Misadventures in Election Forensics: A Note on Choquette & Johnson 2012
14 Pages Posted: 14 Nov 2015
Date Written: November 12, 2015
Choquette and Johnson’s “2008/2012 Election Anomalies, Results, Analysis and Concerns” purports to have uncovered an anomaly that “helped…Mitt Romney win the Republican nomination in 2012 and John McCain gain votes against Barack Obama in 2008.” Specifically, they cite the positive correlation, at the precinct level, between votes cast and Romney (or McCain) vote share. Their conclusions are unsupported: (1) They offer no theory supporting their assumption that the correlation should be zero in the absence of fraud, and therefore no reason to treat a positive correlation as suspicious. (2) Although they assert that the positive correlation is anomalous, it appears in other primary and general elections — including at least one (the 2008 Minnesota contest for U.S. Senate) in which an exhaustive hand recount rebuts any inference of massive fraud. (3) Their method for estimating the true vote counts often yields implausible results, providing further evidence that their assumption is false. (4) Their theory of how fraud created the positive correlation is poorly developed, and is contradicted by evidence. An addendum briefly discusses the controversy about the 2014 Kansas general election, and comments on the 2015 Kentucky general election.
Suggested Citation: Suggested Citation