Elections as Rocket Science and Brain Surgery

270px-STS120LaunchHiRes-edit1Finally some discomfort with and discussion of the idea of ambiguity and the notion of accuracy vs precision in elections is beginning to surface. As the purpose of this blog is to trigger thinking and reflection on the administration of elections, I hope that readers will feel free to disagree, agree, pose questions, answer questions and generally engage professionally and intellectually in this forum. With that in mind, I want to respond to some of the questions and points raised in the various comments in this blog.

We continue, as election geeks and critics, to struggle with the idea that a “good” election may not be an error free election and that an accurate election may not be a perfect election. Bill Kelleher points out that what is a “good” election for one person may be a “bummer” for another. That point is well taken but goes to the core problem in evaluating the “goodness” of an election. The current partisan political environment, enflamed by the media and political parties, wants to judge the quality of an election by who the winners are. In contrast, as professionals and scholars, we should evaluate an election as a process and by its ability to provide an unambiguous outcome.

In a previous discussion of accuracy vs precision, I used the example of rocket science so let me offer a spacefaring example as an illustration. While NASA suffered two tragedies during the three decade lifespan of the Space Shuttle program, it also had 133 successful missions. Success in this example is defined as launching into space and returning safely to earth which, when one thinks about it, is a remarkable accomplishment. NASA engineers knew that there were problems during several of the missions (O-ring blow-by and heat shield damage and missing tiles for example) but the existence of these issues did not cause the missions to be considered failures. Only when the issues prevented the safe launch and return to earth were the missions considered tragic failures. Using a medical example, a surgery is considered a success when the tumor is safely removed and the patient recovers. Pain during recovery or the presence of an uglier than necessary scar does not render the surgery a failure.

Elections should be similarly evaluated as rocket science (shuttle launches) and brain surgery (tumor removal). Did the election process yield an accurate and unambiguous outcome despite any issues or problems, real or perceived, which may have occurred? If yes, then the election, like 133 shuttle missions and countless surgeries must be considered “good.”

This standard automatically begs the question raised again by Bill Kelleher “What is ambiguity?” We know that during several shuttle missions, NASA knew that the heat shield was damaged and it was uncertain whether or not the shuttle would return safely. The situation was ambiguous until the shuttle safely returned to earth. Once the ambiguity was resolved, the mission(s) were considered successful. Until there was some means of demonstrating that an error or issue would not affect the outcome, the result remained in doubt and ambiguous. Interestingly in elections, even when election administrators successfully “land the shuttle” there are some who cling to and make an issue of any issues or errors that occurred during the “mission.”

Stay tuned for further discussion on ambiguity.


[photo courtesy of Wikipedia]

Federal Election Reform and the EAC, a case of “Ready! FIRE! Aim!”?

All indications are, a decade after the passage of the Help America Vote Act (HAVA), the federal government is again seriously ready to tackle election reform.  Clearly the President has reform on his agenda based upon his remarks in his Inaugural Speech when he said, “…our journey is not complete until no citizen is forced to wait for hours to exercise the right to vote.”  In Congress, a half dozen election reform bills have already been filed.   The Nation reports that democratic leadership has made election reform a legislative priority in both the House and Senate.  After years of election legislation at the state level, which notably has given us Voter ID and stricter voter registration regulations, the feds appear ready to weigh in.

Without offering an opinion on the substance of any of the proposed legislation, it seems to me there is something out of sequence in the process.  Congress, through HAVA, created the US Election Assistance Commission (EAC) as a federal agency to address on-going election issues in the wake of the 2000 election.  Congress was hesitant to grant much responsibility to the new agency thus contributing to the Agency falling short of its promise and potential.   For more than a year now, the agency has been languishing with severe budget cuts, without any Commissioners or Executive Director and has not taken any substantive actions.

HAVA limited the role of the EAC to dispersing and accounting for federal funds for new voting technology, setting voting system technical standards, selectively granting research funds for a limited set of election topics and acting as an information clearinghouse.  HAVA funds have been distributed, spent and audited, although the success of the programs is uncertain.  The EAC has created a voting technology bottleneck by failing to issue standards in a timely manner and by encumbering an already lengthy voting system certification program.  Many EAC funded research projects have not been released because of partisan squabbling over the findings.

