Featured Product
This Week in Quality Digest Live
Quality Insider Features
Don’t underestimate the power of telling your story
Oak Ridge National Laboratory
Hafnium oxide key to novel applications
William A. Levinson
Deciding whether you need CAPA or a bigger boat
Matthew T. Hughes
Two engineers explain how heat waves threaten everything from cars to computers

More Features

Quality Insider News
To be unveiled during PACK EXPO Las Vegas at the Hiperbaric booth, No. N-10857
Educating the next generation of machinists with state-of-the-industry equipment
In a first, researchers have observed how lithium ions flow through a battery interface
Air-extend/spring-retract moves probe out of the way
Precision cutting tools maker gains visibility and process management across product life cycles
Expanded offering includes Smartscope E-series for 3-axis video measurement
Accelerates CAM programming time 80% to make U.S. manufacturers more productive
Pioneers new shape-memory alloys
A Heart for Science initiative brings STEM to young people

More News

Larry P. English

Quality Insider

Fixing Broken Election Processes

An information quality mandate

Published: Monday, November 27, 2006 - 23:00

Déja vu
After the 2006 primary elections in the United States, a local newspaper article headline read, “Computer ballots in stage of ‘trial and error.’” Although elections are critical in a democratic society, electronic voting that isn’t transparent has been introduced with a trial-and-error approach instead of a proven process improvement approach.

After the 2000 U.S. presidential election, the “Help America Vote Act” (HAVA) election reform was intended to improve the election processes. Its requirements to replace older voting technologies with information-age electronic voting machines have introduced negative side effects and increased known problems.

The mandates for e-voting and other purported reforms failed to analyze root causes of the broken processes in 2000 and again in 2004. Although I have the utmost respect for the commission leaders of the National Commission on Federal Election Reform, they failed to follow the fundamental process improvement cycle (plan-do-study/check-act [PDSA]).

How broken are U.S. election processes?
In the 2000 U.S. presidential election, there were 4–6 million disenfranchised voters—1.5 to 2 million were due to faulty equipment, 1.5 to 3 million due to registration mix-ups, and up to 1 million lost in polling place operation failure, plus an unknown number due to absentee ballot problems. The 4–6 million votes lost in the 2000 presidential election due to election-process failures is the equivalent of throwing out every vote cast in Florida.

Election process failure 3.0 in 2006
Hundreds, if not thousands, of reported glitches and failures have been noted in the 2006 elections. Here’s a small sampling of the complaints:

  • Williamson County, Tennessee, voters had to revert to paper ballots because Election Systems & Software failed to correctly program the new iVotronic voting machines.
  • The ES&S audit reports in Williamson County showed that 18,989 people had voted at 9:30 p.m. on election day, then 17,377 at 2:30 a.m. the following morning, and finally that 16,507 people had voted as of the next afternoon.
  • Several south Florida voters say the choices they touched on the electronic screens were different from those on the review screen, the final voting step. The Broward County supervisor of elections said that it’s not uncommon for screens on heavily used machines to slip out of sync, making votes register incorrectly.
  • Programming errors and inexperience with electronic voting machines frustrated poll workers in hundreds of precincts, delaying voters in Indiana, Ohio and Florida, and left some voters with little choice but to use paper ballots instead.
  • About 175 of 914 precincts in Indiana’s Marion County turned to paper ballots, because poll workers didn’t know how to run the machines.
  • In Denver, software crashes and overloaded servers were to blame for extraordinary lines, with police being called out in at least one case.
  • U.S. Senate candidate James Webb’s last name was cut off on part of the electronic ballot used in Alexandria, Falls Church and Charlottesville, because of a computer glitch that affects candidates with long names.
  • Election officials in Delaware County, Indiana, extended voting hours because voters initially couldn’t cast ballots in 75 precincts. The cards that activate the push-button machines were programmed incorrectly.
  • In Williamson County, Tennessee, the election procedures posted on my polling place weren’t clear or consistent with the design of the e-voting machine, leading to confusion. I witnessed two voters walk away from their voting machines without the final casting of their vote, with poll workers calling them back to complete their vote.

These problems are trivial compared to the fact that—by design—current certified e-voting systems in the United States aren’t transparent and don’t have a verifiable, controlled audit capability. Nontransparent voting systems don’t produce proof that a voter’s intention was correctly recorded and tallied, nor do they enable a verifiable recount against a voter-verified audit record. This puts the election process integrity at risk. What can be done?

