The King Banaian Saga

If you've read my past several articles, the name King Banaian probably sounds vaguely familiar; if you haven't, prepare to be enlightened. I'll try and let the court documents do the talking, but I'll have to fill in some of the blanks.

Before any analysis occurs, I better provide some background on Professor King Banaian. Prof. Banaian is an economics professor at St. Cloud State University in Minnesota. He is also a contributor to the very conservative SCSU Scholars and co-hosted a radio show with Michael Brodkorb of King Banaian is not an unbiased observer, he is a Republican.

The saga all began on January 20th, as described within King Banaian's deposition; the document is dated January 27th, but was filed with the court on January 29th:


Q. Professor Banaian, my name is Dave Burman. I'm one of the lawyers for Al Franken, the contestee in the election contest that's going on.


Q. You are a social scientist, aren't you?

A. That's correct.


Q. Who was it that gave you your assignment on the 20th, or whenever in advance of the 20th you got it?

A. The attorneys from the Coleman campaign had contacted me, in particular, Mr. Trimble.

Q. And when did Mr. Trimble first contact you?

A. I would say noon on the 20th.


Q. And it came with an e-mail from Mr. Trimble?

A. Yes, that's correct.

Source: Affidavit of Richard D Snyder Regarding Motion in Limine via [PDF]

The Coleman campaign sent Prof. Banaian a series of spreadsheets containing county level data pertaining to rejected absentee ballots on January 20th; he later stated that he verified some of the data using a spreadsheet posted on the Minnesota Secretary of State's website. The exact contents of the SOS provided spreadsheet remain unclear, but there does not appear to be a publicly available rejected absentee ballots table. I have however placed an inquiry with the Secretary of State's office regarding the context of Prof. Banaian's testimony.

Despite Prof. Banaian's own classification as a social scientist, he also testified to his qualifications as a statistical analyst:

Q. Do you consider yourself an expert in statistical analysis?

A. I consider myself a practitioner of statistical analysis. Expert, I believe -- I have difficulty with that word.

Q. Why is that?

A. I have - I have taken several courses in statistics. There are, obviously, people who have degrees in statistics, Ph.D.s even in statistics. It's fair to say that they will call themselves experts.

I -- I call myself a long-time practitioner of -- of statistical analysis. I provide data analysis for the St. Cloud community, and I teach data analysis to my students.

Q. As I understand it, as long as I picked the right test of those offered by Excel, even somebody as unschooled as myself could have, basically, replicated what you've done in Exhibit 2?

A. If you picked the right test.

Q. And the binomial approach is about the most basic of statistical tests, is it not?

A. Am I being asked my opinion?

Q. Yes.

A. In my opinion, most people by default go to the normal distribution. The bell curve is the most well-known distribution in statistics.

Source: Affidavit of Richard D Snyder Regarding Motion in Limine via [PDF]

Prof. Banaian stated that he used a binomial distribution to determine potential anomalies in rejection rates; he also experimented with a Poisson distribution and a Chi-Square test. Prof. Banaian also asserted that each of these regression techniques "require homogeneity of variance;" he did not consider the "socioeconomic qualitative characteristics of people who fill out absentee ballots." In turn, his analysis is based entirely upon the deviations present within the given distribution.

Q. Then there's a column that says -- and I can't tell whether the heading reject too many is just for one column or two. Do you know offhand?

A. It is for two.

Q. So the yes, yes for Aitkin County means that Aitkin rejected too many compared to what you expected in a random distribution?

A. Yes, that's correct.

Q. And is the first column based on the 5 percent calculation and the second on the 1 percent?

A. Yes, that's correct.

Source: Affidavit of Richard D Snyder Regarding Motion in Limine via [PDF]

Based upon my own calculations, Aitkin's rejection rate places 4th among the 132 regions responsible for collecting, categorizing and counting absentee ballots; Hennepin County passes this responsibility onto each municipality. My table uses rejected absentee ballot data provided by each campaign within a multitude of Court filings. The Coleman campaign has identified 4,458 absentee ballots and their associated voters by name, while Franken has listed 792.

