Skip to Content

Support MinnPost

Recount: How Lott misleads

Conservative stat maven John R. Lott got a lot of traction early in the recount when he pronounced Al Franken's early gains evidence of fraud. It turned out they were three simple tabulating errors that Norm Coleman's camp doesn't dispute.

Lott is back and bigger than ever. In this blog post, he presents evidence of state Canvassing Board "mischief" that doesn't tell the whole truth — or in some cases, any truth.

First, he presents four ballots clearly marked for Coleman, noting, "Here is an example where the Minnesota Canvassing Board claims the vote is for no one."

Outrageous? Nope. What you can't see and what Lott doesn't tell you (unless you click through to the Star Tribune's website and look at an enlarged image) is that all four ballots have clearly identifying marks, such as signatures. By state law, they have to be tossed — but Lott doesn't tell you that, either.

Then comes the just plain wrong part. The next two ballots appear to be marked for Coleman; Lott writes the Canvassing Board judged them "clearly for Franken."

Begin the Brooks Brothers riot!

But no. A cross-check of the Secretary of State's extremely helpful spreadsheet (Excel file) indicates neither ballot was called for Franken. The first was tossed for — you guessed it — an identifying mark, and the second was ruled an undervote (for no one) because the Coleman oval was that voter's only cross-out.

To be fairer to Lott than he is to the Canvassing Board, he presents a couple of ballots that went Coleman's way. But that's a fig leaf for his charge of "a systematic bias by the board in favor of Franken."

Lott does a little CYA by calling his post "a rough glance" and pledges to examine things "more scientifically" —cherrypicking, anyone? But you know the game: any moment now, Fox or Ann Coulter, or some other retransmitter will link to this (I found it on Twitter), and the cycle of misinformation will rev up again.

Just because it's typical doesn't make Lott's baseless sliming of Minnesota's efforts any less disgusting.

Get MinnPost's top stories in your inbox

Comments (21)

Here's a bizarre idea: maybe Lott's not sliming anybody, but is just, possibly, wrong about something. A slimer perhaps would not have approved a comment on his website that was critical of one of his calls -- but Lott promptly approved one of mine that was just that.

You might want to think about that possibility.

Full disclosure: I know John Lott slightly, have talked with him on the phone, briefly, on three occasions, have had dinner with him once, and admire him greatly, particularly at the way he's handled himself when some people have slimed him.

Joel - I'm sure John Lott is a lovely fellow, but the guy throws out words like "fraud" and "systematic bias" at a moment's notice, based on sheer speculation the first time and evasions/inaccuracies the second.

In other words, he's wrong AND loud about it. (Systematic bias indeed.) Whatever his intent, the result is sliming the reputations of people who don't deserve it.

Lott does have a certain reputation for pushing data where it doesn't want to go.
He appears to start with a conclusion, rather than concluding with it.

Lott sounds like a right-wing flack. He has as much credibility as Sean Hannity or Ann Coulter. He ranks with them at the bottom of the barrel.

I've met John Lott myself and have admired much of his past work. However, in this instance, he could not be more wrong. This has been an upfront and transparent process. As of now, I am confident that at the end of the day we will know you won this election.

It's the sort of thing that makes me go all Elvis Costello. When somebody you generally agree with gets something wrong, it's either not worth mentioning or something like charming crotchetiness; when somebody on the other political side does, it's just more proof of their rampant immorality.

I wish I thought the obsolesence of the local print MSM was because of that sort of groupthink, as it could be reversed. Alas, I think that Craigslist has done far more to sink the Strib, CP (yeah, they've become MSM) and such than, say, Powerline catching them doing their job so badly.

I would not put Lott in the same bucket as Coulter.
He does have some real chops as a statistician, even if he lets his biases distort them more than he should.
Coulter, OTOH, is a certified dingbat whose only talent is invective.

I think John Lott has been PROVEN wrong more often than most academic researchers, and in being proven wrong has frequently been shown up as a partisan hack who routinely manipulates data in support of positions not born out by any real facts.

For my money no one can touch him when it comes to transgender comment posting. Is Mary Rosh still leaving comments around the blogosphere, or did Lott retire his alter-ego after everyone caught on to his cheap trick for applauding himself from the comments?

