranged things so that there was very little time between decision [determined at random by the experimenters] and actual movement towards home.) A signal was defined as a movement by Jaytee to the porch that was inexplicable. In case of more than one such movement during an experiment, the first one was taken to be the signal. For the second to fourth experiments this definition was altered to be the first such movement that resulted in a stay of more than two minutes at the porch window.

Using this definition of signal and hit, Wiseman, Smith and Milton do indeed conclude that “in all four experiments Jaytee failed to accurately detect when PS set off to return home.”

Items 2, 3 and 4 above do not appear in the WSM paper. Let us look at the subsequent exchanges between Wiseman and Sheldrake to see if they appear there.

First we have the initial response (1999) of Sheldrake to Wiseman. This is extensive, including much more than we are looking at presently. With regard to point (2), there is the contrary. That is, Sheldrake points out that the Wiseman studies were initiated in the train of publicity around the Austrian TV special which came about due to the studies he and Pam Smart had already been conducting (more than 100 experiments at the time). These experiments did not include any mention of signal or hit, but were focused on the relative frequency of visits to the window/porch that Jaytee made in three periods: the main period of Pam’s absence, the ten minutes just prior to the start of the return home, and the first ten minutes of the return home. Sheldrake reports that in September of 1996, in discussions with Wiseman about his 100 and more experiments, Wiseman objected to his method of graphing the results of the main period and suggested that he use 10-minute segments for that period as well. This Sheldrake did.

Far from changing his definition of hit, Sheldrake was simply continuing his experiments as before and commenting on the difference between his own experimental design and Wiseman’s. The issue of which design might better reveal any psychic activity of the dog is highly relevant, but not to Wiseman, who simply uses his own design and completely disregards the Sheldrake experiments. (As Sheldrake invited Wiseman to do the experiments with Jaytee in the first place, and as he reports extensive conversations with Wiseman about his own design and results, it seems strange that Wiseman would write up his own experiments so briefly and without mention of the Sheldrake experiments. Instead, Wiseman looks only at the Austrian TV’s entertainment study of the dog’s behavior.)

Now as to (3) there is no discussion in the first Sheldrake response to Wiseman. However, we know from the 1999 Sheldrake paper (item 5 of the Discussion) that this possibility was controlled for, tested for and rejected by several means, one being a control series of tests on evenings when Pam did not return at all.

Neither does the (4) point have bearing, although Sheldrake does not discuss it in this first response, simply because there are many tests in which Pam does not herself decide when to return, but that is determined randomly.

I will postpone Sheldrake’s own 4-point rejection of the Wiseman 1999 conclusions about Jaytee until we have dealt with exactly what Shermer is saying in his response to Sheldrake at Better Schools, specifically in the two paragraphs he has there about the Jaytee experiments that Sheldrake mentions in his opening statement.


###########


And now for the Wiseman writing that Shermer actually links to in his response to Sheldrake in the Better Schools dialog.

This is a panel. Close it by clicking somewhere not on it. May take two clicks.

May include gaphics and hyperlinks.


(The porch is often referred to as the “window” by Pam and Sheldrake. This discrepancy is resolved by watching the Austrian tape of the house. It looks to me like a small porch has been enclosed with glass windows and a house wall removed so as to expand the living room of the house.)

It is necessary to read the experimenters’ paper carefully to pick up these two definitions. Start reading in the middle of page 6 and go on for two pages and the definitions will become clear. The key word is “inexplicably,” which occurs 7 times. They use “for no apparent reason” once.