Election Forensics

 

AEG’s team of public policy and economic experts combine data analytics with policy expertise to help federal, state, and local agencies ensure a fair and safe voting process, and to deter attempts at tampering and fraud.

Our experts use advanced statistical techniques to determine whether the geographical distribution of votes and other characteristics of election results reveal evidence of geographically localized statistical anomalies that may be a sign of tampering, fraud, or other irregularity.

With AEG election forensics services, our clients can:

  • Identify significant election anomalies within specific regions, which may indicate fraud, hacking, mistakes, or other potential sources of anomalous results.
  • Reduce the need for expensive and time-consuming recounts and audits of a large number of districts, and implement a more focused effort.
  • Enhance public confidence in election results.
  • Deter organized efforts to hack into election data, government systems, and related software and hardware, by increasing the chances that such efforts would be identified quickly

Our Approach

Based on our experience and our previously-published methodology, we have outlined a recommended plan of work for states, cities, and counties interested in deterring election tampering and enhancing voter confidence. That plan of work is as follows:

  1. Work closely with your agency to collect relevant data on the voting behavior of residents in past elections, by specific counties, precincts, or districts, in the months leading up to Election Day.
  2. Analyze this data to establish the historic statistical patterns among districts, for specific types of elections.
  3. Establish a protocol for the assessment prior to the election, specifying methods, timeline, and report format. This protocol will be prepared in a manner that can be published in advance of the election, and is explicit about the purpose of the analyses and the types of activities it is possible to detect.
  4. Once the election occurs, collect election results data from your agency and quickly perform our election forensics analysis.
  5. Prepare a report and visualizations, and answer questions regarding the data, findings, and methodology. We can work closely with your agency to provide customized deliverables and participate in presentations or testimony.

Our methods are non-invasive; do not involve any interference with voting or canvassing; protect voter anonymity; and respect the limited availability of resources and time at government agencies.

Previous Election Forensics Studies

Immediately after the November 2016 Presidential election, a number of computer scientists, activists, political candidates, and competing political factions asserted that fraud and tampering had occurred in certain states.

AEG conducted county-by-county election forensics analyses in December 2016 and January 2017 in Michigan, Wisconsin, Pennsylvania, and California. Our findings, based on that methodology, included:

  • No evidence of systematic tampering, hacking, or changes in vote tallies in Michigan, Wisconsin, or Pennsylvania.
  • Unusual voting patterns in a small number of counties, which we stated merited further inspection, but did not prove fraud had occurred. In one such county, a subsequent investigation and audit demonstrated widespread irregularities among precincts, which were attributed to widespread human error.
  • No evidence of widespread voting by immigrant populations without citizenship in Michigan, Wisconsin, or Pennsylvania.
  • An inability to determine from the aggregated county data available at that time any findings within California, and a need for further disaggregated data.

Presented below are data visualizations from our report on the 2016 presidential election in Michigan. For more information on this report, click here.

For more information, please contact Director of Public Policy and Economic Analysis Jason Horwitz at jhorwitz@andersoneconomicgroup.com or at (517) 333-6984.