Angry Commentators Get the Discussion of Minnesota Ballot Recount Wrong

My piece at Fox News yesterday on the inconsistent rules for counting ballots in Minnesota's Senate race has produced a lot of reaction. Nate Silver over at fivethirtyeight claims in the title of his post that my piece "blames liberal conspiracy." My piece didn't mention the words "liberal" or "conspiracy" (nor did it say that Democrats or Franken were involved in some secret conspiracy or use any similar synonyms). Nor did I say that people had made mistakes intentionally (unlike Nate, I don't claim to know what is going on in people's minds).

My point was simple: "The primary problem isn’t the rules. The real problem is the lack of consistency." Mistakes can be made simply because someone is dealing with so many ballots, but mistakes, even if they are not done on purpose, can effect the outcome of this race because it is so close. Silver doesn't address this concern at all in his write up. Instead, he focuses on the fact that the Star Tribune data base made a mistake in classifying one of the ballots and that I relied on that for one of my examples. The fact that my website had already noted this and corrected that one ballot example before his post was put up is never mentioned by Silver. He also incorrectly implied that I hadn't double checked that ballot, but he didn't make any telephone calls or check this point either. It is not surprising that he wouldn't try to check these points out before making his assertions. For those interested, you can see my original post on this website here.

Others have made similar posts. David Brauer with minnpost also focuses on the one ballot misclassified by the Star Tribune and ignores the issue of rules not being applied consistently. He also incorrectly asserts that I did not check on what the newspaper had, but, ironically, he did not bother to check and contact me before he made this claim.

Others have also latched on to me claiming conspiracy. Emily Kaiser at Citypages writes: "Will he still claim liberal conspiracy despite losing his best evidence? Of course." Again, Kaiser also repeated the claim that no additional checking had been done. Silver's claim about a conspiracy is just being repeated. Why is it so difficult to evaluate whether a mistake was made without having to declare that someone did it on purpose? Why don't these blogs look at the systematic errors made by the Canvassing Board?

The Huffington Post has gotten involved in the act and claimed that I was "Blam[ing] Liberal Conspiracy." For those interested in the sock puppet and identity theft history of those who run the Huffington Post please see this post here. The so-called "For the Sake of 'Science'" also talks about me "trying to blame a typo on a vast liberal conspiracy."

All this is actually pretty funny. One person makes a post after I have already pointed out the problem on this website. He obviously misstates what I wrote and uses terms such as "conspiracy" to inflame readers. All these other liberal bloggers repeat these claims without probably even reading what I wrote (it would be worse if they read my piece and still repeated what was in Nate Silver's post).

Again, the original posting and the correction can be seen here.

Update: The MN Secretary of State's Office has told me that they made multiple mistakes in their listing of ballot decisions. I should also point out that before I published the piece being discussed in this posting I had contacted the MN SOS office a couple of times to double check things but was unable to get any assistance from them.



Blogger Nathan Cook said...


I still have a couple of problems with your position here.

First, you acknowledge the fact that the whole ballot must be looked at to determine intent in some cases, which is good, as your article didn't make that clear at all. But your idea of how the canvassing board applies state law isn't quite right. There is a rule (whether statutory or S.O.P., I don't know) that crosses, crossed ovals, ticks etc. are counted for intent if used consistently over the whole ballot. There is no rule saying that in all other cases those marks have some other meaning. The board, as it encountered new circumstances, determined what standard it would apply in future cases. One of those rules was that a consecutive sequence of races in which the only mark made was a crossed oval against one candidate, followed by a sequence of races with uncrossed ovals, is to be counted as if the crosses were absent. You may think that's a nonsensical ruling, but the board was not abusing its power by making it, and I've seen no evidence that it was not applied equally to both candidates.

Secondly, if, as you just stated, the point of the article was to indicate inconsistency, rather than bias or collusion, why expose yourself to the charge by only pointing out decisions favourable to Franken?

Whether or not you believe bias exists, I doubt you'll make the accusation, even with the proviso that it is unconscious bias rather than fraudulent. Perhaps that is a matter best left alone. On the charitable assumption that there is no bias, and taking up your assumption that there is a lack of consistency, it follows that if the canvassing board swings the election to the wrong guy, it's because of variance. I think a statistical analysis on that basis would be interesting.

12/23/2008 6:38 PM  
Blogger John Lott said...

Thank you for your comment.

1) The point is not whether I like the rules, the point is that they were definitely not applied consistently.

2) Because there were many more mistakes in Franken's favor than the reverse.

3) I don't know whether there is explicit bias. All I can say is that the ballots are that the rules are not applied consistently, and that the different implementation of these rules favors Franken.

12/23/2008 6:56 PM  
Blogger Nathan Cook said...