Information quality (IQ) principles for election reform
We must apply quality principles and error-proofing to all election processes, from ballot access and voter registration to accurate counts and recounts as necessary. Following is a sample of the principles and improvement recommendations.

The election processes should be reviewed in the form of IQ principles and quality improvement (QI) recommendations.

IQ principle 1—W. Edwards Deming’s first point of quality says, “Create constancy of purpose for improvement of product and service.” Management must solve the problems of tomorrow. “The obligation to the customer to constantly improve design of product and service never ceases.” In elections, this means the obligation to the voter never ceases.

Election processes must be improved and controlled to assure the voters’ needs—to have their votes cast and counted accurately—are met.

IQ principle 2—Poor quality causes waste, process failure and adds the cost of scrap and rework. In business, poor quality causes customer dissatisfaction and lost business.

In election processes, if disqualified votes would have changed the outcome of the election, the cost of failure means changes to the course of human history.

Voter bill of rights
Truth-in-lending laws require financial institutions to disclose information about the costs and terms of a loan. Truth-in-advertising laws require retailers to make honest claims about their products or services. The National Highway Transportation Safety Administration, the Food and Drug Administration and other consumer-protection agencies seek to protect consumers from faulty products that can cause injury or death. The U.S. Constitution has a Bill of Rights that guarantees citizens’ freedom in a democratic society. There should be a “Voter Bill of Rights” that guarantees every eligible U.S. citizen the right to vote, and to have their vote counted and tallied accurately. This defines quality for the voter. Every election year, countless voters are denied the opportunity to vote, or their votes are counted incorrectly or not counted at all, which is more than enough evidence that voters need such a bill of rights.

A rejected ballot or an incorrectly recorded vote that doesn’t represent the intention of a voter is disenfranchisement, regardless of the cause. The only acceptable goal for election processes is zero defects—not one denied eligible voter or spoiled vote for any reason.

NASA’s space program has a goal of zero defects in all missions that put people in space. Shouldn’t there be a goal of zero defects in election processes that put people into office to govern us?

IQ principle 3—Deming’s fifth point states “Improve constantly and forever the system of production and service.” The goal of process improvement isn’t just the elimination of the defective products (i.e., disqualified or unrecorded votes); the goal is to eliminate the cost of the consequences. Substituting e-voting technology for punch-card technology without process improvement and control simply exchanges one defective vote rate for another.

  • QI recommendation: Adopt a goal to reach a quality standard of Six Sigma or zero defects in spoiled and disqualified votes, and qualified citizens prevented from voting because of voter registration or other election process error. Philip Crosby correctly states, “There is absolutely no reason for having errors or defects in any product or service.”

This requires a culture of quality improvement to prevent disenfranchisement of any U.S. citizen. Six Sigma quality, the statistical version of zero defects, allows no more than 3.4 errors in one million opportunities—no more than 340 disqualified votes out of 100 million. Each state and voting jurisdiction should adopt an achievable—but stretch—target, such as a 50-percent reduction in disqualified votes in the 2008 election. If every jurisdiction halved the number of disqualified votes every two-year election cycle, zero defects (around 305) would be reached in 2032.

Election reform requires designing quality into the process and technology to make this happen.

IQ principle 4—Quality management means no blame or fault-finding when seeking to solve a problem and improve processes. Defective products result from defective processes, not defective people. Incorrect (wrong selection), unrecorded (undervotes) or disqualified (overvotes) votes aren’t the result of confused or uneducated voters—they’re the result of confusing ballots, error-prone technology, lack of training in the voting procedures and lack of feedback as to how to prevent errors.

  • QI recommendation: Recognize that human error isn’t a root cause. Analyze and discover root causes, then error-proof the process to prevent the conditions that caused the human error.

Ballot access
The high cost of running for office keeps many worthy candidates off ballots. This limits the candidates to those with money or access to money. This prevents the United States from having a government “of the people, by the people, for the people.” Furthermore, visibility and franking privileges make the current system biased toward incumbents.

IQ principle 5—To have a government of peers, equal opportunity must be extended to all qualified citizens to run for elective office, regardless of ethnic origin, gender, creed or financial means. Root causes such as the high costs of campaigning funded by one’s own resources and contributions from others, provides a burden to those without money or sources of money.