The exact data available to Prof. Banaian remains unknown; some of it was provided by the Coleman campaign, and some of it was obtained through the MN Secretary of State's website.

The rejection pattern of Aitkin County seem to be irregular, but this irregularity may or may not be relevant. If you take any set of data and apply any distribution model, you will get outliers; the purpose of a distribution function is to identify a certain number of mathematically significant outliers, not logically significant. The existence of outliers, in and of itself, is completely meaningless without an analysis of their context. Prof. Banaian would later offer a potential explanation for the outliers he identified:

Q. So at least using this data there were between 15 and 23 counties where there seemed to be differences that couldn't be explained by random variation?

A. That's correct.

Q. But we don't know what caused those?

A. Yes, that's correct.

Q. Might be that they had a disproportionate number of people who were blind or didn't see very well and didn't sign in the right place?

MR. RALPH: Objection, form.

Q. That's a possible explanation for those counties?

A. It's possible, yes.

Q. And nothing that you've done tells us whether or not the excess rejections were contrary to law or incorrect in any way?

A. No.

Source: Affidavit of Richard D Snyder Regarding Motion in Limine via [PDF]

Prof. Banaian essentially stated that he was unable to explain any correlation between the random variations he identified and any procedural or demographic information inherent to any of the counties he analyzed. It is very difficult to find context for comparison because applicable historic data is scarce. There is however some county level data available from the 2004 Washington State Gubernatorial statewide recount; King County, containing Seattle appears to have an improper re-rejection rate of about 66%, but King County is not representative of all counties in Washington, and certainly not Minnesota.

The Franken campaign took issue with Prof. Banaian's analysis by filing a motion on January 29th to suppress any potential testimony:


In Contestants' opening statement, they announced that they would not be calling a social scientist to testify to alleged statistical differences between counties. Now they have changed their mind and wish to call Prof. King Banaian, an economics professor and political commentator. They were right before, and wrong now. Contestee moves to exclude the testimony of Prof. Banaian.


Prof Banaian should be excluded for a number of reasons. First is timeliness. The proposed witness was first contacted by Contestants on January 20 (Banaian Dep. at 29, line 6) (attached to Snyder Affidavit). He did not reach "any type of conclusion" until January 22 (id. at 26, line 19), a day after the expert disclosure deadline imposed by the Court, and he was not made available for deposition until after trial had already started, despite his earlier availability. Jd. at 79, line 11, to 80, line 4. The court is within its discretion to exclude the witness for this reason alone.

Banaian has not qualified himself as an expert on the issue he presents. He admits that he is not necessarily an "expert" in statistics,

[The second excerpt contained within this article is then referenced as proof of the previous assertion.]


Similarly, his opinions are not on a subject matter requiring expert testimony. He opines on the rejection rates of absentee ballots in 87 counties. The rejection rates are known facts, ascertainable by a simple calculation. Expert testimony is not required on that factual issue, and the only reason Contestants want to have Prof. Banaian testify is to suggest that there is something behind the variations even though he does not have the expertise and has not even attempted to do the analysis to suggest that.

Source: Motion - Contestees Opposition to Contestants Motion in Limine to Exclude Testimony of King Banaian via [PDF]

The Coleman campaign then responded to Franken's above motion on February 10th, by addressing each point of opposition:


Contestants plan to call Professor King Banaian as an expert to testify that, based on the application of well-established statistical methods, many counties rejected more absentee ballots than expected from the statewide rate. This conclusion is not obvious on the face of the data, and expert testimony is appropriate to support it. Professor Banaian's proposed testimony will rule out, as a matter of statistical analysis, the possibility that the variation in rejection rates across counties was the result of pure chance. As such, his testimony will assist the Court as it considers the equal protection and fairness issue underlying its determination of which ballots are in fact legally cast votes.