The original Commissioners took a timid approach and permitted the culture and personalities of the national elections community in its hiring decisions and in the way it chose to fulfill its charter.  The record of the EAC has been disappointing to supporters and critics of the agency alike.

Nevertheless, it seems reasonable to me, as Ready! AIM! Fire! implies, that Congress and the Administration should look to reforming, re-chartering and re-staffing the EAC as the first step in any federal election reform.  As election reform has proven to be an important and persistent national issue, it makes sense to grant broader authority to the EAC rather than disband it as has been repeatedly proposed.

The EAC is not the authoritative voice representing the functions of and policies for conducting elections in the US.  Imagine if Congress consulted with all 50 states and their National Guard bureaus and professional organizations as the primary source of input and expertise into decisions regarding national defense rather than the federal Department of Defense.   Presently, consulting with the states and professional associations is what Congress does for federal election reform.

It seems reasonable that Congress would look to reforming and re-chartering the EAC to be the authoritative source of input for Congress, regarding election reform.  To do otherwise, would be short sighted and nonsensically juxtapose the well known military command “Ready! FIRE! Aim!”

Stay tuned


Photo courtesy of Reuters/Lucy Nicholson

Thinking Outside the Lines

“Our journey is not complete until no citizen is forced to wait for hours to exercise the right to vote.”

~President Barak Obama, Inauguration Speech, January 21, 2013

When President Obama repeated his election night remark that the nation needs to solve the problem of long lines at the polls in his inauguration speech today, I could hear the collective rolling of eyeballs by election officials across the country.

Every election administrator, at one time or another, has received complaints from voters about the length of time it takes to vote and the frustration engendered by waiting in a line to vote. Old news, right?  Based upon the President’s comments and pundit commentary, lines at the polls have become new news and a likely subject of federal debate and legislation.  I believe that this is unfortunate but not for the reasons one may expect.

I have come to conclude that there are really two issues here.  The first issue is the perception that voting is slow and lines are long.  The second issue is the reality that long lines routinely disenfranchise voters who can’t or don’t choose to wait in lines.  Election officials tend to frame the issue as the former and voters, interest groups and parties frame it as the latter issue.  Paradoxically, both perspectives are accurate; part of the problem is perception and part of the problem is reality.  Lines at the polls are another textbook case of the introduction of ambiguity into the outcomes of an election.

The persistence of complaints about waiting in lines to vote, across time and jurisdictions, has it origin in three sources: 1) the paradoxical perspectives on the nature of the problem and official tone deafness to the issue, 2) the inability to measure and quantify what constitutes a line and what constitutes an unreasonable wait, and 3) the failure to recognize that the problem is adaptive and systemic and not a technical problem with a technical solution.

Perspectives.  The reports of lines and delays on election day are usually unverifiable, anecdotal and tend to be reported second hand.  These reports are also made through the media in the midst of voting on election day.  These sensationalized reports predispose voters to show up at the polls with the expectation of having a problem or discourage them from showing up at all.  When there are delays, they are usually isolated to a small number of locations but reports and allegations often generalize lines and delays to be jurisdiction wide.  From the election official’s chair,  one sees sporadic and isolated issues (including lines) arise and get resolved throughout election day.  Voters, parties, candidates and interest groups hear only the media’s reporting or after the fact complaints.  Is it then any surprise that election officials have a different perspective on the issue?

Definition and situations.  What is a line? Our concepts must be reasonable and consistent.  I heard a recent complaint by a voter (who contacted the media) that he had to wait more than an hour to vote.  When asked for details, it turns out the he showed up at the poll enroute to work at 6:00 a.m.  The poll opens at 7:00 a.m.  He did wait an hour to vote but that wait does not indicate any kind of problem at the poll for him or other voters.  In another situation, the first day of early voting – two weeks before election day, several hundred voters wanted to cast their ballot early.  It took many of the voters an hour to vote that first day.  When a voter chooses an alternative mode of voting and there are other locations, days and times to vote; does an hour constitute an unreasonable burden on the voter?

