Political Climate for
outcome. Although no current procedures available. Legislators
Post-Election Audits
or pending legislation actually might be unwilling to revisit
Some states have enacted vote
requires risk-limiting audits, we election audits repeatedly, so
tabulation audit laws that require
commend states for moving jurisdictions that legislate meth-
comparing machine tallies with a
toward statistically motivated and ods can be stuck with poor meth-
hand count of a sample of ballot
justified audits. A common flaw ods indefinitely.
batches. Other states are consider-
of many bills is that they attempt
Collect election and audit data
ing such laws. Federal legislation
to legislate details of sampling
and make it accessible. A central
requiring audits has been intro-
and statistical calculations, rather
repository for data—including
duced several times. The laws and
than enunciating principles (e.g.,
election results at the precinct
bills vary in their goals and effect
“the procedure shall have at least
level or below, voting system fail-
on election integrity. Some have
a 90% chance of requiring a full
ures, errors uncovered by audits,
no enunciated goal, but require
hand count whenever that count
etc.—would be extremely valu-
that the audit trail receive some
would show a different outcome”)
able. Metadata should include
scrutiny. Some seek ‘quality con-
and leaving implementation
details about the voting technol-
trol,’ in a loose sense. Some seek
details to regulation. The push
ogy, the vote-counting technol-
to ensure that electoral outcomes
for overly detailed prescriptions
ogy, the audit, the hand-counting
are correct. States draw their ini-
seems to arise from mutual sus-
procedures, and so on.
tial audit samples differently;
picion among the stakeholders,
few require additional sampling
including local, county, and state
Help the public and elections
if the initial audit finds errors.
election officials, legislators, and
officials understand basic sta-
Some use a fixed number of
election integrity advocates.
tistical principles that apply to
batches per county across all con-
audits. For instance, it is a com-
Vote-Tabulation Audit
tests, or a number that increases
mon misconception that correct-
with population. Some require
Road Map
ing the error found in a random
auditing a fixed percentage of
From our experience in talking sample of batches corrects the error
batches or votes. Some have
with election officials, auditing in the entire contest—the idea
tiered designs, whereby one of
elections, and working with elec- that each item in a random sample
three percentages of ballots are
tion integrity advocates, we see stands for a larger number of items
audited, depending on the appar-
the following priorities: in the population is not universally
ent margin of victory.
appreciated. Similarly, jurisdictions
A few bills now being con-
Build the information infra- have passed laws requiring that the
sidered seek to limit the risk of
structure for good audits (i.e., audit sample be selected before the
certifying an incorrect electoral
data plumbing). Audits can only preliminary results are published.
be done efficiently if data are Statisticians can help explain why
promptly and publicly available that undermines the security the
in machine- and human-readable audit would otherwise give.
formats at the level of auditable
ASA Science Policy Actions
batches of ballots. Surprisingly, Statisticians possess the exper-
this is rare. Commercial elec- tise to discover and characterize
• The AsA sends a letter regarding
tion management software seems the nature, frequency, and sourc-
effects of IT consolidation on the IRs
to be the bottleneck. Adopting es of problems in our elections
statistics of Income (sOI) Division and a
standard terminology and data and reduce the risk that electoral
letter regarding the next sOI director
formats would help states ask outcomes do not truly reflect the
• The 50 slots for the AsA congressional
for useful data and make it easier will of the voters. The ASA as an
visits filled in just two weeks, with
for voting machine and software organization and several of its
representatives being 25% students
vendors to supply it. National members are working to develop
and coming from 25 states
or international standards should and disseminate model language,
be considered. techniques, and best practices for
• The AsA signs a letter to congressional
our elections. Election officials
Legislate auditing principles,
leaders thanking them for their efforts
have begun to come to us for
not implementation details.
to pass the Edward M. Kennedy
guidance. We welcome your help
Methods for risk-limiting audits
serve America Act (H.R. 1388)
to meet the important challenge
are evolving quickly. Overly pre-
of measuring, improving, and
Further information can be found at
scriptive laws prevent jurisdic-
ensuring the accuracy and integ-
www.amstat.org/outreach/scipolicyletters.cfm.
tions from using the most effi-
rity of U.S. elections. n
cient, reliable, and transparent
40 AmstAt News JUNE 2009
Page 1 |
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 |
Page 93 |
Page 94 |
Page 95 |
Page 96 |
Page 97 |
Page 98 |
Page 99 |
Page 100