Professional Documents
Culture Documents
W
ith the close presidential elections in 2000 and 2004, there has been a call
for election transparency, the ability of the public to have access to data
and documents pertaining to the election process in a timely manner. The
non-partisan, non-profit Election Science Institute conducted data gathering efforts that
assessed the availability of election information and the quality of that information for
both the March, 2004 Super Tuesday Presidential Primary Election and the November,
2004 Presidential Election. The results showed that overall, most jurisdictions are not yet
able to provide the data needed for election transparency.
A review of Super Tuesday states found that responses ranged from an inability to
provide any election information on election night to the provision of complete, timely
information. In some cases, only newspapers, some of which charged a subscription to
access the information, collected needed information.
It is understood and acknowledged that there are time and resource pressures on
election officials to certify the election within the allotted time, and that public requests
Inside for records exert additional demands on their limited resources. There needs to be
3 Exhibit 1: Availability of Primary Data on a mechanism in place that allows the officials to perform their duties while still hav-
Super Tuesday Election Night continued
4 Exhibit 2: Items Necessary for Election
Transparency ELECTION The Election Science Institute is an independent, non-partisan election science
organization formed in 2002 under the name “Votewatch” to monitor and im-
5 Appendix 1: List of Documents SCIENCE prove elections nationwide. For further information, contact Steven Hertzberg:
Requested from Ohio Counties INSTITUTE steven@electionscience.org
ing information available to the public. These systems should land, provided information on the web prior to certification but
be developed so that requested items critical to verifying the not on election night. Results for Maryland were not available
transparency and accuracy of elections can be met with minimal on the website until two days after the election. Paper copies
additional effort on the part of the election officials. of county-level results were also made available by the Board of
Having said this, the review of Ohio Counties following the Elections. The Election Science Institute received this informa-
2004 Presidential Election showed at least half the counties had tion two days after the election.
trouble complying with public documents requests before elec- Neither New York nor Massachusetts provides unofficial
tion certification. None of the requested Ohio counties were results on their websites. Neither of these states makes elec-
able to provide a complete audit trail within the requested time tion results available until after they are certified. Communica-
frame. Only one in eight counties was able to submit documen- tions representatives of both states confirmed that unofficial
tation that was complete enough and of sufficient quality to election results are collected by the election news service and
allow the vote counts to be reviewed. made available by the newspapers. Members of the public inter-
Based on these and other findings, the Election Science ested in election results before they are certified were referred
Institute developed recommendations to improve public data to the news sources.
availability in order to support increased election transparency. Unofficial results for Massachusetts were obtained the
These recommendations include recommendations for both morning after the election from the Boston Globe online edi-
the type of data jurisdictions should supply as well as timeliness tion. Data were available for each city/town; the only infor-
and methods of access. mation missing was the number of ballots cast. County-level
Results of The Election Science Institute election results for New York could not be found on either
Investigation The New York Times or the Wall Street Journal websites. After
Obtaining Election Night Information from Secretaries of certification, New York posts final results to their website and
On Super Tuesday 2004, The Election Science Institute Completeness: California, Connecticut, Maryland and
conducted an experiment to identify in real-time, whether the Ohio provided complete information, including the number of
data needed for independent election monitoring was available votes cast. Georgia and Rhode Island posted information but
from Secretaries of State. Throughout the election night, staff it was not complete. Georgia reported real-time results on its
contacted the offices of Secretaries of State and accessed state website, but information about the total number of votes cast
web sites in efforts to obtain information. The data items that was not available. Rhode Island posted only statewide results
staff were looking for included vote totals for each candidate in real-time on its website, so town-by-town results had to
and the total number of votes cast. Staff were looking for both be obtained from The Providence Journal online, where it was
state-level and preferably county-level data. The total number found on election night. A representative of Rhode Island’s Sec-
of votes cast was included because it was critical to examining retary of State Office said final results, including data on ballots
undervote issues raised in the 2000 election. County-level data cast, were to be available on their website 4 to 5 days after the
can be linked with data such as registration, voting equipment, election. Georgia reported real-time results on its website on
and demographics to perform a variety of statistical analyses. election night, but information about the total number of votes
The eight states that held primaries that day took a variety cast was not available until about three weeks later, even though
of approaches to responding to information requests: Georgia law requires certification of these numbers within 15
election results on their websites on election night: California, Quality/Usability: States varied in terms of the usability
Connecticut, Georgia, Ohio and Rhode Island. One state, Mary- of their data formats. The most complete and usable forms of
page
Election Transparency: How Ready Are States and Counties?