  • QI recommendations:

    1. The solution is to reform campaign finance to public funding only, with adequate maximum funding for specific elective office positions.Special interest funding causes obligations to special interests.
    2. The order of names on ballots should be in sequence such that an incumbent running for re-election or the candidate from the incumbent’s party is last, preceded by the candidate of the next major party, preceded by the candidate of independent parties based on the relative size of the parties in each jurisdiction, with the smallest having its candidate listed first.
    3. Make the campaign season end the day before early voting starts. Those who vote early miss breaking news that may happen after they’ve voted.

IQ principle 6—Campaign advertising must be accurate and nonbiased. Biased ads mislead.

  • QI recommendation: Prohibit negative, attack ads. Ads found to be libelous or slanderous to opposing candidate would be grounds for disqualifying the candidate or party authorizing the false advertising.

Voter registration
Citizens must have easy access to voter registration. The voter registration process must ensure that persons presenting themselves are qualified. Voter registration data must be easily available on election day or during early voting. Processes must be in place to capture updates to citizens’ new addresses and voting jurisdiction to prevent fraud and the denial of voting rights. Currently, American residents have a mobility rate (change of address) of about 15 percent per year (with a high of 20 percent in 1984–85).

IQ principle 7—Process integrity and information quality begin with a clear definition of the data to be known and the integrity of the design of the databases to house that information. States must have well-defined voter registration databases with easy, secure access by each other to check voter movement.

  • QI recommendation: Design the voter registration database and process to capture “identity attributes” such as birth date, mother’s maiden name, last four digits of one’s Social Security number, driver’s license number and digital signature.

    The process and database must also be designed to capture and maintain name changes, and previous addresses of individuals with relative effective dates to handle the information quality decay of address changes.

    1. Processes should be in place to capture current addresses easily. National Change of Address (NCOA) data from the U.S. Postal Services and motor vehicle registration should be utilized to capture new permanent addresses. Access to a database of deceased persons will help prevent voting fraud.
    2. Registered voters should be able to update their name and address data easily. Enable updates through the Internet using secured services and proper identification verification to reduce problems on election day.
    3. Provisional ballots should be made available where eligibility requirements are claimed, but not verifiable at the polling place.

Ballot design and information presentation quality principles
IQ principle 8—There are three components to information quality:

  • Definition: you know the meaning of the data, e.g., “Touching this button causes your vote to be recorded,” or “party” means “an established political group that supports specific principles and to which a candidate may belong"
  • Content: accurate value, e.g., the name of the party by the candidate’s name is the correct party name
  • Presentation: Information, such as instructions for voting, is presented in a way that is clear to the voter, or when actions are required or information is requested, the information is presented in a way that leads the knowledge worker to make a right action. Information presentation design is critical for communicating information, requesting action or collecting information.
  • QI recommendation:

    1. Design ballots with all candidates for one race chunked in a single space with unambiguous placement of candidate names with their selection mechanism. Test voting instructions and procedures against the ballot design to ensure consistency.
    2. Review instruction wording for clarity. Replace the instructions “Vote for Group” with a clear, natural text, such as “Vote for one president/vice president team by shading completely the oval next to their names.”
    3. Use multisensory design techniques. For example, include candidates’ photographs as well as names for visual recognition. Electronic recording devices can use sound to indicate the selection of a candidate or when an incorrect choice is made, and when the final recording of a vote has been made.

The ballot designers or the people who signed off in approval of the ballot cannot be blamed. A basic principle of accountability is “We cannot hold someone accountable if they haven’t been trained or taught the principles.”

IQ principle 9—Deming’s sixth point states, “Institute training on information quality.” People cannot do a quality job if they don’t know how.

  • QI Recommendations:

    1. Provide training in information quality principles for election officials (ballot designers, procedure writers, poll workers, vote counters, etc.) Standardized training includes general principles with modules that address different voting systems. A key focus teaches error-proofing techniques to prevent common errors inherent in the specific voting technologies.
    2. Provide training and guidance in how to conduct “usability tests” of ballots for voter-friendliness with likely voters. This includes the design of e-voting screens, including the candidate or referendum presentation, the verification screen, and the final “cast your vote” screen.

IQ principle 10—The absence of information doesn’t mean, “I chose to leave out information.” There must be positive indication of a planned absence of information.

  • QI Recommendations:

    1. Implement a “No vote” or “Abstain” option for the voter to select to confirm that voter intent was to abstain from voting in a race, especially for voting methods prone to high degree of failure to sense a voter selection. This is the only way to assure that a “No vote” (undervote) was intentional.
    2. The term “residual vote” means all uncounted votes, undervotes (whether the voter intended to abstain or not), spoiled votes and overvotes. Create a new term, “defective vote,” or redefine the meaning of “residual vote” to include only defective votes (unintentional undervotes, spoiled or overvotes). This is a more meaningful metric for measuring voting process integrity.

