Nate Silver debunks graph allegedly proving fraud by Ahmedinejad

[AlexanderProgressive]

Ahmedinejad of Iran may well have rigged the elections in his country. Who knows? but at least one graph circulating the internet (including here in DU) is poppycock, according to progressive statistician Nate Silver of www.fivethirtyeight.com.

This is the graph:



Silvery says: "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."

http://www.fivethirtyeight.com/2009/06/statistical-evidence-does-not-prove.html

[WeDidIt]

1. Nate's debunking is debunked

Nate used an alphabetical listing of final results for his waves.

This differs significantly from how results are reported in real time. For example, compare Indiana 1 hour after the polls closed, 2 hours after they closed, 4 hours after they closed, etc. then the final results. They are all over the board.

[AlexanderProgressive]

3. I thought you were going to link to an expert who debunked Nate's debunking n/t

[BlooInBloo]

2. Silver acknowledges that he hasn't shown the election to be clean.

[AlexanderProgressive]

5. This thread does not make the claim that the election was clean

It is related to one specific graph.

[BlooInBloo]

6. That's good, since I didn't say it did.

[AlexanderProgressive]

7. Then why did you bring up the fact that Silver didn't say the election was clean?

It was unnecessary.

[BlooInBloo]

8. No more unnecessary than your post.

[AlexanderProgressive]

9. To state the opinion of an expert disagreeing with a widely circulated graph is unnecessary?

Wow.

By the way, you are changing the subject. We were discussing the fact that I never said that Silver claimed the elections were clean.

[BlooInBloo]

11. And I never said you said that Silver claimed the elections were clean...

I'm not sure how vehemently you want my agreement with you on that point to be.

[AlexanderProgressive]

16. And I asked you why you said that Silver didn't say the elections were clean, but you ignored me

Because the only reason why you would bring that up is if you thought I said that Silver said that the elections were clean.

[BlooInBloo]

17. You criticized my saying that on the grounds of it being "unnecessary"...

I challenged those grounds, based upon the fact that there was absolutely no "necessity" in your post, either.

[AlexanderProgressive]

24. That's not a challenge. That's the "you did it too" fallacy

I asked you why it was necessary to say what you said, because in the absence of an explanation, I'd have no other option than to think that you were claiming that I said Silver said the elections were clean.

But you did not address my question. You simply replied that my Original Post was unnecessary, as if that were relevant to my question.

And here's why your observation (about Silver not claiming that the elections were clean) was unnecessary: The first thing I said in my OP was that we don't know whether Ahmadinejad stole the elections. People reading this thread then knew immediatly that I didn't necessarily consider the elections to be clean. But you made the observation for a reason that is yet to be explained.

By the way, you also ignored my question about why you think that linking to an expert disagreeing with a widely circulated graph is unnecessary.

[LittleBlue]

4. HAHAHA, so much for the morons posting graphs they don't even understand

lololol

[alcibiades_mystery]

32. Why not? They don't understand one thing about Iran. Why stop there?

ROFL.

[grantcart]

10. Silver's comparison is with 6 waves

But in Silver's waves the groups are put together by state alphabetically rather than geographically.


We have seen the phenomena where leads gain and are lost during the night when large numbers of rural and then urban votes come in.

In Iran this did not happen. It started at 63% and ended up at 62% with the numbers being consistent regardless of the areas that were reporting in.

[nsd]

21. I'm not sure grouping states geographically would change things that much.

Just now, I took the 2008 popular vote totals, divided states into seven groups by geography (New England, Middle Atlantic, Midwest, Southern Coastal, Southern Central, Mountain West, Pacific), and made the same plot using the groups in that order. The resulting graph looks pretty linear, though not quite as linear as the Iranian one. The r^2 is 0.992.

Maybe the rural-urban divide matters, as you say, but I think Silver's essential point -- we can't from this graph alone conclude anything -- still holds. We'd need to know something more about Iran's political demographics and in what order the votes came in.

ETA: A little off topic, but it would be interesting to know what the networks' running totals on the U.S. popular vote were on election night, to see quite how much they zigged and zagged. Does anyone know if that's available online?

[grantcart]

22. no but taking all of the urban vote in, say CA would give a completely different

percentage than the rural areas.


The point of the graph was that the results, that normally 12 hours to assemble were released within 3 hours and that the rate was absolutely consistent regardless of where the vote was counted.

Now there are even more compelling arguments, like the rate of Ahmadinejad's victory in Mousavi's home area is a landslide, and again very similar to the percent in each of the 6 stages.