Contestee disparages Professor Banaian's selection of the binomial distribution and chi-square tests, Mot. at 4, but "both the chi-square and binomial tests of statistical significance are recognized by courts as being valid means of assessing the representativeness of a sample." Reynolds v. Giuliani, 118 F. Supp.2d 352,374 (S.D.N.Y. 2000) (citing cases). The binomial and chi-square tests are routinely admitted in employment discrimination cases to show disparities in treatment of protected groups. See, e.g., id.; Barnes v. GenCorp Inc., 896 F.2d 1457, 1463, 1466 (6th Cir. 1990); Coates v. Johnson & Johnson, 756 F.2d 524, 536 n.11 (7th Cir. 1985); Lilly v. Harris-Teeter Supermarket, 720 F.2d 326,336 n.18 (4th Cir. 1983).

The binomial and chi-square tests would be helpful here for the same reasons they are helpful in employment discrimination cases:

When a plaintiff demonstrates a significant statistical disparity in the discharge rate, he or she has provided strong evidence that chance alone is not the cause of the discharge pattern. The statistics do not and cannot determine whether the more likely cause is the defendant's bias or a legitimate selection criterion.

Barnes, 896 F.2d at 1468. Employment statistics may thus establish a prima facie case, subject to rebuttal or proof by additional evidence. Id. at 1466; Lilly, 720 F.2d at 336 n.20. Here, Professor Banaian's proposed testimony would rule out chance as a possible explanation of the different rejection rates and lay a foundation for the Court to consider inconsistent application of the statutory standard as the cause.


Contestee argues that Professor Banaian should be disqualified for hesitating to proclaim himself an "expert in statistical analysis," Snyder Aff. at 37:2-3, but an expert's modesty is not a basis for excluding him. Although Professor Banaian does not have a degree in statistics, he testified that he has a Ph.D. in economics, teaches economics and data analysis to college students, frequently uses statistical analysis in his work, and considers himself a "long-time practitioner of statistical analysis." Id. at 37:4-17. Professor Banaian has been a professor of economics at St. Cloud State University for nearly 25 years, and he has numerous data-intensive publications. Affidavit of F. Matthew Ralph ("Ralph Aff.") Ex. 5.


(quoting Dennie v. Metropolitan Medical Ctr., 387 N.W.2d 401, 405 (Minn. 1986) ("The general rule in Minnesota is expert testimony should be suppressed for failure to make a timely disclosure of the expert's identity only where 'counsel's dereliction ... is inexcusable and results in disadvantage to his opponent.").

Contestee suffered no prejudice whatsoever. Contestee does not claim that Professor Banaian's deposition testimony differed prejudicially from his disclosed opinion, because it did not. Compare Ralph Aff. Ex. 5, with Snyder Aff. at 12:20-13:5. Contestee complains that he was obliged to depose Professor Banaian after trial commenced, but he does not explain how the timing of the deposition harmed him, especially when it occurred several days before Professor Banaian's trial testimony. In sum, Contestants' disclosure of Professor Banaian was timely, and Contestee has suffered no prejudice whatsoever to justify exclusion of Professor Banaian's trial testimony.

Source: Contestants' Memorandum of Law in Opposition to Motion in Limine to Exclude Testimony of Professor King Banaian via [PDF]

It's interesting to note that absentee voting in Minnesota is considered a privilege and not a right.

Franken then responded to Coleman's response on February 11th; the memorandum presents two main ideas. First, the Franken campaign asserts that rejection rates are irrelevant to the Election Contest Court's duty and secondly, they stated that the techniques used by Prof. Banaian are uneffective:


Contestee Al Franken respectfully submits that the Motion in Limine to Exclude Testimony of King Banaian ("Motion in Limine") should be granted. As his deposition transcript makes plain, Professor Banaian's proposed testimony is irrelevant to the only proper question before the Court: which candidate received the greatest number of votes that were "legally cast" in the 2008 U.S. Senate election. Evidence of alleged statistical differences in rejection rates between Minnesota counties is especially irrelevant in light of Professor Banaian's own admitted inability to offer any evidence, testimony or opinion on the underlying cause for the observed differences between the counties. As a result, this evidence is irrelevant under Minn. R. Evid 402 and 702. It would add nothing to this already over-burdened litigation except additional cost, delay, and confusion, and should also be excluded under Rule 403.


