Election Data Standards Requirements: Getting on with what we - PowerPoint PPT Presentation

About This Presentation
Title:

Election Data Standards Requirements: Getting on with what we

Description:

Election Data Standards Requirements: Getting on with what we ve got John L. McCarthy, volunteer Verified Voting Foundation Common Elections Data Format Workshop – PowerPoint PPT presentation

Number of Views:72
Avg rating:3.0/5.0
Slides: 17
Provided by: JohnMc123
Learn more at: https://www.nist.gov
Category:

less

Transcript and Presenter's Notes

Title: Election Data Standards Requirements: Getting on with what we


1
Election Data Standards RequirementsGetting on
with what weve got
  • John L. McCarthy, volunteer
  • Verified Voting Foundation

Common Elections Data Format Workshop National
Institute for Standards and Technology Gaithersbur
g, Maryland 29-30 October, 2009
2
Overview Review Background for election data
standards
  • Who needs uses election data? (clients)
  • What kinds of election data are required?
  • When are election data needed for what purposes?
  • What objectives would data standards help meet?
  • How are these needs currently being met?
  • in the United States?
  • In other countries?
  • What characterizes good data format standards?
  • Why cant we simply use EML ( extend as
    necessary)?
  • OASIS Election Markup Language (dialect of W3C
    XML)

3
Who needs uses election data ( how)?Potential
clients for election data standards
  • Voting systems vendors and system developers
  • component communications, system integration,
    testing, reporting
  • Election officials local, state, and national
    (EAC, )
  • ballot definition, testing, reporting,
    aggregation, auditing
  • Election management consultants contractors
  • systems integration, contract work for election
    officials
  • News media (TV, radio, print, web)
  • reporting results, predicting outcomes
    analyzing trends
  • Candidates, political parties organizations
  • deciding whether to concede, claim victory or
    dispute results
  • Citizens, citizen organizations academic
    researchers
  • pre and post election auditing, analyzing
    detailed results

4
What kinds of election data are required?
  • Election districts district boundaries
  • Voter registration information eligible voter
    lists
  • Candidate nominations approved candidate lists
  • Referendum options and approved options lists
  • Ballot definition information (for each
    jurisdiction)
  • Election vote records, counts, results, and
    statistics
  • Cast Vote Records (CVR) for each individual
    ballot
  • including outcomes for each voting opportunity
    (choice)
  • e.g., vote recorded, blank, too many choices,
    unrecognized
  • Logs from each individual piece of voting
    equipment
  • Audit information pertinent to all the above
    categories

5
What detailed components are needed for vote
tabulation audits?
  • GEOGRAPHIC IDENTIFIERS
  • State, County
  • Sub-county jurisdiction(s), if any (e.g., city,
    township)
  • Precinct
  • Other Aggregation Unit Identifiers (e.g., state
    assembly district, water district)
  • Voting Method (early, absentee, in-precinct,
    provisional,
  • Ballot Type and/or party (for primary elections)
  • FOR EACH CONTEST
  • Contest (e.g., Governor, State Assembly, City
    Council, Water Board)
  • Choice (candidate or position Y/N)
  • Summary records typically contain counts for each
    choice and some systems cast vote records for
    individual ballots may show how each choice was
    counted -- vote, blank, too many choices
    (overvote), or unrecognized mark.
  • SoSs others also need standards for various
    types of election audit reports

6
When are election data needed?
  • Preceding an election
  • system development testing
  • logic and accuracy testing test results
  • jurisdiction boundaries, ballot types, voting
    places
  • ballot design and contents (candidates, ballot
    measures, etc.)
  • registered eligible voters
  • During an election
  • problem reports
  • individuals who have voted
  • Election night
  • detailed vote counts by polling place, type
    (in-person, absentee), candidate, ballot measure
    choices, overvotes, undervotes
  • Individual Cast Vote Records (CVR) for each
    ballot
  • Before certification of final results
  • audit results, including resolution of any
    discrepancies found

