Sai Baba Dies
From New York Times (nytimes.com:)
I received a forwarded note from Grzegorz Osipiak of Poland, who said,
"I don't know if you are aware that On 24th April 7.40AM Indian time Indian holy man Sri Sathya Sai Baba died(left his body). He has estimated 50,000,000 - 100,000,000 devotees around the world. For many millions he has been worshiped as an incarnation of God. His passing is enormous tragedy and emotional trauma for millions.
"I have noticed that on 23th April between 21.00 and 22.00 Global Consciousnes Project graph showed very, very strange behaviour. Could you please check it for yourself if you have few spare minutes?"
The GCP event was set for 7:00 to 13:00 Indian time, beginning about 40 minutes before Sai Baba died, and continuing for 6 hours to capture some of the immediate reaction around the world. The result is Chisquare 21057.486 on 21600 df for p = 0.996 and Z = -2.629. The figure shows that this result is the culmination of a steady negative trend. It is opposite to our standard prediction, but it is worth noting that such trends have been seen in numerous meditation and religious events.
Because Grzegorz Osipiak mentioned that the preceding day was apparently unusual, I decided to look at the larger context. Sai Baba's death was near the beginning of the UTC day of the 24th, so taking the 23rd and 24th of April provides roughly a day before and a day after the death. The resulting graph is very interesting, with a huge and unlikely rise and fall on the 23rd, followed by a continuing fall (low Network variance) during the first part of the 24th -- specifically during the 6 hour period I had chosen for the formal hypothesis test. Since Osipiak was not clear about the time zone for his observation of strange behavior between 21 and 22:00, we can't be sure how it relates to this figure. In any case, the high peak (where Netvar changes from strong positive to strong negative deviation) occurs at about 21:30 India time.
It is important to keep in mind that we have only a tiny statistical effect, so that it is always hard to distinguish signal from noise. This means that every "success" might be largely driven by chance, and every "null" might include a real signal overwhelmed by noise. In the long run, a real effect can be identified only by patiently accumulating replications of similar analyses.