Thanks, that comment looks much less fishy than the original article. I wonder whether the Washington Post authors tried to run their test on past known rigged and non-rigged elections. If not, that would be a good idea.
There are a number of causes for caution here, such as the possibility of publication bias. There could be 100 other statistical tests that the election passes, but who would publish those boring results?
Another problem is how reasonable the null hypotheses are. Should we really expect Iranian vote counts to fit Benford's Law in a fair election, for example? I don't know.
I have no reason to believe the election was fair, but I think any scientific results reported directly in news media without peer review a week after a very controversial election are highly suspect.
This is actually pretty poor statistics. They are basing their entire argument on ~100 provincial results. The deviations they are talking about with this few number of samples isn't likely to be significant.
I thought that too until I re-read the article and started trying to do the math myself.
They sum it up nicely but it's too easy to miss:
"The probability that a fair election would produce [these number distribution disparities] is less than .005. In other words, a bet that the numbers are clean is a one in two-hundred long shot."
Key thing to take away from the article if you're going to perform election fraud, use a computer to produce your random numbers.
"The probability that a fair election would produce [these number distribution disparities] is less than .005."
Careful with that reasoning. Suppose there are four candidates, and the last digits of their total vote counts are 8,7,2,8. Fewer than 1 in 9999 fair elections would produce a vote count with exactly those digits!
Does this mean that the vote was rigged?
Post-factum searching for (perceived) oddities and calculating probabilities of their occurences in uniform distributions shows some misunderstanding of statistics.
Also note: "But many districts where the excess votes were recorded are small, remote places rarely visited by business travelers or tourists, analysts said"
And so on, problem after problem.
I dare say, the next election there will be "handled" better... :-)
The problem with that assessment is the more tests you do the larger your chance of finding something odd.
In isolation you can look at the odds for 17+ and 4-, but two <5, or two > 16 would also look odd. As would a single >20 or <4. And, they started looking at 2 digits but the pattern is only in the last digit so there are looking at 3 groups.
PS: Also how did you calculate the odds? A single number of 17 or more increases the odds of a number of 4 or less. I am not really sure how you would calculate the odds outside of a simulation.
> They sum it up nicely but it's too easy to miss:
> "The probability that a fair election would produce [these number distribution disparities] is less than .005. In other words, a bet that the numbers are clean is a one in two-hundred long shot."
Man, reading that makes me roll my eyes and wonder how the average person reading this site has such poor statistical intuition as to mod this up.
It's like rolling a die six times, and arguing that getting the exact numbers I got has an exceptionally low chance of occurring, so the die must be rigged.
Taleb makes an interesting argument regarding this in The Black Swan: If you flip a coin five times and you get five heads, what is the probability that the next flip is a head?
You have two options: either you trust your (unproven) assumption that the coin is fair and that you just witnessed a very unlikely event and say "the past outcomes do not affect the probability of the next throw, it's 0.5."
The other option, on the other hand, is to conclude that your assumption that the coin is fair is false, at which point you estimate the next throw based on the previous sequence and say "1.0".
None of these are right or wrong, it just depends on what prior you trust.
I wonder if the first digits comform to Benford's Law. Basically, the first digits of the vote counts should be 1, 2, 3 much more often than they should be 7, 8, 9
Indeed. It's quite non-intuitive, but the naive assumption that the last digits are randomly distributed is simply wrong. Here is a V.I. Arnold essay about Benford's Law: http://www.math.ru.nl/~mueger/arnold.pdf
At this point it doesn't matter if it was rigged or not. The regime has voided the social contract with the people of Iran and is therefore illegitimate.
The question whether the election was rigged matters for the same reason that the question of Iraq's WMDs mattered: this information is influencing Western public opinion and government decisions right now. Also, going "it doesn't matter 'cause you're bad anyway" whenever someone calls attention to the substance of your accusations strikes me as a bad faith tactic.
Ha. To be blunt, I don't think most folks on Hacker News are thinking about this in good faith. If some other country this community doesn't like a priori, say the US, killed a bunch of its citizens after an election, there'd be zero chance everyone here would be rationalizing away statistical improbabilities in the election results.
> Others have pointed to the surprisingly poor performance of Mehdi Karroubi, another reform candidate, and particularly in his home province of Lorestan, where conservative candidates fared poorly in 2005, but where Ahmadinejad allegedly captured 71 percent of the vote.
If, for example, Broward County had come up 71% for Bush in an election, that would have been considered proof of election fraud around here. In this case, it's rationalized away. Poof.
Debating the validity of Iran's election results at this juncture is kind of like worrying about whether Caesar is too ambitious after he's already crossed the Rubicon. Shooting peaceful people to death in the streets and cutting off any means for the outside world to learn about these atrocities tells us all we need to know about the Khamenei/Ahmadi regime.
The validity of the election results still has the potential to change Iran's government, so yes, it's still worth talking about. If there is enough evidence of fraud, and the world and all the Iranian people know it, then there's a higher chance the current regime will relent. If the evidence is inconclusive, then the regime is more likely to remain in power and preserve the status-quo.
How does that follow? Would anyone worth consideration seriously believe this government is legitimate even if the election wasn't fixed? What sane person says, "oh, they've shot dozens of people to death and tried to cover it up, but at least they didn't rig an election"?
