EAST LANSING, Mich., December 16th, 2016— Anderson Economic Group, LLC has completed a forensic analysis of the 2016 election results. In a letter to the Secretary of State written by Patrick Anderson and Alex Rosaen, AEG concluded that the outcome of this analysis shows no evidence of fraud or tampering with the Michigan 2016 general election results of the office of the President. This analysis was motivated by three sets of assertions of alleged fraud, mistake, and tampering including the recall petition of Jill Stein alleging “fraud or mistake” by county and state boards of canvassers, as well as other election officials; the public statements that there is “evidence that vote totals could have been manipulated”; and the statement by the White House press secretary that the President has ordered a review of an alleged “pattern of malicious cyber activity related to our presidential election cycle.”
AEG examined the certified election results from the State Board of Canvassers with exploratory data analysis tasks, and two sets of hypothesis tests. These methods were employed to detect evidence of fraud, or indicate no such evidence, consistent with the purpose and the limitations of election forensics methods.
The statistical technique and tests in this routine are intended to detect episodic tampering with voting results in certain voting districts. Alex Rosaen, Director of Public Policy and Economic Analysis at AEG, stated: “It would be very difficult to engage in large-scale tampering with vote counts and have the results look like these, with no unexplained outliers and a smooth distribution of county vote totals.” Such episodic, county-specific tampering is implied by at least two of the three sets of public assertions recorded as motivating factors here. For example, if foreign hackers infiltrated the voting machines in some counties in the 2016 general election and changed the voting results in those counties, but did not do so in the 2012 election and could not or did not do it in all counties, it would cause the types of patterns that these routines are designed to detect.
Note that these statistical tests attempt to find deviations that are unusual, which usually means they are larger than would be expected given the distribution of other votes. As a consequence of this focus on significant deviations, these statistical tests cannot be expected to detect small numbers of irregularities, especially if they occur regularly in elections or are spread around the entire state.
The exploratory data analysis tasks included creating and examining histograms of vote shares by county in two elections, scatter charts showing same-county vote shares across two elections, and empirical cumulative distribution functions for the same set of candidates and elections. The results are all consistent in showing a pattern of votes, and shifting of votes across elections, that is not unusual. There is no evidence from these results that suggests no irregularities or episodic tampering.
The results from both sets of hypothesis tests indicate no evidence of fraud or tampering with the Michigan 2016 general election results for the office of President. Thus, inspection of the pattern of voting in 2016, and basic statistical analysis of the 2012 and 2016 election, and two hypothesis tests, all provide no evidence of episodic tampering or fraud in the Michigan 2016 general election results for the office of President.
About Anderson Economic Group
Since its founding in 1996, Anderson Economic Group has become one of the nation’s premier boutique consulting firms, specializing in business valuation, expert testimony, and public policy. AEG has assisted governments, universities, large corporations, and businesses across the U.S. and abroad. All of the firm’s work is based on its core values of Professionalism, Integrity, and Expertise. AEG is headquartered in East Lansing, Michigan, and has offices in Chicago, New York City, and Istanbul. For more information: www.andersoneconomicgroup.com
CONTACT: Tyler Marie Theile, Anderson Economic Group, LLC