Although I am no longer auditing post-election audits in Connecticut, a new report finds significant faults in post election audits in all swing states after the 2024 election. From the report <Full Report>:
In Michigan the post-election audits are conducted months after the election, and
elections are certified based entirely on computerized totals before examining any
paper ballots. To date, Michigan has not published its state-wide post-election audit
report.
Georgia requires all in-person voters to use computerized ballot marking devices
(BMDs) statewide. In Pennsylvania, 30% of in-person voters must vote on these
devices. These computerized devices do not provide the trustworthy record of voter
intent necessary for an effective audit.
Nevada primarily uses direct record electronic voting machines (DREs) with paper
audit trails and BMDs to record votes. These computerized devices also do not
provide the trustworthy record of voter intent necessary for an effective audit.
Georgia, Pennsylvania and Nevada all claim to conduct “risk-limiting audits” (RLAs),
which are generally acknowledged as the gold standard of auditing. In fact, none
meet the requirements for a true RLA.
Pennsylvania is bound by court settlement to conduct risk-limiting audits on all
election contests. But in apparent non-compliance with the settlement, not only
does it not adhere to the requirements of a risk-limiting audit, in 2024 Pennsylvania
only conducted its “RLA” on one contest, the state treasurer’s contest.
Current Nevada statute directs the state to audit 2% of ballots for all contests, in
addition to its “RLA” but there is no evidence the 2% audit was conducted.7
Arizona audits include many effective provisions, and are more transparent than
other states, providing evidence to support the election outcome, but they could be
enhanced.
North Carolina’s audits adhere to several effective practices, but like Arizona, its
audits could be improved to provide stronger evidence that its election outcomes are
correct.
Wisconsin audited far more ballots than any other state, but its audit results were
published months after the election was certified. Moreover, officials published no
source documentation or details of
elections are certified based entirely on computerized totals before examining any
paper ballots. To date, Michigan has not published its state-wide post-election audit
report.
Georgia requires all in-person voters to use computerized ballot marking devices
(BMDs) statewide. In Pennsylvania, 30% of in-person voters must vote on these
devices. These computerized devices do not provide the trustworthy record of voter
intent necessary for an effective audit.
Nevada primarily uses direct record electronic voting machines (DREs) with paper
audit trails and BMDs to record votes. These computerized devices also do not
provide the trustworthy record of voter intent necessary for an effective audit.
Georgia, Pennsylvania and Nevada all claim to conduct “risk-limiting audits” (RLAs),
which are generally acknowledged as the gold standard of auditing. In fact, none
meet the requirements for a true RLA.
Pennsylvania is bound by court settlement to conduct risk-limiting audits on all
election contests. But in apparent non-compliance with the settlement, not only
does it not adhere to the requirements of a risk-limiting audit, in 2024 Pennsylvania
only conducted its “RLA” on one contest, the state treasurer’s contest.
Current Nevada statute directs the state to audit 2% of ballots for all contests, in
addition to its “RLA” but there is no evidence the 2% audit was conducted.7
Arizona audits include many effective provisions, and are more transparent than
other states, providing evidence to support the election outcome, but they could be
enhanced.
North Carolina’s audits adhere to several effective practices, but like Arizona, its
audits could be improved to provide stronger evidence that its election outcomes are
correct.
Wisconsin audited far more ballots than any other state, but its audit results were
published months after the election was certified. Moreover, officials published no
source documentation or details of