Sticking with the 1 2 3:

1) I recognise the point you are making. The pattern of marks I referred to is present in all of the Franken crossed-oval votes you linked bar one (the second one). I surmised that you were unaware of a ruling reported only in an uncommentated webcast, and didn't spot the pattern yourself because of its odd nature. I didn't mean to suggest you were allowing your opinion of the ruling to influence your calls. My point is that there is only one inconsistent ballot in your list of crossed-oval Franken marks. Admittedly that one ballot is clearly not kosher, and if I find that the ruling in this case was correctly recorded, you'll get no complaints from me if you wave it around on Fox.

2) You have not proven this. This ballot was judged for Coleman despite there being only weak evidence of a cross in Barkley's oval. This one resembles your between-the-lines example in that it touches Coleman's oval and has a small intersection with and descending line into Franken's. In this case, however, the voter has added a small arrowhead to the descender, pointing from his scrawled oval into Franken's. Why he did this is beyond me, but the intent is probably to vote for Franken. Yet the board judged the ballot unclear. These two examples, which you did not list as favourable to Coleman, were not hard to find. Either you used a small sample size when looking at the board's judgements, in which case you do not have evidence that more putative mistakes were favourable to Franken than to Coleman, or your sampling methodology was incorrect, which does the same damage to your argument. I don't fault you for this - your article was time-sensitive, and applying peer-review standards to an online opinion piece would be ludicrous. Nevertheless, your claim is unsupported.

3) There is certainly some inconsistency. The between-the-lines ballots demonstrate that. How much inconsistency is a very interesting question. Congratulations, you've semi-converted me! I do not believe the inconsistency led to bias, based on my reasoning above, and the integrity I saw the board members display while watching the webcast I mentioned. I am 99.9% convinced that no explicit bias, i.e. fraud, occurred in favour of Franken. If it had, the ballot with an arrowhead attached to its descender would have been accepted for Franken along with your very similar but less clear example ballot.

Sorry for taking up so much real estate. Hopefully this comment is in any case more of what you are looking for than not. What needs to happen now is for an estimate to be made of how much variance has been introduced by inconsistent rulings. It may be that if the final score is very close, there will be so much irreducible uncertainty that a re-vote becomes attractive.

12/23/2008 9:37 PM  
Blogger Me, not you said...

Some of your problem comes from the Star Tribune LINKING THE WORNG BALLOTS! It may not be the government making the mistake John, it might be you and the Star Tribune. That is why you need to leave this very well analyzed recount to people actually IN MINNESOTA!

"Meanwhile, I believe I can explain the exact nature of the STrib's link errors---indeed it seems two links have shifted since he (last) posted the Fox piece. The URL of the first STrib vote review link for a ballot in his Fox piece ends in


This is a database-query link, written in the scripting language called PHP which supports the so-called SQL queries. The query does not identify the ballot by name or by any of its instrinsic labels or characteristics---as was the case with their ballot links when they first set up their Ballot Challenge. Rather it is identified "implicitly" by the review_date and an index number. What I infer is that when 1 or 2 challenges were added or withdrawn, a change was made to the underlying database, but the URL did not know how to track the ballot after the change. This sort of thing is what I meant by an "off-by-1" error, and indeed at least 1 bad link is fixable by subtracting 1 from the index in the URL, jumping you from an incorrect Le Sueur County ballot to the correct Dakota County one. (I won't link it---the links may change!)"

12/23/2008 11:21 PM  
Blogger Martin G. Schalz said...

Dear Mr. Cook. Where, praytell, did you come up with the concept that Dr. Lott made any statement to the effect of "the fact that the whole ballot must be looked at to determine intent in some cases, which is good, as your article didn't make that clear at all."

The whole ballot, Mr. Cook? Where does any statement referring to this, exist within the Fox News Opinion that you refer to? Please be so kind to show the text to which you base your argument upon, Sir.

Dr. Lott, did not refer to state law in any way that I can see, so I fail to see how your argument of "But your idea of how the canvassing board applies state law isn't quite right. There is a rule (whether statutory or S.O.P., I don't know) that crosses, crossed ovals, ticks etc. are counted for intent if used consistently over the whole ballot. There is no rule saying that in all other cases those marks have some other meaning."

There is a rule, whether it be statutory or Standard Operating Procedure, and you do not know? This is what you base one segment of your argument upon?

Dr. Lott clearly points out inconsistency of a process. That is all. To 'see' anything beyond that, based upon the entirety of your argument, would be to place the bias at your doorstep, Mr. Cook. After all, there are many who can recognise the lack of a clear analysis of a problem without bias.

12/23/2008 11:58 PM  
Anonymous Anonymous said...

Mr. Schalz, what an extraordinary response! I'll take your points backwards. When you place quotes around the word see, you intend either to quote me, or mock the idea that what you see and what I 'see' are similar. If the former, you are being too elliptical for me to reply sensibly, although in any case you are too elliptical here.

The rule is statutory. It deals with a particular circumstance and is silent on other circumstances, and I may knew that when I wrote my comment. Note the absence in the statutes of "if marks are not all consistent, those which are not filled ovals are to be discarded."

Your third paragraph is missing a crucial verb phrase. Are you attempting to lead me up the garden path, Mr. Schaltz?