In a general election, a jurisdiction with a bond proposal on the ballot failed to inform voters or provide them with information with which to make a decision.  Most of the voters saw the lengthy bond language for the first time in the voting booth and spent 20-30 minutes, in some cases, to read and consider the bond.  This resulted in a line of voters, who had already been checked in and who had a voting card, waiting for a voting machine for 45 – 60 minutes.  In another election, I observed a line at an early voting location that wrapped around the building but only one or two of the 15 voting machines was ever used at the same time.  The wait was at the check in table, getting voters signed in, not at the voting machine.  On another occasion, I heard a complaint that some voters were being processed immediately while others had to wait a half hour or more to vote.  Instead of some voters being favored over others, that location had a cumbersome method for processing provisional voters.  The delay was limited to voters whose eligibility could not be immediately determined.  Rather than hold up everyone, these voters were diverted to a separate table for processing.  To be fair, many cases in which I have observed or heard of delays in voting were not as nuanced; there was, in fact, a resource or process issue that created delays.

All of these situations represented voter delays but all the situations are not equal in their cause, their solution or, most importantly, in their impact on an individual voter.  At times, when faced with such reports, the issue of lines is willingly acknowledged and investigated by officials intent upon finding a solution.  Sometimes officials deny there was a problem and/or that complaints could not be verified.  Sometimes officials acknowledge a problem but claim they are resource constrained and can’t do anything about it.

Adaptive and systemic problem-solving.  So I pose the question once more, what is a line?  The absence of definitions, criteria, standards and without data or reliable observations, the question cannot accurately be answered.  Similarly, without this information, even the best-intentioned and conscientious election officials can’t do much more than apply band aids and short term fixes.

This leads to the third source of the persistence of the issue of voting lines and delays and the reason I think it would be unfortunate if the federal government were to mandate an ultimate, silver-bullet solution.  The problem is not a technical problem.  Mandating additional machines will not solve the problem when it is caused by check-in processes.  More workers won’t solve the problem when there is a shortage of functioning equipment.  Neither will solve delays that are caused by legislative requirements (voter ID), language barriers, election schedules, ballot crowding, lack of voter information, voter mis-information, unsatisfactory polling facilities and so on.

The problem of lines and voting delays is an adaptive and systemic problem that can only be resolved by solutions which acknowledge a willingness for change, which address system-wide variables and which are based upon facts, data and measurable outcomes.  The solution for each jurisdiction is inherently different as the causes and the magnitude of lines differs from large to small jurisdictions, from urban to suburban counties and from jurisdictions with permissive absentee and early voting regimes to those who rely mainly on election day for voting.  A one-size solution, whether imposed by the federal or state government, will fail and further contribute to the on-going problem of election day lines and the potential disenfranchisement of voters.  In allowing others to solve local issues, we will continue to suffer from the ambiguous (mis)perceived impact of lines on the outcome of elections.

Most of the recurring challenges and problems faced by election officials (and public administrators generally) are adaptive and systemic problems which can only be resolved by long-term solutions.  These types of problems are highly resistant to technical fixes like those described above.  There are a range of approaches, collective and individual, for addressing issues systemically to achieve specific outcomes and foster continuous processes: Lean, Six Sigma, TQM, TOC, Adaptive Leadership, etc.  Each approach has its strengths and I have found are used best in combination to develop quality management and learning environments.

It would behoove all jurisdictions to develop competencies and practices of solving problems systemically, using data and  performance metrics, and by developing a willingness to do things differently to get different (and better) outcomes.   In fact, adaptive leadership and continuous improvement approaches in election administration are the only way to eradicate ambiguity and improve the “goodness” of elections.

Stay Tuned.


January 23, 2013 Update:  Doug Chapin in today’s Election Academy blog post discusses a report issued by the Orlando Sentinal on the potential impact of line in Florida during the 2012 election.  I agree with Doug’s analysis when he says:  “These numbers are pretty spectacular (and not in a good way), and yet it would be helpful to have a little more detail on the conclusions than a very short – and very general – methodological link so we can dig into the analysis.”

Professor Allen’s estimate of the number of voters that failed to vote as a result should be looked at with skepticism until other scholars have the opportunity to review the data and methodology.  Hopefully Professor Allen will make the data available.

While this study describes the problem with delays in voting and suggests the potential magnitude of the issue, it offers little in terms of solutions.  Because such lines were not uniformly experienced across precincts in the state, the next step would appear to be a study of the cause(s) of lines in some places and not in others.  Such a study would provide election administrators the empirical data they need to craft adaptive solutions to the problem.