page
eight counties (12%) responded in less than two weeks. While a few counties may have had
difficulties in complying, even cooperative counties had trouble assembling the requested EXHIBIT 2:
documents in a timely manner. Items Necessary for
Election Transparency
Completeness: None of the counties that supplied documents before election Data to be readily available
certification were able to supply all of the documents requested. From the original 49 and posted on the web
Before the election
counties, only one in six counties (16%) supplied more than half of the requested docu-
• Full list of precincts and polling locations
ments before certification and the remaining four in ten counties (39%) that had submit- (not just a look-up by address) with split
precincts noted (available 1 month before
ted documents, had submitted less than half the documents requested. Only one county election)
provided at least three-quarters of the requested documents. When election experts • Final voter registration by precincts (avail-
able as soon as possible after registration
reviewed the documents that were supplied, they found there was also incorrect identi- closes, where registration closes, but not
later than 10 days before the election.)
fication of documents submitted, and statements that certain documents (E.g., problem • Machine distribution statement by pre-
reports, audit logs, zero total reports from voting or central tabulating machines) did cinct (available 1 week before election)
As soon as the unofficial count is
not exist. In one case, a county informed us that an audit log from one of their central completed
tabulating machines had been destroyed. Another county indicated that they destroyed • Precinct level summaries (also posed at
the precinct, where equipment permits)
old cumulative reports as new ones were created. including ballots cast, votes recorded and
over- and under-votes
Quality: None of the counties were able to provide accurate copies of original
• Unofficial Canvass and precinct-level
documents that provided a complete audit trail of the election process. However, six statement of votes cast, including over-
and under-votes for each race (available
counties, comprising 12% of those reviewed, did supply sufficient documents so that on the web the day after election)
most of an audit could be completed. These counties submitted documents that were • Absentee Ballot Accounting Report
(by precinct, if absentees are tracked in
primarily copies of originals and the missing items were not critical ones. Critical items precincts, available on the web 10 days
after election)
included Ballot Accounting Charts, audit logs, accumulated totals reports, the Unof- • Provisional Ballot Disposition report and
ficial Canvass, as well as precinct totals from precinct and voting machines (for counties Accounting Report, by precinct (available
on the web 3 days after the provisionals
are completed and checked)
• Statement of ballots added or subtracted
post-Election Day (overseas absentee,
remade ballots, rediscovered ballots etc.)
by precinct; (one week prior to Official
Canvass certification)
• Problem statements and responses to
problems (by precinct, where applicable)
(web-based and posted on an on-going
basis during the election period)
At election certification
• Official Canvass and statement of votes
cast by precinct including over and
undervotes for each race (one day after
certification)
Data available quickly upon
request
• Pollworker training manual/guidelines
• Ballot Accounting Charts and records of
corrections to the charts
• Audit logs from any computerized equip-
ment
• Cumulative reports created during elec-
tion night
• Sample ballot books for punch cards
• Formula or guidelines for machine
distribution
• Written guidelines for ballot tracking and
vote counting procedures
page
Election Transparency: How Ready Are States and Counties?
Official canvasses, provisional, absentee and recount reports were obtained after the original
2
were also incomplete. For example, records request, as they became available.
Election Transparency: How Ready Are States and Counties?
many counties do not include the serial number of the transfer auditability, most will need to supplement their current tracking
case (which contains the voted paper ballots) on the Ballot Ac- systems and increase availability of the documents they do have
counting Chart. to ensure public confidence and process transparency. By pro-
Overall, the results of the Ohio 2004 public records viding the items listed above, counties and states can improve
requests show that counties have many areas in which they may the transparency of the election process and provide the public
need assistance to meet public records requests supporting with the information it needs.
election transparency before election certification. In general,
but especially where elections are expected to be close, coun-
ties need to be prepared to respond quickly and accurately
to public documents requests. Ohio counties showed trouble
doing this in all areas, whether in terms of timeliness, complete-
ness or quality.
Summary
The Election Science Institute’s investigations
show that, for the most part, local jurisdictions are
not ready for the increased scrutiny of close elec-
tions. While a few jurisdictions have almost complete
page