Tuesday, March 23, 2010

Error limits at CITES

At CITES meetings governments vote on whether or not to support a ban, or a limitation, in the trade of an endangered species.

Here in Doha, there is an electronic voting system. About 137 governments (of a total of around 180) have turned up. They put a card into the voting machine, then they press the buttons to vote. I've not been here for the full meeting, but from memory I'm absolutely certain that if you press '2' you vote yes. If you press '3' you vote no. And if you press '4' you abstain.

Earlier today, during the Porbeagle vote, the CITES voting system failed and two nations were not able to record their vote. The system was reset. The chair asked everyone to vote 'yes' to test the system. Of the 137 nations that voted yes, 7 hit 'no' and 2 hit 'abstain'.

Unless I'm greatly mistaken, this gives us an error level of 9 for the voting at CITES. In other words, if 100 parties intend to vote yes, about 6.5% will accidentally vote another way. (I know, hard to believe it is so difficult to press a button.)

The interesting observsation is that this is an argument for proposing the hammerhead shark again for a vote at the Plenary meeting tomorrow and Thursday. The hammerhead failed by only five votes.