In this case, we reviewed the "108% voter registration in Wood County, Ohio" allegation that implied that Obama won the Wood County election thanks to the 8% more people registered to vote than voting-age people who lived in Wood County according to the 2010 census.
My article shows how one could use the 2012 opinion polls to determine whether to accept or reject the hypothesis that over 8000 excess votes were cast for Obama in Wood County.
Promoting my 108% Voter Fraud Article
One Democratic Writing Tip taken from "108% Registration Fraud"
When I began researching this article, I saw a double allegation. First was that the 108% voter registration was fraudulent. The second allegation was that the extra 8,000+ registrations led to Obama's victory in Wood County, Ohio.
One might write a short book to cover that pair of topics, but Decoded Science prefers to finish within about 1,000 words.
Therefore I decided to avoid deciding whether the 108% registration was an unusual event, or consistent with previous elections. As well, any discussion of the legal processes of the US Census, or Ohio voter registration, were explicitly put aside rather than explored.
Instead, I crafted one "hypothesis" about the actual versus expected election result in Wood County for 2012. That became the sharp focus of my article.
Both statisticians and writers benefit from clear and concise statements.
Incidentally and slightly off the topic of this writing tip, I re-used some of my calculations from my recent post-election article in Decoded Science, "Reliable Pollster Report Card in 2012 Presidential Election". If I hadn't already crunched numbers from opinion polls, I might have reported on other numbers: census, versus voter registration, versus total votes cast, versus margin of victory. The method would be extremely similar, but since I already had reviewed some data, I made use of it. This also saved some room in the "108% Fraud" article, since the "Pollster Report Card" had determined values such as the standard deviation which I needed for the recent article.