No explicit reference to looking at the whole ballot is present in the op-ed, except for the last sentence in the fifth paragraph. Therefore that concept is not as clear as it might be. Quoting an absence is paradoxical (although I seem to have managed it in this comment). Now, it so happens that in comments on another post on this blog, Dr. Lott says "even if some [ballots] include an X in the presidential race, the vast majority of the races do not have Xs." This indicates that he is looking for voter intent in the whole of the ballot, albeit with what I believe are incorrect views on what the rules are for this. I gave the rule which governs many of the ballots he uses. It is non statutory, ad hoc, and may well be de novo. Returning to your last paragraph, it might be said that when the process is not agreed on, the inconsistencies are well named.

12/24/2008 8:26 AM  
Blogger Martin G. Schalz said...

"Mr. Schalz, what an extraordinary response! I'll take your points backwards. When you place quotes around the word see, you intend either to quote me, or mock the idea that what you see and what I 'see' are similar. If the former, you are being too elliptical for me to reply sensibly, although in any case you are too elliptical here."

Yes Mr. Cook. The use of quotation marks does indicate that I am refering to what you posted in it's context. Nothing more, nor less Sir. To mock you would require that I treat you with contempt or ridicule, which did not take place. Using your own words as a foundation for an argument against the point that you were trying to make, does not in and of itself rise to the level of ridicule.

To be obscure, would mean that I am trying to hide something, and if that be the case, you obviously have yet to find it. After all, I merely used what you posted, against you.

"Your third paragraph is missing a crucial verb phrase. Are you attempting to lead me up the garden path, Mr. Schaltz?"

Explain please. I did not leave out a crucial verb phrase, Sir. I submit that I occasionaly foul up a sentence by means of misplaced punctuation, or an syntax problem.

As far as the Garden Path goes, I have enough fertilizer, thank you.

To read into Dr. Lott's words, and to apply an interpretation that is wholly wrong, is an abomination, and a willful attempt to mislead.

Prove to us all here that you do posess some form of a moral compass, and stick to what was actually said, rather than twist it to conform to your own version of what you would like it to be.

Extracting concepts that do not exist in a man's statement, is simply a deliberate falsehood that is meant to obfuscate the original text. Those who engage in such behaviour, do so for their own gains, at the expense of others.

12/24/2008 5:34 PM  
Blogger Nathan Cook said...

Mr. Schalz, firstly please note that my last comment was edited slightly. Look to the left of my name: unless the orange B symbol is present, I cannot claim full ownership (technical details on request). Please look up 'garden path sentences' if you want to know what I was talking about.

I did not accuse you of being wilfully obscure, Mr. Martin G. Schalz, but rather of being too brief. The verb 'to see' occurs in "I've seen no evidence that it [a rule] was not applied equally." I refer here not to the rule Lott mentioned, but the one I introduced to the discussion. By quoting a single, common word, you made yourself difficult to understand, and so I thought you might have been using 'scare quotes', which do not identify a quotation. That is the mocking I was referring to. I wonder how my quoted remark indicates my bias. Is it false?

You request an explanation of a missing verb phrase. I would not have pointed out your error if it did not hide your meaning.

'... I fail to see how your argument of "But your idea [...] some other meaning."' You need something more at the end, e.g. 'is applicable to what he wrote.' I could tell that you disagreed with my quote, and thought Dr. Lott did not refer to state law, but you did not specify the disagreement. I think I can deduce what you intended from your second comment; my reply follows.

You seem to believe that because Dr. Lott did not describe the actions of the Board by referring to state law, he could not have been talking about how the canvassing Board applies state law. No - from beginning to end, the article deals with that subject. The board was "determining voter intent", as provided for by 204C.22 subd. 1. The "claimed exception" is 204C.22 subd. 10. The "inconsistencies" he believes exist are, if present, examples of the Board applying the rules, both statutory and other, correctly in some circumstances and incorrectly in others. I'm not digging hidden meanings out of his words by pointing this out. How can I be said to be 'extracting' concepts, e.g. the ad hoc rule, that aren't present in his article, when I don't use his article as the source of those concepts? In case it is not clear from my second comment, I know of the rule because I watched a webcast of the recount, during which it was formulated. Someone who didn't watch that webcast would probably not know of it. Numerous such rules exist.

12/25/2008 1:02 AM  
Anonymous Anonymous said...

Mr. Cook, I don't really see what you are saying.
Take the examples provided by Lott. Look at this ballot that went for Franken. There are no other X'ed votes, only the one for Franken, and yet Franken picks up this vote.
Look at this ballot that went for Franken. There is only one other X'ed vote, and yet Franken picks up this vote.

I can't find anything similar for Coleman that counted for him. In fact, the ones that Lott shows are very telling.


12/26/2008 1:01 PM  
Blogger John Lott said...

Comment section closed.

12/27/2008 8:51 AM  
Blogger Nathan Cook said...


12/27/2008 12:39 PM  

Post a Comment

<< Home