Post-Election Audit Standards Working Group, Report, Evaluation of Audit Sampling Models and Options for Strengthening California’s Manual Count
This report moves the technical and political conversation to a whole new level.
“ The literature does not frame the statistical problem in the best way: Most of the papers address essentially this question: ‘If the machine count named the wrong winner, what is the chance we will see at least one error in the sample?’ However, the Working Group believes the right question to ask is: ‘If the machine count named the wrong winner, what is the chance we would have seen more errors in the sample than we actually saw in the sample?’â€
“If audits are effective, then the public can have confidence in the outcome of elections even if the voting systems used are imperfect, because the audit can detect and be used as the basis to help correct human and voting system errors…
The complexity of these systems means there are many more ways in which voting systems can fail to capture votes correctly, lose votes, miscount votes, and be manipulated to yield incorrect results…
Auditing a small percentage of precincts is not effective for finding problems that affect only a few precincts. Moreover, no fixed percentage (short of 100%) suffices to give high confidence that the apparent outcome of the election is correct. For that goal, the number of precincts that should be tallied manually depends on the margin in each precinct, the number of ballots cast in each precinct, and other factors, including the number of discrepancies found in the precincts that are manually counted.â€