John Lott's sins against the truth are well worth mentioning because they're not tabulation errors, they're ideological manipulations up to and including falsification of data used in academic studies. But don't take my word for it, just google ["john lott" "mary rosh"] and you'll have all the evidence you could ever want.

Your comment critical of this piece was approved, wasn't it?

Yeah, I see your point...

One of Lott's recent posts (see here: ) features speculation that a run-of-the-mill cell phone number harvesting scam is in fact a Soros-funded effort to subliminally boost support for Obama.

I suppose it's possible. A more likely possibility: John Lott has a habit of seeing sinister motives where there are none.

People can have different sides, and appear quite charming sometimes and quite the opposite others. Coulter might never be anything but grating, but look at Bill O'Reilly. The man can be seem erudite and likable, and then become a bully in an instant. Maybe, if Lott is is sometimes wrong, always wrong in a partisanly useful way, and doesn't retract, then he might not be a basically good guy. Maybe during the times he wasn't caught getting his facts wrong, he wasn't so much right as not caught. Like someone I[m not sure who said, a person who is nice to you and rude to the waiter is not a nice person. Lott might be nice when you're having dinner with him, but if he tells half-truths repeatedly, then his is dishonest.

Comments here ignore the most important point made in the original post, which is not about Lott's sublime civility or his mere susceptibility to human error. By publishing wrong data and characterizing it in inflammatory terms, Lott guarantees his error will be repeated endlessly as truth in the neo-con media echo chamber and therefore has an excellent chance of being reported by main stream media as credible. That's how it works. Make up stuff, get it to a Fox talking head, purr to Cokie Roberts or Brian Wilson, and you've moved the debate, clouded the issue, and poisoned the well again. That's how our country is governed.

First, let me note that I have removed a couple of the questionable ballot calls from my list (those involving whether identifying marks were made on the ballots). These points weren't always clear to me (whatever the board claims was the note "TO" an identifying mark?), but they aren't worth debating and apparently distracted from the other problems. However, I appreciate the discussion on those examples.

As to the content, take the first example. The explanation on the decision to count the ballot for Franken and not what appears to be a clear vote for Coleman is "No Dup." I assume that this means that there is no duplicate ballot, but how does this cause a clearly marked vote for Coleman to be counted for Franken? I admit on this one I am not completely sure what is going on, but it sure seems like a questionable decision.

More important empirically in terms of the large number of votes involved, I understand the notion of not counting the "X's" through the vote marks, but the examples that I provide indicate that this rule is clearly not applied equally. I provide plenty of those examples and there are many more that could be provided. David dismisses the point but fails to respond to the fact that similar situations for Franken resulted in different conclusions.

As far as I can tell, no discussion is provided on all these other examples. If there is something that I should know, please provide it.

Finally, I don't think that David's discussion here has been particularly accurate for other reasons. He writes: "But that's a fig leaf for his charge of 'a systematic bias by the board in favor of Franken,' " leaving out the qualifier that it "seems to show." Besides point out examples on both sides, which David backhandedly acknowledges, there were other qualifiers both in the write up and the title that he ignores.

Also, no absolutely no evidence is provided for his claim of me selectively reporting examples.

Not only does he exaggerate the statements that he discusses. In the comment section David says that " the guy throws out words like 'fraud'," but the discussion of "fraud" was brought up in a discussion on ACORN. If he wants to defend ACORN's actions as not being fraudulent, he should do so directly. I did use the term "systematic bias," but especially with the qualifiers that I used it is hardly an inflammatory term similar to the terms used by David.

John Lott is the one being slimed? This is a guy who could not produce any of the data for his gun study for which he is best known, claiming there was a computer crash. This is a guy who blogs under fake names praising the work of John Lott.

This is not an innocent mistake. John Lott is a serial liar. Joel, did you check to see if there was any money missing from your wallet after you had dinner with Lott?

A few specific responses to John:

1. I'm glad he's changed examples, given the low quality of the originals. But NO ballots where Coleman's oval was the only one marked was counted for Franken. Period. I understand he's a bit bewildered by the SoS mark, but there's no mistaking the spreadsheet column of who the vote went for - not Franken.

