The Court ordered that counsel and their forensic document examiners could review 100
randomly selected envelope/affidavits to do a signature comparison. These were
envelope/affidavits as to which election officials had found a signature match, so the ballots were
long ago removed and tabulated. Because voter names are on the envelope/affidavits, the Court
ordered them sealed. But because the ballots were separated from the envelope/affidavits, there
is no way to know how any particular voter voted. The secrecy of their votes was preserved.
Two forensic document examiners testified, one for Plaintiff and one for Defendants. The
process forensic document examiners use to testify in court for purposes of criminal guilt or civil
liability is much different from the review Arizona election law requires. A document examiner
might take hours on a single signature to be able to provide a professional opinion to the required
degree of certainty.
Of the 100 envelope/affidavits reviewed, Plaintiff’s forensic document examiner found 6
signatures to be “inconclusive,” meaning she could not testify that the signature on the
envelope/affidavit matched the signature on file. She found no sign of forgery or simulation as to
any of these ballots.
Defendants’ expert testified that 11 of the 100 envelopes were inconclusive, mostly
because there were insufficient specimens to which to compare them. He too found no sign of
forgery or simulation, and found no basis for rejecting any of the signatures.
These ballots were admitted at trial and the Court heard testimony about them and
reviewed them. None of them shows an abuse of discretion on the part of the reviewer. Every
one of them listed a phone number that matched a phone number already on file, either through
voter registration records or from a prior ballot. The evidence does not show that these affidavits
are fraudulent, or that someone other than the voter signed them. There is no evidence that the
manner in which signatures were reviewed was designed to benefit one candidate or another, or
that there was any misconduct, impropriety, or violation of Arizona law with respect to the
review of mail-in ballots.
The Court ordered that counsel could review 100 duplicate ballots. Maricopa County
voluntarily made another 1,526 duplicate ballots available for review. These ballots do not
identify the voter so, again, there is no way to know how any individual voter voted. Of the
1,626 ballots reviewed, 9 had an error in the duplication of the vote for president.
Plaintiff called a number of witnesses who observed the duplication process as
credentialed election observers. There was credible testimony that they saw errors in which the
duplicated ballot did not accurately reflect the voter’s apparent intent as reflected on the original
ballot. This testimony is corroborated by the review of the 1,626 duplicate ballots in this case,
and it confirms both that there were mistakes in the duplication process, and that the mistakes
were few. When mistakes were brought to the attention of election workers, they were fixed.
The duplication process prescribed by the Legislature necessarily requires manual action
and human judgment, which entail a risk of human error. Despite that, the duplication process for
the presidential election was 99.45% accurate. And there is no evidence that the inaccuracies
were intentional or part of a fraudulent scheme. They were mistakes. And given both the small
number of duplicate ballots and the low error rate, the evidence does not show any impact on the
outcome.