|
On Wednesday October 17, 2001 An Israeli minister was shot dead in a
Jerusalem hotel today by a
suspected Palestinian gunman, prompting fresh fears for the shaky
truce between Palestinians and
Israelis that was agreed only three weeks ago. The Syria-based
Popular Front for the Liberation of
Palestine immediately claimed responsibility for murder of the
far-right tourism minister, Rehavam
Ze'evi.
I received a prediction from Joseph Menahem, who hosts an egg in Beer
Sheva, Israel. He said,
"It created a big emotional reaction here."
Joseph provided the needed information, and we agreed that
the formal prediction would be for the single Israeli egg, and that we
would do explorations to include the whole network response.
Note added June 26 2004: For consistency, since this is the only
instance of a single egg being specified, we have decided to use the standard
full-network specification for the re-analysis being done by Peter
Bancel in 2004. This changes the result somewhat, to a stronger outcome.
Joseph wrote: "The local time in Israel is 2 hours ahead of GMT. The
assassination occurred on the 17th at about 7:00 a.m.
I think the first major news came out at 8:00 a.m. --
at least I heard it the first time. The final and
formal announcement about the minister's death came at
10:00 a.m.
"The second major related event in my opinion is the
funeral which took place on the next day and which was
broadcast live on the T.V. and radio from the
morning till late afternoon. Huge number of crowd
came to visit the coffin on the parliament yard from
9:00 a.m. (18th) till about 12:00, where also some
significant speeches took place and then on the
cemetary his son called directly to the prime minister
who attended there to do what he has to do
(interpreted as a call for revenge - 4:00-5:30 p.m.)."
Joseph specified deviations in data from the egg he hosts. The formal
prediction was for a deviation in the local egg's data beginning at
07:00 and continuing to 10:00 (05:00 to 08:00 GMT).
We specified the standard analysis (but for one egg).
In addition we have an exploratory analysis (which is exactly like the
usual standard analysis) of the full network reaction
to the assassination events. We plan an exploratory analysis of the
funeral event, from 07:00 to 10:00 on the 18th (not yet available).
The formal analysis shows a positive trend, with Chisquare 10919 on
10800 df, for p = 0.209.
The full database analysis of all eggs also shows a positive trend, with
Chisquare 11038 on 10800 df, and p = 0.053. This would be considered
marginally significant if the full network response had been an a
priori prediction. The comparison with the single egg response
is interesting in the context of the question of location. No single
case can answer it, but this analysis and comparison lends an increment
ot the perspective that distance does not matter, and that all eggs
around the world may respond to an event anywhere in the world. But,
again, there are other cases with the opposite implication, so we have
more to learn.
|