Skip to content

Unconvincing defense of the recent Russian elections, and a problem when an official organ of an academic society has low standards for publication

Last month we reported on some claims of irregularities in the recent Russian elections. Just as a reminder, here are a couple graphs:

Yesterday someone pointed me to two online articles: Mathematical proof of fraud in Russian elections unsound and US elections are as ‘non-normal’ as Russian elections.

I know nothing about Russian elections and will defer to the author and his commenters on the details. That said, I don’t find the arguments to be at all persuasive. The protesters show drastic differences between the patterns of votes of Putin’s party and the others, and the linked articles seem a bit too eager to debunk.

I wouldn’t necessarily blog on this but I was unhappy to see this material on the website of Significance, which is an official publication of the American Statistical Association and the Royal Statistical Society.

The quality control at this site seems low. I clicked through the links and found this:

Barring the revelation of a hoax or a savage Al-Qaeda retaliation, President Obama will be able to stand proudly at the helm during the re-election campaign. Even if unemployment numbers remain against him, other numbers are strongly on his side: zero serious Republican contenders (as of now) to compete for the seat in the oval office, 5 to 1 odds in favor of past presidential incumbent, and one historic executive order that eliminated the single most pernicious foe to ever threaten American security.

First off, it’s ridiculous to say there are “zero serious Republican contenders.” What does it take to be a “serious contender”? Several of the Republican candidates have experience that’s comparable to the eight years in the state legislature and half a U.S. Senate term, which is what Obama had when he ran for president.

Second, Osama Bin Laden was really “the single most pernicious foe to ever threaten American security”? Tell that to George III, Jefferson Davis, Tojo, and Stalin.

Again, I have no desire to pick on the people writing this stuff. I have purposely avoided naming them (you can click through the links to see), as a way of emphasizing that I’m not trying to slam them personally. They are not professional journalists, they’re academic researchers who are trying to do a public service by submitting material to this statistics site.

It’s just that, as a member of the American Statistical Association, I don’t like this sort of poorly-thought-through material being the public face of our profession. I understand the motivation to raise the profile of Significance magazine via frequent postings on topical issues, but I worry about our organizations appearing to endorse some misinformation.

P.S. More on the Russian election here.


  1. Xi'an says:

    I also was unhappy with the tone of those posts, thanks!

  2. Jon M says:

    Couldn’t the differing distributions just be explained by the fact that United Russia is at least somewhat competitive all over Russia whereas the other parties have particular regions of strength and close to zero support elsewhere. So in United Russia’s case you are combining several different regional distributions (which may be Gaussian) but would not be when they were combined.

    This isn’t to say there wasn’t fraud going on: there’s ample domestic and international observer evidence to show that.

  3. Antonio says:

    In my opinion your complaint seems excessive. The article you cite a paragraph is no worse than many other ones published on Significance. “zero serious Republican contenders” is a personal (questionable but rightful) interpretation of the linked source (“The Republican field is taking shape, but right now, there is no one that excites me,” said Gov. Nikki Haley of South Carolina; Jim Jordan, a Democratic strategist, said the GOP field was remarkable for having “no center of gravity, no front-runner, no establishment choice, and no premium at all being put on electability.”). Similarly for Bin Laden as “the *single* most pernicious foe to ever threaten American security”. I want to point out the “single” attribute: there would had been Confederation even without Davis, Japan even without Tojo, and CCCP even without Stalin, while the identification between Bin Laden with Al Qaeda is more tight (and it seems to me that US government did contribute to this feeling).
    On the other hand, on monkey cage blog I read sometimes some terrible and mystifying articles about politic situations in other countries; even if I cannot complain very much because the blog is not associated with any official society (I suppose).
    Maybe the article about Russian elections was not as persuasive as someone could expect, but surely it is more reasonable than many other ones which claim to prove fraud with farcical arguments (even if not published on Significance). So I’m inclined to understand its eagerness, and I would like to read more articles similar to this one.

    • Andrew says:


      If you see anything on the Monkey Cage that seems terrible, please explain why in the comments. I don’t agree with everything on that blog!

      Regarding the articles in Significance: the one on U.S. elections was just empty, sub-op-ed speculation with zero statistical or political content. The article on Russian elections bothered me by its tone, which seemed more like a lawyer trying to argue a case rather than a scientist trying to uncover the truth. In both cases, as a lifetime member of the American Statistical Association, I’m unhappy with our leading professional organization being represented in that way. (I’m also unhappy that Ed Wegman has an award from the ASA.)

  4. JL says:

    Perhaps interestingly, Simkin has also published a similar article at the far-right site Alternative Right:

  5. Bobito says:

    Another possibility, that folks in the US seem unable to even contemplate, is that US elections are (perhaps for different reasons) as bad as Russian elections.

    At any rate, it is not a hypothesis that should be dismissed a priori.

  6. bbzippo says:

    Found this page via Google. Was expecting to see any statistical arguments, but see none. Disappointed.
    Thought you might be interested to see my take on this:

    I’d like to emphasize that the correlation of turnout and party support, which is relied upon as the main indicator and measure of fraud by Ordeshook, Myagkov and the amateure Russian bloggers, is observed in many honest elections. See e.g. A Test For Conformity In Voting Behavior by Stephen Coleman :

  7. […] This reminded me of a series on blog posts Andrew Gelman ran a few years ago on using statistics to detect elections fraud (e.g. here, here, and more recently here). […]