The 3 Principles ASA Endorsed
Principle 5: Risk-Limiting Audits Principle 6: Addressing Discrepancies and
Post-election audits reduce the risk of confirming an incorrect outcome. Continuing the Audit
Audits designed explicitly to limit such risk (risk-limiting audits) have
When discrepancies are found, additional counting and/or other inves-
advantages over fixed-percentage or tiered audits, which often count
tigation may be necessary to determine the election outcome or to find
fewer or more ballots than necessary to confirm the outcome.
the cause of the discrepancies.
a. Risk-limiting audits have a large, pre-determined minimum chance of
a. Audit protocols must clearly state what will result in counting more
leading to a full recount whenever a full recount would show a different
audit units. Such factors might include the number of discrepancies and
(Correct preliminary outcomes are never overturned.
their distribution across the sample. Protocols must also specify the meth-
After any audit, this chance should be calculated and published as part of
od to determine how many additional audit units will be selected and
the audit results to promote continuous improvement.
under what circumstances a full recount will be conducted. For a risk-
b. Audit units (precincts, machines, batches of paper records) should be
limiting audit, the decision of whether to count more audit units is based
selected using appropriate random sampling methods.
In a risk-limit-
on a calculation of the risk; the number of additionally selected audit
ing audit, the sample size will depend on the margin of victory and other
units depends crucially on the discrepancies that have been uncovered.
factors; these other factors may include the number of ballots in each
b. The plan for continuing the audit must ensure that all stages in count-
precinct and the overall number of ballots in the contests.
ing take place before reporting final results. Moreover, the plan should
smaller margins of victory and smaller contest require auditing a larger
aim to control the cost of post-election audits while achieving any speci-
percentage of the audit units.
fied risk limit.
c. To reduce the burden of counting ballots while still auditing a vari-
ety of contests, it may be appropriate to use different rules for auditing
some contests than others. For example, it may be appropriate to allow Principle 7: Comprehensive
more risk for non-statewide contests.
Jurisdictions may require audits
All jurisdictions and all ballot types, including absentee, mail-in and
in some contests and randomly select others to be audited, so that every
accepted provisional ballots, should be subject to the selection process.
contest has some possibility of being audited. For smaller contests, it may
also be appropriate to use alternative audit methods such as targeted sam-
a. Ballots from different jurisdictions and ballot types can be divided into
pling (see Best Practice 8) or random sampling based on a fixed number
distinct groups that are audited in separate phases. In each phase, the
or percentage of audit units.
random selection of units to audit must not commence until prelimi-
nary results for each audit unit in that group have been reported to the
d. The selected audit units must be fully and manually counted.
each selected audit unit, the audit must compare vote count subtotals
from the preliminary reported election results with hand-to-eye counts
b. All types of ballots, even those used by few voters, should be subject
of the corresponding paper records.
to the selection process.
These might include overseas or military bal-
lots, faxed ballots, telephone ballots, ballots transmitted over the Internet,
e. For efficiency, large groups of ballots can be divided into batches,
ballots cast through accessible interfaces “voter-verified paper audit trail”
each comprising an audit unit. In this case, the subtotals for each batch
ballot images, and ballots cast using any other future technology.
must be reported prior to the audit as part of the election results. For
instance, absentee ballots (if not sorted and counted by precinct) can
be divided into batches.
“Outcome” refers to which candidates or ballot propositions won or lost,
Discrepancies found during the audit can also affect the sample size, as dis-
not necessarily a specific vote tally. Here we refer to the outcome as “correct” or cussed in 6a.
“incorrect” depending on whether it corresponds with what would be the out-
come from a complete manual recount. Note that the outcome from a com-
All else being equal, contests spanning fewer audit units—for instance, local
plete manual recount may not always match the will of the voters. To ensure
contests as opposed to statewide contestsv—require proportionally larger audits
that outcomes reflect the will of the voters, additional conditions must be met
to ensure that the chance of confirming an outcome that is incorrect is low.
including rigorous ballot accounting, accurate registration data, elimination of 
“Manual counting” or “hand counting” refers to human visual inspection of
unreasonable delays at the polls, good ballot design, and controls on chain of
paper records to interpret voter intent, followed by a tabulation of the individ-
custody for all election equipment and materials.
ual vote interpretations. Only the tabulation portion is sometimes assisted by
Fixed-percentage samples are inadequate for risk-limiting audits, because
independent and well trusted equipment such as calculators and spreadsheets.
the audit size needed to verify an election outcome depends on the apparent
All hand counts should be done blind to the expected result.
margin of victory, as well as the number of audit units and the amount of error 
When auditing less common ballot types or very small precincts, care must
each audit unit can harbor. However, auditing some minimum percentage of
be taken to preserve voter anonymity and the secrecy of the individual voter’s
votes or audit units regardless of jurisdictional size or election margin may be
ballot. Also, it may be possible to confirm the election outcome without sam-
useful to monitor election accuracy. Generally, requiring a smaller chance of
pling some types of ballots, if these types do not contain enough ballots to alter
error (e.g. 1% versus 5%) will entail auditing more ballots.
the outcome (see footnote 4). However, for fairness and to provide valuable
If audit results indicate that the initial outcome is incorrect, ultimately a
information about the quality of the election process, all ballot types should
full recount would be required to determine the final outcome. Preliminary
be routinely audited.
outcomes cannot be overturned based on audit samples alone. 
In all cases, voter-verified paper ballots or records must be available for the
In the selection, some units may be weighted more than others based on
audit. Auditability—the ability to conduct reliable and efficient audits—should
their size and the amount of error they could harbor. Random sampling is
be a crucial criterion when selecting voting technologies.
unnecessary if all audit units will be manually counted, or if so many audit
units are counted that the remaining units cannot change the outcome.
Please view the full list of principles at http://electionaudits.org/principles
OCTOBER 2008 AMSTAT NEWS 7
| Page 2
| Page 3
| Page 4
| Page 5
| Page 6
| Page 7
| Page 8
| Page 9
| Page 10
| Page 11
| Page 12
| Page 13
| Page 14
| Page 15
| Page 16
| Page 17
| Page 18
| Page 19
| Page 20
| Page 21
| Page 22
| Page 23
| Page 24
| Page 25
| Page 26
| Page 27
| Page 28
| Page 29
| Page 30
| Page 31
| Page 32
| Page 33
| Page 34
| Page 35
| Page 36
| Page 37
| Page 38
| Page 39
| Page 40
| Page 41
| Page 42
| Page 43
| Page 44
| Page 45
| Page 46
| Page 47
| Page 48
| Page 49
| Page 50
| Page 51
| Page 52
| Page 53
| Page 54
| Page 55
| Page 56
| Page 57
| Page 58
| Page 59
| Page 60
| Page 61
| Page 62
| Page 63
| Page 64
| Page 65
| Page 66
| Page 67
| Page 68
| Page 69
| Page 70
| Page 71
| Page 72
| Page 73
| Page 74
| Page 75
| Page 76
| Page 77
| Page 78
| Page 79
| Page 80
| Page 81
| Page 82
| Page 83
| Page 84
| Page 85
| Page 86
| Page 87
| Page 88
| Page 89
| Page 90
| Page 91
| Page 92