If you have to arrange false numbers in a very short period of time and make sure that it all comes out as you want, a simple method would be to simply make sure that all parts of the country and each of the stages, had the same percentage of voting.

But there are more compelling arguments including the extremely low numbers for Ahmadinejad's conservative opponent, even though he scored some impressive shots in the debate, and conservatives who don't like Ahmadinejad personally, but would eventually support him over the moderate, could vote for the conservative and vote for Ahmadinejad in the final.

It is universally perceived that Ahmadinejad's base of support has declined, but he received 4 million more votes in a 4 way run off than he did in the 2 way run off in 2005.

NPR had an interesting series of interviews, some of which were in rural areas supporting Ahmadinejad. Those interviewed, even those that supported him agreed that his percentage of support had declined from being nearly unanimous in 2005 to having a plurality in the region this year.

The most compelling evidence is the resignation of Rafsanjani and other Ayotollah's asking that the election be set aside.

[nsd]

23. You're right on all this, but I think Silver was making a much more modest point.

He was simply arguing that you can't look at a graph like this, without knowing whether rural and urban areas (or different ethnic regions or whatever) did indeed report at very different rates, and conclude that the election was rigged. He simply meant statistics alone, absent other knowledge, can't get you that far.

Your other arguments -- about the results among the Azeri, the speed with which the totals were reported, etcetera -- are solid, but they're different in character and don't have much to do with the FiveThirtyEight post.

[camera obscura]

12. I heart Nate Silver but he was wrong here. It's been debunked in his comments.

*edited because the heart symbol caused my subject line to screw up. Weird!*

[BlooInBloo]

15. Less-than and greater-than are used in html tagging, which is disallowed.

[ClarkUSA]

13. Here's further evidence of a coup (link and quotes -->)

On the basis of what we know so far, here is the sequence of events starting on the afternoon of election day, Friday, June 12.

:bluebox: Near closing time of the polls, mobile text messaging was turned off nationwide

:redbox: Security forces poured out into the streets in large numbers

:bluebox: The Ministry of Interior (election headquarters) was surrounded by concrete barriers and armed men

:redbox: National television began broadcasting pre-recorded messages calling for everyone to unite behind the winner

:bluebox: The Mousavi campaign was informed officially that they had won the election, which perhaps served to temporarily lull them into complacency

:redbox: But then the Ministry of Interior announced a landslide victory for Ahmadinejad. Unlike previous elections, there was no breakdown of the vote by province, which would have provided a way of judging its credibility

:bluebox: The voting patterns announced by the government were identical in all parts of the country, an impossibility (also see the comments of Juan Cole at the title link)

:redbox: Less than 24 hours later, Supreme Leader Ayatollah Khamenei publicly announced his congratulations to the winner, apparently confirming that the process was complete and irrevocable, contrary to constitutional requirements

:bluebox: Shortly thereafter, all mobile phones, Facebook, and other social networks were blocked, as well as major foreign news sources.

All of this had the appearance of a well orchestrated strike intended to take its opponents by surprise – the classic definition of a coup. Curiously, this was not a coup of an outside group against the ruling elite; it was a coup of the ruling elite against its own people.

Link: http://andrewsullivan.theatlantic.com/the_daily_dish/2009/06/the-mullahs-last-stand.html#more

[BlooInBloo]

14. Evidence doesn't matter to this one, for some reason. But to those to whom it does...

Here's a little bit more. From one of the pre-eminent Middle East scholars in the country.

http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=389&topic_id=5842550&mesg_id=5842550


EDIT: Forgot linky. Heh.

[ClarkUSA]

18. Yes, and there are a couple of others, too. Interesting, to be sure. Hmm...

:wtf:

Anyway, thanks for the good read. You should bring it over here as an OP tomorrow, for maximum exposure. Juan Cole's voting graph analysis is mentioned in the evidence I linked to.

[BlooInBloo]

19. Sully is all over the shit with questions, as well.

[ClarkUSA]

20. Sully is going to be a pit bull with this. Having the nation's #1 blogger on your side is good.

[AlexanderProgressive]

25. Sullivan #1 blogger?

That's as subjective a declaration as one can make. He's clearly your favorite. I can name about 10 I'd put ahead of him.

[invictus]

26. It's funny how all the neo-cons come out to support Ahmadinejad.

nt

[spiritual_gunfighter]

35. It was telling

That Fox News had one of their "question" banners today that said "Does it even matter who wins the Iran election?" so obviously there is some concerted effort on the part of the neo-cons to get behind Ahmadinejad on this one. It is important to have a boogey man.