2. His defense about the Nov. 10 FoxNews piece is too clever by half. The headline (Minnesota Ripe for Election Fraud) and first eight grafs all discussed the recount. ACORN only made an appearance in the 16th of 17th graf, when John wrote, "Minnesota was facing vote fraud problems even before the election." It's obvious his point was they were facing it AFTER the election, too.

3. It's true, I did stop after reviewing six ballots - there are only so many hours in a day. But to recap, the first four ballots he claimed were Coleman votes being thrown out for no reason WERE thrown out for a reason (identifying marks), and the next two were listed as Coleman votes for Franken, which was factually wrong. I have not reviewed his point about inconsistency - I may circle back, though am not sure if his examples are the same. However, the first six assertions had nothing to do with inconsistency. They were objectively wrong or misleading.

4. I will grant John that I was less than charitable about his qualifiers. I would be more charitable about this had this been an isolated example, but it is not.

Dear David:

So are we to assume that you agree with the examples that are reported on my website or that you are just not going to discuss them?

You are in the news media business. You know that writers don't pick the titles for their pieces, that editors do that, right? the firs eight paragraphs discussed the recount, but I did not use the term "fraud" in that discussion. The term "fraud" was used when I got down to the discussion of ACORN. This doesn't seem like a very honest response by you. Do you believe that ACORN hasn't engaged in fraudulent behavior? If you do, what is inaccurate with what I wrote.

You are selectively quoting parts of my discussion is disappointing. Your response is also disappointing that it is justifiable for you to do so because you claim that I have supposedly made inaccurate statements about something else.

Will you provide some evidence to back up your claim that I am providing selective examples?

Dear Dan:

You might want to check out the inaccuracy of your claims before you repeat them. More than almost any academic I have long made all my data available to others. All the data for my research on guns is available to other academics. For example, see

Well John, its good to see that you are defending yourself instead of having "Mary Rosh" do it for you. By the way, are you still Mary Rosh's favorite professor?

So this (obviously reported elsewhere as well) is where I got the idea that you haven't been able to produce your data:

"The major research on defensive gun use, Duncan objected, had shown firing rates ranging from 21 percent to over 60 percent. Lott replied that "national surveys" actually referred to his own heretofore unknown survey of 2,424 households. When Duncan pressed him for the survey data, Lott demurred, saying a hard drive crash had destroyed his data set and the original tally sheets had been lost. In fact, there seemed to be no record at all of the study, nor could Lott recall the names of any of the students who he said had worked on it. Some people began to suspect the study, which is tangential to Lott's conclusions in More Guns, didn't exist."

So have you, in fact, found the survey data for this? Is that available on your website?

Dear Dan:

You may not care, but there are many errors in what you post.

1) During the entire 1990s there were a total of four surveys that broke down exactly how guns are used defensively, two of which were mine. Kleck and Gertz’s estimate is 92 percent when brandishing and warning shots are added together, and since I didn't have any examples of warning shots in my data it is the closest to their claim.
2) My discussion was that media are much more likely to cover defensive gun uses that result in the death or at least the wounding of the criminal. The point of the discussion was that the higher the rate that guns are either brandished or a warning shot was fired without harming the attacker, the easier it is to explain the lack of news coverage of defensive gun uses without resorting to media bias. If you believe that I picked a number that was too high and you believe that I think that the media doesn't cover defensive gun uses as much as it should, I would be biasing the argument against me.
3) The point of science is replication. I have replicated the survey, producing similar results, and the survey questions have long been available to others who what to examine the issue. The data are available here: As to the original data set, I have provided the names of 10 academics who have contemporaneous knowledge of the hard disk crash.
4) Again, there are many other errors in your short write up.

Unless I am completely blind, there is no distingushing mark on this ballot that was rejected for Coleman:

It wasn't even challenged on the basis of having a mark. So unless I am missing something, Lott is right on this one.


it wasn't challenged as an identifying mark but as an undervote.

the Coleman vote is the only one with a line drawn through it leading one to be believe they may have changed their mind.

John, I'm sorry, where were the "many errors" in my write up? As I pointed out, when people asked for your data, you couldn't produce it and claimed that your computer crashed and that you lost all records of the study.

Are you denying that you blogged under the name "Mary Rosh" and that "Mary Rosh" praised John Lott as her favorite professor? Any errors there?

Your claim that there were "many errors" in my post has about as much merit as your criticisms of the Minnesota recount.