~~ Evaluating Sources for the Casebook ~~

Printer Friendly Worksheet - Blank (PDF)

Printer Friendly Worksheet - Example (PDF)

Title of Source: Study of the 2004 Election Exit Poll Discrepancies
Author(s): Josh Mittledorf, Ph.D., et. al.
Type of Source: Online article, uscountvotes.org
D ate of Pub. February 9, 2005

Informative: Yes or No?
Imagine you are someone who knows very little about your topic. What will you learn by reading this particular article?

This article is definitely informative. Without getting into partisan politics and angry accusations, it examines unemotionally what I haven't heard discussed since election night-that there was a big difference between the "exit polls" and the official election results. I remember on election night, early in the evening, how the exit polls indicated that John Kerry was leading in the key battleground states, but when official tallies began to be reported, they didn't correspond to the exit poll information. I might have thought that was just acceptable statistical error, but I learned from this article that exit polling is known to be extremely reliable, and that it's been very accurate for the past 20 years. I learned that the concern is not that the exit polls differed from the official results-there's always a margin of error-but that they differed by such a wide margin, which is well above the standard, accustomed margin of error. It's the wide margin of error that concerns these authors. They analyzed the report which was issued to explain the error margin and found that it does not provide sufficient evidence (or any evidence really) to explain why this error occurred. The explanation given in the report is that more Kerry supporters responded to the exit polling, yet the actual data, as these authors show, indicates that the opposite may be true: more Bush supporters responded. So why did exit polls put Kerry ahead, and the official results left him behind-by a much wider margin that the standard margin of error? The authors maintain that to really answer this question satisfactorily we have to undertake a serious investigation of the data-not just the exit polling data, but the official election tallies at the precinct level. They want to create a database to properly analyze the election results.

I was a little taken aback by this article. I didn't realize that the exit poll discrepancies were such a big red flag, but now I realize that it's possible these discrepancies may indicate fraud or a defect in our voting system serious enough to have skewed the election results. If that's the case, then we need an investigation so we can discover how to fix that defect or prevent that fraud (if such exists) by the time the next election rolls around. If we do nothing, people may begin to lose faith in the integrity of our voting system, which is a very serious consequence, since voting is at the heart of our democracy.

Credible: Yes or No?
Explain whether or not you're confident of the source's credibility.

I'm confident this is a credible source for a number of reasons:

• The "Contributors and Supporters" listed at the end of the article are all professional experts, professors in the field of statistical analysis. They hold Ph.D.'s from prestigious universities all across the country: Temple University, University of Pennsylvania, University of Wisconsin, University of Utah, University of Notre Dame, Cornell University, Case Western University, and Southern Methodist University. Given their credentials, it's easy to trust their evaluation of the Mitofsky-Edison report, as well as the statistical evidence they themselves provide.
• The article was peer reviewed by a core group of statisticians at Uscountvotes.org as well as independent reviewers
• The article appears at "UsCountVotes.org," a seemingly non-partisan organization which "proposes to objectively investigate voting patterns through the creation of a database of precinct level election and demographic data for all states." This site states that its goal is "to develop analytical and statistical techniques capable of pinpointing probable errors in vote counts worthy of investigation, regardless of the parties involved." The fact that it's a non-partisan mission makes me less concerned about bias.

Current: Yes or No?
Do you think the text is recent enough to be of use? Do you have any reservations about its currency?

Yes, the source is very current-just published on February 9, 2005.

Persuasive: Yes or No?
Analyze the text's use of argumentative technique as a means of persuasion. What is the claim? What is the line of reasoning in support of the claim? What evidence is presented to validate the reasoning and support the claim? Are opposing views taken into consideration? Are opposing views effectively refuted?

I find this source extremely persuasive for a number of reasons:

• The claim is not the least bit hysterical; on the contrary, it's cool-headed and rational: the authors claim merely that given the range of peculiarities and the wide margin of error between exit poll results and official tallies, a full investigation is warranted. If the authors had claimed that Bush "stole the election" I might suspect them of being partisan.
• The logical line of reasoning is valid-airtight, in fact. The authors reason thus: the report that was published to explain the margin of error between the exit polling and official results did not consider all the possible reasons for the discrepancies, and the reason they did provide (more Kerry voters than Bush voters probably responded to the poll) was not supported by concrete evidence. Given that exit polling is normally very accurate, we need to keep investigating until we find a satisfactory explanation for the unusually wide margin of error between exit polling and official tallies. Although the authors don't say this, the implication is that without this adequate explanation, people might believe there is enough fraud or defect in our system to skew election results, which would be very undemocratic. I find that a very compelling line of reasoning.
• The evidence they provide is statistical analysis of data at their disposal, which seems sufficient to me to make their case. They show that the "WPE" ("Within-precinct Error") came in at a whopping, unheard of margin of 6.5%, and that this margin of error was associated only with certain types of voting equipment. Since they are very credible statisticians all working in academia, I trust their data and their conclusions-it seems like very strong evidence to me.
• They very respectfully give the Mitofsky-Edison report the benefit of the doubt, calling it extensive and well-intentioned. However, they do refute the report's conclusions by demonstrating how it fails to even consider the possibility that voters intentions were inaccurately recorded. It also refutes the report's declaration that the discrepancy is "probably" due to more Kerry than Bush voters responding to the exit poll by pointing out that the report has no evidence to support this conclusion, and that, in fact, the actual data suggests the opposite may be true.

Explain why you believe this is or is not a good source for your casebook.

This is a great source for my casebook on Election 2004 Fraud because it demonstrates very unemotionally and objectively that there are alarming and unexplained discrepancies surrounding the election results that professional statisticians agree need to be studied very carefully. That the information which would permit such a study has not been released suggests that someone has something to hide.

This article helps me establish that election fraud in 2004 is a possibility.