Election voter verification
IQ principle 11—No matter how well defined or how accurate and complete, data is poor quality if it isn’t available to processes when required. Processes should have access to the authoritative source of data, or downloaded or printed data from that authoritative source on a timely basis.

  • QI Recommendations:

    1. Jurisdictions that had election-day registration problems should analyze causes of missing information. If printed lists are used, ensure they’re from the correct source and are made after all updates have been applied. If electronic files are used at the polling place, assure they’re downloaded with current data. The best solution is to have secure access to the single authoritative voter registration database.
    2. Allow “provisional” voting to allow properly registered voters to cast a vote, pending verification. Provisional ballots should have appropriate identity attributes to allow next-day verification. If the provisional voter is found to be eligible, their vote should be included.

IQ Principle 12—“Haste makes waste” is true in election-day processes. When voters have to fit voting into their workday schedules, they have a tendency to rush, increasing opportunity of error, or they leave without voting. Polls tend to have rush hours that likewise contribute to operational errors.

  • QI Recommendations:

    1. Make election day a national holiday. This reduces the pressure on poll workers and voters.
    2. If voter lists are used at polling stations, an additional scheduling technique can be used to suggest staggered times for voters based on the second letter of their last names. For example, names in which the second letter was A, B, C, or D might have a suggested voting time of 7–10 a.m. Persons with names Bach, Talbert, Edison and Udall, for example, would be encouraged, but not required, to go to the polls during that time to even out the demand.
    3. Pass out voter instructions and sample ballots as people wait. Provide a videotape on continuous loop monitor(s) with on-the-job training that shows someone voting and validating their ballot. This should show problems the voter might encounter with the voting method used, and how to prevent or correct them. Provide audio headsets for different languages.
    4. Make expert poll workers available for people needing additional assistance, with a sample voting booth to “practice” the voting process. This should be required for first-time voters.

Voting process
IQ Principle 13—Standards, including process and data definition, and technology standards are required to produce consistent results. Defining the election process is currently a state’s right charged to local jurisdictions. The more that processes and technologies can standardize across the 3,140 county jurisdictions, the more consistent the results of election processes and recounts.

  • QI Recommendations:

    1. States should create standards for technologies and process guidelines with the most error-proofing capabilities. States should have the same standards in every voting jurisdiction for state elections.
    2. The Federal Election Commission should promote quality standards and guidelines for federal elections. The FEC should provide training with such guidelines, standards and error-proofing techniques so that all jurisdictions have access to the best practices and quality improvements.

The existence of a recount process isn’t because of close elections, but because the voting and vote counting processes are defective.

IQ Principle 14—A recount is an assessment process. It’s inspection of ballots to ensure that the official vote-counting process recorded an accurate vote total result. The best way to deal with recounts is to design quality in to minimize errors that require recounts.

Currently, there’s no process assessment accuracy of a cast ballot because of the secret ballot requirement. However, optical technology that’s scanned while the voter is present allows voters to fix unintended undervotes (could not read the mark) or overvotes (did not fully erase one mark). This should be used to allow the voter to confirm they marked the intended choices.

In e-voting, error-proofing can prevent a voter from voting for more than one choice in a race. A verification screen can summarize the selection choices for all the races one voted in. The machine can—and should—produce a paper, human-readable summary of the voter’s choices that can be verified and used for recounts.

  • QI Recommendations:

    1. Design voting systems with voter-friendly presentation and error-proofing techniques. Design screens and forms that prevent overvoting. Even the use of optical technology with the one percent overvote rate in Florida (2000) would produce 1 million overvotes without other error-proofing techniques applied. Many direct recording electronic (DRE) devices that are designed like mechanical lever machines lack voter-friendly features. E-voting systems should be tested for usability and intuitiveness with voters.
    2. For e-voting, design an independent audit capability and recount capability into the system. This entails a single human-readable/machine-readable ballot that will be used for sampling to test the accuracy of the electronic recording results and will be used for any required recount. This machine-readable audit and recount record must the same record as the human-verifiable record. No printed choices for selected candidates coupled with a bar code to electronically read, for there can be a mismatch between the bar code and the candidate. It isn’t possible to perform a recount by re-reading the machines’ aggregated vote count. This is only a confirmation of the machine count display. A recount requires the ability to re-total the number of counts from the individual legal ballots.
    3. Use this human-readable ballot for voters to verify it represents their intended votes before leaving the voting booth.
    4. Measure the accuracy of precincts by selecting a statistical sample of audit ballots, manually count the totals and compare to the electronically reported totals. If the results are outside the confidence interval, a full recount must be made.