7
Objectives that election data standards can help
us achieve
  • Timely Transparent Reporting
  • aggregation within local jurisdictions from
    local to state
  • to media, interested organizations the general
    public
  • to help support pre and post-election auditing
  • Lower costs improved Accuracy
  • Improve transparency testing of ballot
    definition
  • connect registration, pollbooks, and reporting
  • facilitate transition to electronic
    record-keeping
  • Interoperability
  • between components from a single vendor
  • among different components from different vendors
  • Auditability
  • detailed data available immediately following
    each election
  • machine-readable reports broken down in arbitrary
    ways

8
How are these needs currently being met?
  • In the United States
  • very little standardization
  • data exchange via poorly documented proprietary
    formats
  • election management systems produce human
    readable reports
  • some exceptions
  • CA SoS media feed 2008, 2009
  • IL translation programs for EAC data collection
    grant program
  • In other countries
  • Council of Europe recommends EML for
    interoperability (2004)
  • Australian Electoral Commission EML Media Feed
    (since 2007)
  • UK e-voting pilots and CORE registration project
    use EML
  • Belgium uses EML for local elections in Flanders
    (2006-7)
  • Others?

9
What kinds of data and metadata do current
commercial vote tabulation systems provide?
Human Readable Reports
e.g., Hart-Intercivic (Crystal Reports)
10
What would characterize good election data
format standards?
  • Machine-readable, structured components
  • separate elements for each distinct type of
    information
  • (e.g., state, county, precinct, type, contest,
    candidate, undervotes)
  • easy to render into different formats
  • modular structures/schemas for different kinds of
    data
  • (e.g., ballot definition, geography, tabulation
    results, )
  • Well-defined and documented data elements
    structures
  • preferably defined by data verifiable via
    formal schema
  • Quasi-human-readable
  • data volume does not require serious compression
    (e.g., ASN.1)
  • easy to render into different human-readable
    machine formats
  • Compatible with tools for translation, rendering
    storage
  • e.g., XML style sheets, schema, databases web
    services XSLT
  • Developed through standards consensus process
  • input and discussion from all stake-holders,
    trial use, etc.

11
Doesnt EML (Election Markup Language) meet most
if not all of these requirements?
  • Dialect of XML (current lingua franca for data
    exchange)
  • Developed by OASIS Technical Committee (since
    2001)
  • participation by vendors and election experts
  • currently completing work on version 6.0 (still
    time for feedback!)
  • OASIS will propose EML 6.0 as ISO standard early
    in 2010
  • Flexible, extensible, modular framework
  • version 6.0 includes new elements features to
    support US voting
  • V 6.0 meets most known election requirements
  • Already used by a number of organizations
    jurisdictions
  • California Australia media feeds, etc.
  • ESS, Hart-Intercivic (EDX XML variant), EDS,
    IBM, more in Europe
  • For more info, see http//www.oasis-open.org/commi
    ttees/ tc_home.php?wg_abbrevelectionexpository

12
What are primary objections, barriers, and
counter-arguments to use of EML?
  • Too new ?
  • development of multiple versions since 2001
  • used successfully in growing number of
    jurisdictions
  • Competing approaches standards ?
  • IEEE Voting Systems Electronic Data Interchange
    Project 1622
  • temporarily deactivated because TC "failed to
    achieve balance"
  • Comma-delimited spreadsheet format
  • No schema to enforce data input requirements
  • Require multiple tables to supported nested
    repeating groups
  • Would have to develop table and column
    definitions, etc.
  • Too complex and/or missing features ?
  • can ignore modules that are not applicable
  • Easy to extend and add new features using XML
    (e.g. audit reports)
  • Implementation costs ?
  • 3 major vendors already use EML or XML in
    significant ways
  • Lots of tools to support XML development and use