Ambiguity and Transparency in Elections

This morning, Doug Chapin blogged on the recent decision of the U.S. Court of Appeals for the Third Circuit, based in Philadelphia, in its opinion PG Publishing Company, Inc. v. Aichele.  The case involves media access to polling places on Election day to report on the implementation of the state’s new and controversial voter ID requirement.  Election officials blocked access to polling places to reporters from the Pittsburgh Post-Gazette.  Pennsylvania law contains limitations of who may be in the polling place on Election day and the media is not one of those granted access in the statute.  In the decision, the court ruled that the media had no special right which would grant them access to the polling place. (for a more complete  summary and analysis of the case, please refer to Doug’s Election Academy blog.)

I mention this case because it represents an example of the type of ambiguity that is often inadvertently and unnecessarily introduced into election administration.  My “No Chads” definition of a “good” election argues that anything that adds ambiguity to an election undermines the credibility and “goodness” of an election.

Pennsylvania election officials were on sound legal footing when they made the decision to use the statute to ban the press from the polls.  An argument could also be made (and I am sure it was) that it made good administrative sense to limit the potential disruption of the polls by banning the press. 

So then, what is the problem and where is the ambiguity?  Election administrators understand that election statutes can be interpreted literally, permissively, restrictively or ignored altogether.   (One would be surprised how many outdated or contradictory code sections exist in every state’s election law which are ignored out of necessity.)  The need to interpret statutes based upon the facts of a given situation is an implicit, and sometimes explicit, expectation in the code and of a legislature’s intent.  Election administrators are granted a great deal of discretion in how they apply and implement the law.  Administrators often deny or are reluctant to acknowledge the degree of discretion they actually have preferring instead to defend their policies by claiming the ministerial nature of their jobs.  Put bluntly, we, as election administrators, often can and do hide behind statutes to defend our policy decisions rather than defend the policies on their merits.

There is a large body of political science and public administration literature that studies the different administrative frameworks and the role that discretion plays in elected and non-elected administrators performing the people’s business.  The classic example that analyzes the surprisingly broad discretion of public officials is Lipsky’s well-studied “Street-Level Bureaucracy.” (Lipsky 1980)  This scholarly understanding is sometimes at odds with the way administrators describe the way they perform their duties.  It is widely accepted among political scientists that discretion plays a large role in decision making by administrators which makes the question not “Do administrators have discretion?” but “How should administrators use the discretion they have?”  This latter question has been taken up as an ethical question for public servants by John Rohr. (Rohr 1998)  The ethics in the use of discretion will be taken up in a future post.

Voting is one of the archetypical communitarian rituals and practices that exist today in American society.  Across the broad nation, in multiple time zones, and on the same day of each even-numbered year; Americans trek to their local polling place to record their voice and fulfill their obligation to participate in governing the republic.  There is hardly a more public event and rite in our society (except maybe the Super Bowl?)  Because of the implications of an election for each citizen and for the republic, citizens want to know, and have a right to know, that the election was properly conducted.   The “Help America Vote Act” (HAVA) and the unprecedented activism, on the right and on the left, surrounding elections since 2000 are evidence enough of the universal desire to have confidence and transparency in our electoral processes.

The stated purpose of the voter identification requirement was to foster confidence that votes were being cast only by eligible voters and to prevent fraud.  There has been widespread concern that, while the requirement may prevent fraud, that it will also prevent eligible voters from exercising their franchise.  Each of these closely held perspectives is legitimate even though each represents a competing priority and objective.  The best resolution, from my experience, for such a situation is transparency.   In the absence of transparency, each faction is free to let their imaginations and conspiratorial theories run unrestrained by facts and reality.  Competing realities are the essence of ambiguity.  Transparency is a cure for ambiguity.  When competing groups both see the same reality at the same time ambiguity is eliminated.  When one side is denied access, suspicion, doubt and mistrust are the inevitable consequence.

The press (and media generally) has always been the “eyes and ears” of the American public even though, throughout our history, its objectivity and veracity is fairly questioned.  It would not be possible for all of us to linger and observe voting at our polling place to satisfy our need to be confident in the election.  Even if we could, our observations would be limited to a single location.  We rely on the media, with all their shortcomings, to make and report the observations on our behalf.  The press is our only practical window and portal for transparency.