The Tru-Vote Voting System has this feature, creating a printed receipt that is verified by the voter and becomes the legal representation of their votes. This audit capability assures voters that their votes are recorded and can be used to identify potential voting irregularities.

Election process improvement
To solve any problem, it’s important to apply a sound improvement technique such as the Shewhart Cycle of plan-do-study/check-act (PDSA or PDCA) for process improvement. The Plan component must analyze and identify the root cause(s). Without understanding the root cause, one may only attack the symptoms and fail to accomplish permanent quality improvement.

IQ principle 15—The various problems exposed in the 2000, 2002, 2004 and 2006 elections confirm chronic problems in U.S. election processes. PDSA is required to get to zero defects. Sample precincts with different technologies as prototypes should be selected for full and rigorous PDSA improvements before the 2008 elections. Prototype improvement initiatives can be overseen by neutral quality improvement experts, or election commission personnel trained in quality improvement methods. Included in this are:

  • Conduct cause-and-effect analysis to understand root causes. A simple, effective technique is to ask, “Why?” until your get to the root cause.
  • Define specific improvements that eliminate recurrence of defects (failure to know a voter has already registered, unintentional undervote or overvote, etc.)
  • Conducting the improvements in a controlled test environment using real voters to ensure improvements eliminate the defects without creating negative side effects. This can be done at places where crowds of typical people are gathered. The irony of the 2000 Palm Beach butterfly ballot was that this ballot design represented a “planned improvement” to increase the print size for ease of readability. Unfortunately, the “negative side effect” it produced threw the election into chaos.
  • Implement the improvements and put them in control for the next election. Train ballot designers and election workers in the new procedures. Share experiences among all jurisdictions for lessons learned and best practices.
  • QI Recommendation:

    1. Design tests to measure accuracy of votes in which the voter actually cast a vote for the intended candidate. This must include having the voters record their votes and then verify that the actual vote produced—whether punch card, optically scanned ballot or electronic vote—was what was intended. Conduct this in a way to maintain the voters’ right to a secret ballot.
    2. Have quality personnel observe registration, voter verification, voting and counting and recounting processes, to help analyze problems. Attention should be paid to instances where voters request an additional ballot because of spoiling the first. Quality-trained observers who are not part of the traditional voting processes can see problematic procedures that seem perfectly normal to those who have “always done it this way.”
    3. Provide voter satisfaction cards with standardized questions and places to allow for complaints and suggestions for improvement. This should be a regular practice for every voting precinct, regardless of whether this is a sample voting-reform study. Voter-satisfaction cards should be standardized so that the data can be analyzed across a broad set of voting environments. Analyze key quality indicators centrally that will allow general continuous improvement recommendations in the next election and at the local level for local improvements. This should be mailed in or dropped in a separate box to allow independent assessment and then provide feedback to precincts and election jurisdiction officials.

Election process integrity
Once the election processes are improved, put them in control to sustain the integrity.

IQ principle 16—Once improved, processes must be put in control to hold the gains of the improvements.

  • QI recommendations:

    1. Put regular quality assessments of voting accuracy in place to ensure consistent process performance, especially in e-voting machines.
    2. Maintain absolute control of the chain of custody over the ballots, ballot boxes, memory cards, cartridges and personal electronic ballots that contain the votes and tallies along with the audit ballots that are used for vote recount. This control must be sustained without any breaks from the origination (of printed ballots) and empty memory cards, etc., to assure no opportunity for tampering is possible.

If these recommendations are implemented, voters can all go to the polls and vote with confidence and assurance that they shall have elected their officials by a real majority of the voters.


About The Author

Larry P. English’s default image

Larry P. English

Larry P. English is co-founder of the International Association for Information and Data Quality and president and principal of INFORMATION IMPACT International, Inc. English is an advisor of the Election Assessment Group and author of the article “Information Quality Mandate for Election Reform”, published in DM Review. He also authored the book Improving Data Warehouse and Business Information Quality: Methods for Reducing Costs and Increasing Profits (Wiley, 1999). English provides consulting, education and management coaching to enable organizations to increase business effectiveness through total information quality management.