13
The need is urgentNow is the time to act
  • Election auditing requires a single standard set
    of formats
  • statement from last weeks meeting on election
    auditing at ASA
  • States are beginning to implement electronic
    reporting
  • California 6 county experiment plans to expand
    to statewide
  • Illinois plans statewide integrated voting
    elections system
  • Need for national archive of election data
  • for policy makers, legislators, academic
    researchers
  • current election day survey data is inadequate
  • not timely, detailed data not easily available in
    standard formats
  • EAC data collection grant project results can
    provide insights
  • If EML is deficient, we can propose revisions for
    v6
  • but should do so in the next couple of months

14
Opportunities for participation
  • Election Data Standards Email list ( google
    sites wiki)
  • electiondatastandards_at_verifiedvoting.org
  • Try new election data software help improve it
  • Auditing software from CO (McBurnett), UC
    Berkeley (Stark),
  • VTS translation software from IL?
  • EML enhancements for version 6
  • OASIS Elections Voter Services Technical
    Committee
  • Joe Hall, David Webber, others
  • www.oasis-open.org/committees/election/
  • NIST, TGDC, VVSG
  • Urge EAC and/or NIST to become active members of
    OASIS TC
  • create documentation guidelines to facilitate
    adoption

15
Thanks to .
  • Verified Voting Foundation President Pam Smith
  • Election Data Standards and Auditing Lists
  • American Statistical Association Steve Pierson
  • David Webber, OVS/OASIS
  • John Sebes, Open Source Digital Voting Foundation
  • Neal McBurnett, Boulder, Colorado
  • Scott Hilkert Catalyst Consulting associates,
    Chicago
  • participants in last weeks election auditing
    meeting

16
Example XML data fragment
  • lt?xml version"1.0" encoding"utf-8" ?gt
  • - ltelection type"GE" name"General Election"
    date"11/4/2008"gt
  • - ltstate id"IL" name"Illinois"gt
  • - ltjurisdiction id"2402" name"Alexander County"
    federalId"1700300000"gt
  • - ltcontest id"4" name"12TH CONGRESS"
    polling"3167" absentee"0" early"487" grace"0"
    provisional"0" total"3654"gt
  •   ltspecialCount type"blankVotes" polling"0"
    absentee"0" early"0" grace"0" provisional"0"
    total"0" /gt
  • - ltspecialCount type"underVotes" polling"283"
    absentee"0" early"88" grace"0" provisional"0"
    total"371"gt
  •   ltprecinct name"CAIRO 1" polling"57"
    absentee"0" early"14" grace"0" provisional"0"
    total"71" /gt
  •   ltprecinct name"CAIRO 2" polling"37"
    absentee"0" early"21" grace"0" provisional"0"
    total"58" /gt
  •   ltprecinct name"CAIRO 3" polling"16"
    absentee"0" early"14" grace"0" provisional"0"
    total"30" /gt
  •   ltprecinct name"CAIRO 4" polling"22"
    absentee"0" early"14" grace"0" provisional"0"
    total"36" /gt
  •   ltprecinct name"CAIRO 5" polling"19"
    absentee"0" early"9" grace"0" provisional"0"
    total"28" /gt
  •   ltprecinct name"CACHE" polling"9" absentee"0"
    early"5" grace"0" provisional"0" total"14" /gt
  •   ltprecinct name"SANDUSKY" polling"7"
    absentee"0" early"2" grace"0" provisional"0"
    total"9" /gt
  •   ltprecinct name"TAMMS" polling"39"
    absentee"0" early"1" grace"0" provisional"0"
    total"40" /gt
  •   ltprecinct name"MCCLURE" polling"29"
    absentee"0" early"2" grace"0" provisional"0"
    total"31" /gt
  •   ltprecinct name"THEBES" polling"19"
    absentee"0" early"1" grace"0" provisional"0"
    total"20" /gt
  •   ltprecinct name"OLIVE BRANCH" polling"29"
    absentee"0" early"5" grace"0" provisional"0"
    total"34" /gt
  •   lt/specialCountgt
Write a Comment
User Comments (0)
About PowerShow.com