For starters, the task before the Court is to determine which of the two candidates received the highest number of votes legally cast in this election. The relative rejection rates between counties sheds no light on this subject. A rejected absentee ballot either was, or was not, properly rejected and the fact that it was or was not rejected by a county with a higher or lower rate of rejection that its neighboring county (or even as compared to all of the other counties) hardly makes that rejection any more or less wrongful or the vote more or less "legally cast." Rejection rates by county are, in short, simply irrelevant to the task before this Court.

Moreover, the use of relatively unsophisticated binomial distribution analysis and chi square tests are hardly the preferred means of statistical proof in discrimination cases. Instead, such rudimentary statistical tests are frequently rejected in favor of a regression analysis that can eliminate other causal explanations for observed differences - precisely the analysis that Professor Banaian admits that he cannot offer.

Source: Contestee's Reply Memorandum in Support of His Motion in Limine to Exclude Testimony of King Banaian via [PDF]

The King Banaian issue is next in the Election Contest Court's queue, but they may have already addressed it's resolution. We know the Election Contest Court canceled twelve of their nineteen categories last Friday, but they also asserted the accuracy of Minnesota's vote counting system; from the introduction of the memorandum attached to the Order Establishing Hearing on Motions:

The facts presented thus far do not show a wholesale disenfranchisement of absentee voters in the 2008 general election. Of the approximately 286,000 absentee ballots cast by voters, roughly 4,800 are at issue in this contest and remain unopened and uncounted. This is less than 2% of the absentee ballots cast in the general election. Further, the Court notes that, many of the voters whose absentee ballots remain unopened may have cast ballots that have already been counted either because they voted in person on Election Day or they submitted a subsequent absentee ballot that was opened and counted. These voters were not disenfranchised by the rejection of their absentee ballots.


All rejected absentee ballots were reviewed both on Election Day and through the Supreme Court review process. The Court is confident that although it may discover certain additional ballots that were legally cast under relevant law, there is no systemic problem of disenfranchisement in the state's election system, including in its absentee-balloting procedures.

Source: Order Following Hearing via [PDF]

Based upon the above expression of the Election Contest County, it will be interesting to see if the Coleman campaign continues to push for the testimony of Professor King Banaian.

These icons link to social bookmarking sites where readers can share and discover new web pages.
  • bodytext
  • Furl
  • NewsVine
  • Reddit
  • SphereIt
  • Technorati
  • YahooMyWeb
  • Ma.gnolia
  • StumbleUpon
Your Ad Here

5 Response(s) to The King Banaian Saga

2/16/2009 10:29:47 PM CT

Social scientists routinely use statistical methods to analyze data. Just because someone is a social scientist (anthropologist, economist, psychologist, etc) does not mean that they have no need for math. Statistics are essential for analyzing data, whether one studies physics or economic theory.
2/16/2009 11:28:57 PM CT

haha, i learned about chi square test, binomial and normal distribution in my stats class :D

anyways, thank you for posting these!!! thanks you so much =] very interesting read
2/17/2009 3:49:46 PM CT

He's already admitted that he has nothing to offer as to the underlying cause of rejection, it would seem that all that remains to ajudicate is whether they were accurately rejected, or not. Coleman's blowing smoke out of his ass, AGAIN>
2/18/2009 2:49:02 PM CT

I would NOT expect a normal distribution. Frankly, I'd try to impeach anyone using other than non-parametric tests since they _presume_ things about the data that is not placed into evidence.

The Franken folks should nail him on that... That this is NOT a bell curve.
Stephen Cleveland
2/20/2011 8:24:41 PM CT

I wasn't paying attention, um, did America win? I voted for him. King Bananaman was saying something about bellcurves and shit. The poijnt being, King Banaian is now cited in Speaker Boehner's letter that purports to be signed by 150 economists (we presume they are not biased, and this man is clearly a GOP sympathizer). So I ran King Bananaina's name in a search and hot damned if I didnt end up skimming through some shit here. How bout that? Well I guess I'll find somewhere around here on the internet to masturbate before I sleep. Ta ta my pretties.

Leave a Reply:

Name: (Defaults to Anonymous)
Type the characters you see in the image below:
(Word Verification)
Electoral College Projection Map
Senate Projection Map