I can't read that because paywall, *but*...this isn't new in the sense the CEPR was making more or less the same points two months ago. I don;t know what this one says, but the two different interpretations in part hinged on which districts were still to report at the time of the interruption of the count (CEPR said they were mostly rural and hence very pro-Morales, so no surprise numbers higher, OAS said the still-to-come districts included La Paz and (IIRC) Cochabamba, which are pretty urban, never saw either side try to resolve that discrepancy, would be interesting if this report did, because that has a major impact on you interpret the numbers.
*Also*, if you read the investigation report, the heart of the OAS claim about fraud was not a statistical claim at all, but a series of claims about poorly-handled ballot boxes, loads of problems with ballot chains of custody, electronic voting reports being suddenly diverted to an unknown server before being counted, etc etc. It was a pretty damning series of allegations. Arguing over two different statistical interpretations doesn't really absolve anyone of interference with the ballots. (and of course the thing about the mystery server is you can't prove that votes were manipulated, only that someone put themselves in a position to manipulate them)
I mean, it seems to me you could build a pretty plausible case for both sides. Here's a scenario that fits the known facts (not saying it;s what happened, just that it is consistent with data presented by both sides):
1) Evo's people *really* wanted to win on the first ballot because they didn't think they would get 50% on the second (his losing the referendum on an extra term being a pretty good barometer)
2) Polls and early results showed he was *really* close to that 10%. But too close to be sure.
3) The Evo camp swings into action, delays count so it can put itself in a position to tamper if need be (hence the issues with chain of custody, strange changes in servers, etc
4) They look at the raw data. Actually, they are over 10%! perfecto, no need to tamper. The count continues. Evo declared the victor.
Would that have been a stolen election? No. Would that reveal malevolent intent on the part of Evo's team? Yes. Enough to disqualify the result? Eye of the beholder, I guess.
*Also*, if you read the investigation report, the heart of the OAS claim about fraud was not a statistical claim at all, but a series of claims about poorly-handled ballot boxes, loads of problems with ballot chains of custody, electronic voting reports being suddenly diverted to an unknown server before being counted, etc etc. It was a pretty damning series of allegations. Arguing over two different statistical interpretations doesn't really absolve anyone of interference with the ballots. (and of course the thing about the mystery server is you can't prove that votes were manipulated, only that someone put themselves in a position to manipulate them)
I mean, it seems to me you could build a pretty plausible case for both sides. Here's a scenario that fits the known facts (not saying it;s what happened, just that it is consistent with data presented by both sides):
1) Evo's people *really* wanted to win on the first ballot because they didn't think they would get 50% on the second (his losing the referendum on an extra term being a pretty good barometer)
2) Polls and early results showed he was *really* close to that 10%. But too close to be sure.
3) The Evo camp swings into action, delays count so it can put itself in a position to tamper if need be (hence the issues with chain of custody, strange changes in servers, etc
4) They look at the raw data. Actually, they are over 10%! perfecto, no need to tamper. The count continues. Evo declared the victor.
Would that have been a stolen election? No. Would that reveal malevolent intent on the part of Evo's team? Yes. Enough to disqualify the result? Eye of the beholder, I guess.
Comment