VANITY - Analysis of Minnesota Recount

11/29/2008 | vanity

[ex-Libertarian]

The recount is around 88% complete with Coleman ahead in the official tally by 282 votes. At the same time, there are 5,623 ballots which have been challenged by one or the other campaign, and these ballots are removed - temporarily - from the official tally. Therefore the 282 vote lead is not very meaningful; what matters is the dispensation of the challenged ballots. For example, if these challenged ballots break 52.6% to 47.4% in favor of Franken, he wins.

Note also that the vast majority of challenged ballots will go against the candidate making the challenge. That's because most of these ballots are ballots which otherwise would have been decided against that candidate, which is why they challenge it. The corollary to this is that the more challenges a candidate has made, the more votes you expect his opponent to gain once those challenged ballots are resolved and included in the counts. So making lots of challenges is a bad sign for a candidate, since those challenges will tend to go against him.

Since Coleman to date has made most (51.3%) of the challenges, this is a cause for concern. However, the final key factor in attempting to predict the outcome is an understanding of the QUALITY of the challenges. In other words, if one candidate's challenges are materially more frivolous than those of the other candidate, we can expect a higher percentage of that candidates challenges to go for the opponent than those challenges made by the other candidate.

Fortunately, the Star Tribune web site has a feature that lets you view the challenged ballots and "vote" on how each ballot should be decided. I took the time to sample 116 of those ballots and reviewed them according to the stated standards of Minnesota (i.e., clear intent of the voter; no identifying marks). I'd also appreciate feedback from any statistics experts as to whether this sample size (just over 2%) is sufficient.

Here's what I found:

* of the 60 Coleman challenges I reviewed, 71.7% should count for Franken, 5.0% for Coleman and 23.3% for neither

* of the 56 Franken challenges I reviewed, 78.6% should go for Coleman, 7.1% for Franken, and 14.3% for neither

The conclusion from the above is that there is a significant difference between the quality of the Franken Challenges vs. the Coleman challenges. Specifically, the Franken challenges are more likely to be frivolous challenges of clear Coleman ballots, while Coleman challanges include a higher percentage of true questionable ballots which will ultimately be decided for neither candidate.

Don't get me wrong - both side are making alot of frivolous challenges. It's just that Franken's are slightly more weighted towards the frivolous.

Although these numbers seem pretty close, with a race this close they are very signficant. Despite the fact that Coleman is making more challenges, the actual challenges in my analysis break slightly for Coleman (40.8%/40.2%/18.9%). If this sample holds up over the universe of ballots challenged to date, Coleman will actually pick up votes from the challenge process, despite the fact that he has made more challenges. Specifically, these rates would yield a Coleman lead of 315 votes versus the current 282 once all the current challenges are resolved.

Now, this does not take into account any changes that may occur from the recount of the remaining 12% of votes which have yet to be recounted. But since Coleman will have seen a significant net gain from the recount so far, there is no reason to believe that the final 12% will yield a significantly different result from the first 88%. Especially since the remaining uncounted votes are around 51.5% in favor of Coleman and are mainly in Republican counties.

Of course there still remains the X factors of court challenges and potential challenges by Harry Reid and Senate Democrats, but the stronger Coleman is at the end of this recount, the harder it is for those challenges to gain traction. A 300+ vote lead at the end of the recount would be very difficult politically to overturn.

So overall I'd rather hold Coleman's cards than Franken's right now.

[centurion316]

Interesting analysis, a good “Vanity” IMO. Several of your paragraphs were truncated, a phenonmenom that has been frequent of late on FR, something to do with the posting software, I suspect.

[YdontUleaveLibs]

I don’t live in Minnesota anymore, but I still follow the politics there more closely than in Tennessee, old habits die hard.

[TenthAmendmentChampion]

Great work. I pray Coleman defeats Franken. Does anyone know how much Franken spent on his campaign?

[Norman Bates]

“Since Coleman to date has made most (51.3%) of the challenges, this is a cause for concern.”

A cause for concern for Franken perhaps.

[MainFrame65]

I can’t find the post, but earlier someone had theorized that the Frankenvolk had begun making frivolous challenges for the specific purpose of removing opposing ballots from the recount to create a lead, however temporary and tenuous, during the recount process in order to build their case for a court challenge. Coleman’s team countered that by using the same tactic, until they came to a tacit agreement that no Franken advantage could be gained that way.

[tickles]

While I understand that the US Congress can have input into the selection of a President if there are problems in the election and there is no clear winner, is there really anything in the Constitution that allows the Senate to challenge a state election? I thought that the state was the one to choose it’s own representatives and senators, not those in DC.

[goldstategop]

The Senate can decide the qualifications of its own members. In practice, its rare for the Senate to refuse to seat a person legitimately elected by the voters unless they were convicted of a felony. The Senate would have been sure to deny Ted Stevens his seat if he had won and Republicans would have backed the decision. Not so so in the Minnesota case in which the only reason to refuse to seat Norman Coleman would be raw partisan considerations.

"Show me just what Mohammed brought that was new, and there you will find things only evil and inhuman, such as his command to spread by the sword the faith he preached." - Manuel II Palelologus

[Krankor]

I look at it another way. I have a feeling almost all of the challenges are frivolous and will be disallowed so, at the end of the day, the challenged votes will just be added to the totals of each candidate. Since, Coleman has challenged 147 more ballots than Franken, you subtract 147 from Coleman’s lead of 282 and get a real lead of 135 votes.

[ex-Libertarian]

That’s a decent “back of the envelope” way to look at it, but that method assumes that the challenges from both sides are equally frivolous. As per the above analysis, I think that is not a valid assumption. With an election as tight as this one, a small difference in the “frivolity factor” makes a significant difference.

[J40000]

Franken(on the floor kicking and screaming): “It’s NOT FAIR!!! He cheated BETTER than I DID!!! It’s NOT FAIR!!!”

[Krankor]

Believe me, I hope you are right. You gotta give Franken credit for a pretty good strategy- frivolous challenges in order to build up a big lead in the recount so as to plant the idea in the public’s mind that Franken was the winner and therefore build public opinion for any court challenges or an intervention by Harry Reid. Of course, Coleman was sharp enough to catch it and respond in kind. I think Coleman is trying to stay safely ahead in the challenge department just to fend off any last minute barrage by Franken.

[jmaroneps37]

Analytic work like THIS is why FR is such a great place to go for the truth. Thank you for this post.

[pistolpetestoys]

I totally expect Reid to refuse to seat Coleman.These are Dhims-laws that are not to their advantage are to be ignored.Who’s going to stop them?