I am not questioning why election officials in Pennsylvania used their discretion in interpreting the law to prevent the press from observing the polling places on Election day.  The point I hope to make is that lawful, justifiable and reasonable policy decisions made with the discretion enjoyed by administrators have consequences which can either enhance or undermine the “goodness” of an election.  Simply avoiding bad press, embarrassing disclosures or poll worker error does not make a “good” election.  This point is important for election officials to consider as they plan and administer elections.  The use of discretion by officials and its resulting transparency or ambiguity is the take away point for scholars and others seeking a metric for evaluating the “goodness” of elections.

Stay tuned.


Works Cited

Lipsky, Michael. Stree-level Bureaucracy: Dilemma of the Individual in Public Services. New York: Russel Sage Foundation, 1980.

Rohr, John A. Public Service, Ethics and Constitutional Practice. Lawrence, KS: University Press of Kansas, 1998.



Really Scott? Really?

Based upon feedback on the previous post from someone whose judgment I trust implicitly and whose opinion I value greatly, I want to briefly clarify a point I attempted to make.

There is nothing acceptable or excusable about an election that is mismanaged, rife with errors and which creates doubt about the outcome of the vote. A poorly conducted election in one jurisdiction has immediate short-term and long-term repercussions for other jurisdictions; there is a decrease in confidence by the public, increased scrutiny by the media and legislatures inevitably pass new laws to solve isolated problems of the past. Put more plainly, when errors occur; the public is outraged that their vote may not have been counted, the losers are not convinced they lost, the media has a heyday and political leaders sanctimoniously pontificate and then legislate.

As a result, it is understandable for election officials to conscientiously set a high yet unobtainable standard of perfection and to choose not to see or admit to ever falling short of that standard. The price is too high. Yet the adoption of this seemingly noble and highly responsible standard -perfection- has two paradoxically negative and unanticipated outcomes. First, it reinforces an unreasonable and unattainable expectation, among the public, media and politicians, that an election is only acceptably “good” when conducted without issues or errors. Second, the façade of perfection often adopted by election administrators truncates the feedback loop that is a necessary part of the cycle of learning and improving.

The personal cost of publically acknowledging “learning moments” can be unnecessarily high and painful. The unrecognized organizational cost of ignoring or hiding the “learning moments” is even higher.

Back to the point I want to clarify; errors and mistakes in elections seldom affect the outcomes and should not be considered inherently fatal to acknowledge. The healthy and constructive approach, which I am advocating working toward, is one in which it is safe to acknowledge mistakes and failures for the purpose of learning and improving from them. To do otherwise casts unjustified suspicion on elections and election administrators and inhibits a culture of continuous learning and improvement from which the profession can greatly benefit.

In many ways, this clarifying post is a case study in the learning and improvement cycle I have tried to describe above.

Stay tuned.


Accuracy vs Precision

In my last post, I offered a definition of a “good” election: A “good” election is an election without ambiguity. Some concepts are so broad or complex that they are best defined by what they are not. This is the case, I believe, with elections but further explanation is appropriate. In saying that a “good” election is an unambiguous election is not to say that a “good” election is a perfect election.

There are distinctions we easily make in everyday life such as cleanliness does not require sterility or accuracy does not demand precision. For most things in life, cleanliness and accuracy are all that are needed and expected. Brain surgeons and rocket scientists do, however, expect sterility and precision and they need to know when there is an exception to those states.

Elections are not really brain surgery or rocket science but we tend to think of them in that way in as much as we unreasonably expect the higher standard of sterility and precision over cleanliness and accuracy. No election administrator wants to set an expectation of anything less than perfection. To do otherwise appears lazy, sloppy and irresponsible not to mention politically dangerous. The standard of sterility and precision are sirens songs which lead to false priorities and the acceptance of ambiguity.

I know many hold the view that errors should never be acceptable when it comes to elections. That is an unreasonable standard to which that science doesn’t even hold itself. While recently reading a scientific text on social science research, I came upon a quote that I think is relevant to this topic. The authors, in a discussion of the nature of science, disclose that “…theories do not have to make precisely accurate predictions to be judged as scientifically useful….There is always some degree of error.” (Royce A. Singleton 2010)

Every election- even “good” elections have issues, errors or oversights. I like to call l these things euphemistically “learning moments.” The difference between a baimage014d election and a “good” election is whether or not the learning moments introduce ambiguity in the outcome of the election. Florida in 2000 is the obvious case of ambiguous election results created by chad and the butterfly ballot but ambiguity is not always so obvious.