The regime had their chance to relent, and they only cracked down. It's not up to them anymore.
I believe you are interpreting what I'm saying in the wrong manner.
The Iranian government is autocratic and has probably abused its power to maintain that power. I take that as a given in any discussion of recent events. But that fact alone does not mean they will lose their power. Unfortunately, it probably is up to them, since they have most of the power in Iran.
In order for the Iranian government to lose its power, it must either relent and cede some power in the hopes of maintaining what some of what they have, or the Iranian people have to revolt. Both are more likely to happen if there is solid evidence that the election is rigged.
It's not so simple as you make it out to be. The Iranian government is a complex mix of democracy and theocracy. If enough mullahs and enough of the populace are turned against Khamenei and Ahmadinejad, Khamenei and Ahmadinejad will fall. The murder of Neda Agha-Soltan, and any murders that follow hers, will turn more mullahs and more common Iranians than a fixed election ever would. (Everyone thought the elections were fixed anyway. That's how Ahmadi got elected the first time--the reformists boycotted the election.)
I don't disagree with anything you said. But, if I understand you correctly, you don't think that demonstrably fixed elections would effect that complex mix at all. I think they would.
My simple rational is that you're going to have more support from the people and lower levels of the government if one can provide compelling evidence the elections were fixed. There's a difference between most people thinking the elections are fixed, and actually having evidence to support the conclusion. Such knowledge emboldens people.
I don't think the argument is that "you're bad anyway". Regardless of who you are, how you got into that situation and whether the protesters have a point, with a protest movement that big and motivated your options are same: crush the protests, re-run the election or form a unity government. If you choose to crush the protest then you give up the "government by consent of the people" type contract that you have with society. I think it's that fact that will drive western opinion more than the actual legitimacy of the result.
exactly. It matters little anyways because it seems that Ahmadinejad and Yazdi were just going to prepare a coup anyways -- http://www.ireport.com/docs/DOC-279948
At this point I can't sincerely pick any side as the good guys. On one hand, I'm already sick of cookie-cutter "color revolutions" where "nonviolent" protestors oh-so-conveniently invite the support of Western nuclear superpowers to topple regimes unfavorable to the latter. On the other hand, I'm already sick of the only offered alternative, the emerging "multipolar world" fraternity of Ahmadinejad, Putin, Chavez et al.
I kind of envy you Westerners. You have a simple moral principle that tells you which side to pick, a never-failing axiom: democracy is always best. But the recent history of my country (Russia) has given me a huge and readily observable data point against your axiom. So I can't honestly share your faith, even though I'd like to.
So what's your point? That the people who want a say in how they're governed shouldn't express a view because the people who don't like to have a say are sick of hearing it?
I assume you're referring to the second sentence of my comment.
The idea of protesting against bad governments, all the way to armed rebellion, is perfectly okay with me. But I intensely dislike the idea of appealing to a foreign power to settle an internal dispute, especially if the appeal gets prompted by same foreign power. The classical conception of sovereignty, now sadly defunct, had an apt name for this technique: treason. Today it's called "color revolution".
So protest is right if people go to kill their government by themselves?
And when they can't do it by themselves and call for help, this is treason even when they didn't wanted the government that rules them in the first place?
Now that's black and whit simplistic view of world.
Is the concept of treason even relevant when you live in a your country and somebody rules it without your (and other people living there) agreement?
I mean - I live in Poland, after WW2 Soviet Union ruled here (ok, not directly but by proxy:), is treason of such proxy government really treason? For me it's heroism, because people in Poland didn't wanted that government in the first place.
Also - for me peaceful revolution is better than blood bath revolution any time.
From here, it seems like Russia hasn't quite gotten how democracy works yet. Which is fine--maybe democracy doesn't work for Russia. But Iran's a more promising case.
I hope you'll agree that there's another moral principle at play here: don't shoot peaceful people to death in the streets. I couldn't care less about elections when a government stoops to that.
You're assuming that whether the reported totals matched the votes is the only relevant point.
It is not the responsibility of critics to show that an election was fraudulent. It is the responsibility of a government to convince the populace that the election was legitimate. If a large portion of Iran thinks that the election was fraudulent, that's a good enough reason to support their efforts to replace that government.
Al Gore should have gotten his Nobel for preserving the republic by letting the 2000 election die in an orderly fashion rather than for his climate scaremongering.
This article is pure shit. Anyone with a reasonable math education should know why, and should know about data mining. That you can get away with such things in major newspapers is simply shameful.
Uh, forgive me if this doesn't make any sense, but why would the numbers of the actual votes be generated fradulently?
Wouldn't it be easiest to say, "Okay, Ahmadi is going to win 57% to 65% in all reporting districts," and so on and so forth and then use those fractions and the known eligible voter numbers to come up with the actual counts?
In which case, shouldn't we be looking at the last digits of the percentage of votes won by district?
Well, the thing about percentages in an election is that a lot of them don't _have_ last digits. Further, if the percentages came out to nice numbers like 57% and 65%, we wouldn't even be having this discussion - it would be really, really obvious.
Use more detailed results from the Iranian election and you don't see this effect.