[PretzelWarrior]

27. why are you around here trying to debunk the obvious?

I notice many posts and threads denying the Iranian elections were stolen are coming from you.

[AlexanderProgressive]

37. That's false. I haven't "denied" anyone stole the elections

This thread is about a particular graph, and in my first sentence I said we don't know whether the elections were stolen or not. That's far from denying theft. It's asking for sufficient evidence.

[jeanpalmer]

28. According to Mousavi's movie drector campaign spokesman

the Interior Ministry called Mousavi and told him he had won the election. And yet each time the vote was reported by the Interior Ministry, as shown in the graph, Ahmadinejad was winning. So the IM must have called Mousavi, if they called him at all, before the results were released, i.e., before the polls closed. Once the polls were closed and the results began to release, Mousavi would have been able to see that he was losing and the call from IM would have made no sense.

So did the Interior Ministry play a prank on Mousavi by telling him before the vote tallies were announced that he had won the election? When in fact he had gotten trounced. That looks like a good possibility. They were just teasing him. That's why the IM told Mousavi to be careful not to announce the information so as not to rub it in to Ahmadi's supporters. And then after the IM had wound up Mousavi and his supporters, they announced the results and the bottom fell out. They were just funning him.

[Name removed]

29. Deleted message

Message removed by moderator. Click here to review the message board rules.

[jeanpalmer]

30. Well, it's hard for me to believe

if the Interior Ministry was rigging the election for Ahmadinejad, that it would call Mousavi and tell him he had won. Come on, if it were true that he had won, they'd never let that cat out of the bag. That information would have been suppressed. They certainly wouldn't have told Mousavi.

So if you believe Mousavi that he got a call from IM telling him he had won, when in fact IM was getting ready to announce a landslide for Ahmadinejad, well you have to conclude they were playing a prank. And then they threw in the dig, "Oh and by the way, don't tell anyone else about it, you don't want to rub it in." LOL. A classic prank. The IM was laughing its ass off, having a good time.

There's no other way to interpret that, given the sequence of events.

How am I a fraud? I really don't care how this turns out. The main thing I want to see is our country keep its nose out of Iran's internal affairs and get over the notion that it has the right/obligation to interfere in those affairs. Judging from what I'm reading tonight, it looks like people haven't learned that lesson. Iran would probably be better off without Ahmadinejad. But that's their decision and I'm willing to respect it and them.

[PretzelWarrior]

31. what would be your viewpoint if it COULD be proved a fraud?

Oh boy, too bad? Just curious.

[jeanpalmer]

33. If it turns out the election was rigged

I hope they do something about it. I don't know how that process will work in Iran. Is there any review of election results? I can't believe a government could keep a fraud of that magnitude quiet. The word would get out.

But OTOH, let's say Ahmadinejad won fair and square, what would be your viewpoint on that? See, I'm willing to let the chips fall. If the Iranian people want MA as their president, so be it. I have no problem with that. If they want to overthrow their government, that's ok too. Although it would be unprincipled for a party that lost an election fair and square to then resort to force to try to win. So it is important to get the facts to find out what happened.

What I found amusing tonight was how so many people assumed that it was a fraud because MA won. And I have yet to see any hard evidence. I've seen speculation and allegations, but no hard evidence.

I'm more worried about my own government that allows torturers to remain in its employ, that still routinely lies to its people as it did a few weeks ago when it killed 140 innocent people in Afghanistan. To me, the biggest threat to the world is the USA, not Iran. I find it wacky that people want to reform Iran's government when our own government is out of control, heading over a cliff.

[johnaries]

34. Well, I'm looking at the graph and I'm thinking

it doesn't say anything. It's useless and meaningless. :shrug:

edit to add: Personally, I think the election was fixed, but this graph doesn't prove anything.

[northernlights]

36. I think the point is that each data point sits exactly on the line

In a real tally, the data points would lie on either side of a "perfect fit" line. Assuming each data point represents a region of counting (or a different elapsed time, as one voting block reports in and then another), each data point would show a different relationship between the numbers of votes. At one point, the gap between them might narrow, one might overtake the other, and then it would reverse.

As it is, the data points suggest a functional relationship between the candidate's votes -- for every vote Mousavvi got, A-jad got 1.x -- each and every time, no matter which town or city or whatever each data point represented.