Here is a list of common situations that often lead to ambiguity in elections:
• Ballot errors and poor ballot design,
• Delays in mailing ballots,
• Last minute changes to polling locations,
• Poorly trained poll workers, inadequate staffing,
• Absence of adequate and documented procedures for key election activities,
• Disregard of documented procedures,
• Inconsistent application of signature verification standards, provisional ballot adjudication and voter intent interpretations,
• Poor testing protocols and testing documentation,
• Inadequate security and chain of custody,
• Weak or absent capability to audit or reconstruct election events,
• Inability to statistically describe and document the election,
• Inability to replicate election processes from one election to the next,
• Inability to consistently estimate or calculate the actual cost and expense of an election,
• Responses to unforeseen election day events.

The presence, itself, of any one or more of these examples does not prevent an election from being judged as “good.” An election cannot be considered “good” when the effect of these situations introduces doubt and ambiguity and raises questions that cannot be adequately addressed.

An election that avoids learning moments and ambiguity by sheer luck and good fortune (rather than by effective procedures and management) is not, by definition, a “good” election.

A discussion of leveraging and overcoming “learning moments” is always in order.  I’d like to hear your thoughts.


Work Cited
Royce A. Singleton, Jr. and Bruce C. Straits. Approaches to Social Research. 5th. New York: Oxford University Press, 2010.

No Chads and “Good” Elections

My first experience with punch card ballots came at the beginning of my election career as the election warehouse manager. One of my first projects was cleaning pounds of chad from the votomatic voting machines. Chad are the tiny, multi-colored rectangles of paper released from the punchcard when a vote is cast. In other words, chad are the result of punching a pre-scored punchcard- every free-ranging chad represents a vote cast.

My next encounter with free-ranging chad was when counting ballots on the electro-mechanical card readers. As cards were sucked into the noisy machines, scanned, and then fed into the outstack hopper, I saw chad shooting into the air, falling to the ground and lodging themselves into the crevices of the card reader. Fortunately most of the chad stayed intact on the punchcard but it was unsettling to see the errant chad all around. Were the loose chad the result of a voter making a choice? Did the machine accidently dislodge them? Were the voters selections accurately read by the card reader? When I asked, the answer was “We don’t know. And we can’t know.” This was my first realization that the outcomes of an election could be ambiguous. Ambiguity is being defined as “a situation in which something can be understood in more than one way and it is not clear which meaning is intended.” (Encarta)

Over the years, like most election officials, I learned to accept the ambiguity. As long I used punch cards, what choice did I have? For better or worse, over-worked and resource starved election officials often feel compelled to accept limitations and weaknesses in their equipment, processes and statutes – and ambiguity from many other sources.Chad Eyeball Then came the 2000 Presidential election and the Florida pregnant, hanging chad and butterfly ballots. Now the secret was out publically. Punchcard ballots could and did lead to ambiguous election results.

From that point on, I began to see ambiguity in an election, ambiguity from any source, as antithetical to a “good” election. In my last post, I promised to offer a definition of a “good” election. This is it: A “good” election is an election without ambiguity.

When anything occurs during an election which leads to a credible alternative narrative about the outcome of the election (that cannot be proved or disproved) the election cannot be considered to be a “good” election. Conversely, an election rife with errors and mistakes but whose outcome is unambiguous and not in doubt is must be considered a “good” election. As if to put an exclamation point on my argument, I came across this headline on electionline.org this morning: “… election officials call liquor referendum results ‘incurably uncertain'”.

As I write this, I hear, in my mind, my administrator colleagues muttering incredulity and I hear my academic friends asking “how does one operationalize ambiguity and what are the variables?” I can’t just throw out the proposition that an election’s “goodness” is measured by the absence of ambiguity and expect it to be accepted as truth without skepticism and challenge. In the next posts, I hope to make a pragmatic yet theoretical and scientific argument in support of this proposition. In the meantime, I welcome your comments and input.

Stay tuned.