Nate Silver chimes in on the Iranian elections, sort of.
Statistical Report Purporting to Show Rigged Iranian Election Is Flawed
Like most Americans, there are few things I would like to see more than Mahmoud Ahmadinejad, Iran's hateful President, to be voted out of office. Elections in thuggish, authoritarian states like Iran need be treated with the utmost skepticism and scrutiny. I can't say I have any real degree of confidence in the official results, which showed Ahmadinejad winning with some 62 percent of the vote.
There is a statistical analysis making the rounds, however, which purports to show overwhelmingly persuasive evidence that the Iranian election was rigged. I do not find this evidence compelling.
Iran's election results were reported by its Interior Ministry in six waves. The first wave covered about one-third of the total vote; there were then two relatively large waves that reported about 20 percent of the vote each, and then three smaller waves that reported the remainder of the vote. What other observers have found is that, over the course of the six waves, there is an extremely strong, linear relationship between the number of votes reported for Ahmadinejad and the number reported for his principal opponent, Mir Hussein Moussavi (who had declared victory before any results were officially announced):
This relationship is superficially very impressive -- an R-squared of .998, which suggests a nearly perfect relationship.
Just how remarkable really is it, however?
Read the rest and see the graphs at the source.
Like most Americans, there are few things I would like to see more than Mahmoud Ahmadinejad, Iran's hateful President, to be voted out of office. Elections in thuggish, authoritarian states like Iran need be treated with the utmost skepticism and scrutiny. I can't say I have any real degree of confidence in the official results, which showed Ahmadinejad winning with some 62 percent of the vote.
There is a statistical analysis making the rounds, however, which purports to show overwhelmingly persuasive evidence that the Iranian election was rigged. I do not find this evidence compelling.
Iran's election results were reported by its Interior Ministry in six waves. The first wave covered about one-third of the total vote; there were then two relatively large waves that reported about 20 percent of the vote each, and then three smaller waves that reported the remainder of the vote. What other observers have found is that, over the course of the six waves, there is an extremely strong, linear relationship between the number of votes reported for Ahmadinejad and the number reported for his principal opponent, Mir Hussein Moussavi (who had declared victory before any results were officially announced):
This relationship is superficially very impressive -- an R-squared of .998, which suggests a nearly perfect relationship.
Just how remarkable really is it, however?
Read the rest and see the graphs at the source.