“… crises precipitate change…
… secretly plotting your demise…
I wanna devise a virus
To bring dire straits to your environment
Crush your corporations with a mild touch
Trash your whole computer system and revert you to papyrus”
As mentioned in our preview posting, it seems that, as humans, it often takes a crisis in order for us to really change our ways or our technologies. Like the Deltron 3030 lyric above, in 2008 we appear to be “reverting toward papyrus” in a significant way in order to administer this general election with increased confidence. Election Data Services has predicted that about 56.2% of registered voters this year will be using paper ballots, up from 49.6% in 2006 and 35.6% in 2004.
Now, the reason for this is not necessarily an ominous virus bringing down the computer systems of the nation’s election administrators, however the threat of misdoings does exist with current systems, as many security and computer science experts have demonstrated, published, and attested to.
We’ll be posting a couple more installments in the next few days to follow this topic of the election and the voter experience, beginning today with Part 1: Voting Technology History. Prior to Election Day, we will follow up with:
- Part 2: Voter Needs, Usability, and the Larger Voting Process
- Part 3: Trends and the Future of the Voter Experience
Introduction
As recently as a few days ago, reports of problems in the 2008 general election began to emerge regarding direct recording electronic (DRE) voting systems in West Virginia which “switched” or “flipped” some voters’ selections onscreen before their eyes. These types of problems were most likely due to miscalibration of the touch screen systems – possibly from poor poll worker training or a lack of proper procedures for double-checking equipment prior to opening the polls – rather than computer viruses or other devious hacking of the machines.
It’s hardly comforting either that standards and guidelines for voting machines are weak and voluntary. Lack of robust engineering practice, testing, and certification also contribute to machines that are far from failsafe.
But, let’s face it. Vulnerabilities have always existed in our voting systems, whatever type they may be. And there will always be those who try to manipulate or game the system to their advantage.
It’s not just political parties, special interest groups, private enterprises, or powerful individuals who might be guilty. Even you, the innocent voter, may take advantage of loopholes in order to influence election outcomes (for example, maybe you moved this summer to California, were previously attending graduate school in Illinois for two years, but prior to that lived in Florida, where your driver’s license is from… and now you are voting absentee in the swing state of Florida because you believe your vote for Obama will have more impact there. Anyhow, I digress…).
According to Douglas W. Jones – a computer science professor at the University of Iowa who has testified many times before Congress as an expert on voting technology – despite the different types of voting technologies we’ve used in the last couple centuries, “when you factor in real-world variables, like ease of use and proper administration by poll workers, accuracy ends up being similar,” about one error in 10,000 votes.
So, let’s take a quick look back at where we’ve been.
From the Voter’s Mouth to the Voter’s Fingertip
Prior to the mid-19th century, voting was mostly done orally. A voter swore an oath to a judge and then stated his choices (in public, mind you) to election clerks who kept a tally. Partisan paper ballots or “party tickets” also existed in various forms throughout the 19th century. These tickets could be acquired before the election and listed only one party’s candidates. If you were okay with a straight party ticket, you simply stuffed your ticket into the ballot box without making any additional marks, or you could cross out those you didn’t want to vote for and write in who you preferred instead.
These ballots were neither standardized nor provided by the government’s election administrators, so voter fraud in the form of ballot box stuffing was a common problem. You could even write your own ballot on your own paper if you wanted! Other issues with this system included legibility of candidate names and the lack of voter privacy (since each party ticket looked different, in its color or paper type, and you could be seen with it in your hand).
Party Ticket (left) and Australian Style Ballot (right) (source: Smithsonian)
Eventually, the “Australian” style paper ballot – printed by the government and distributed at the polling place with all candidates’ names pre-listed on them – was adopted in the U.S. around 1888. This kind of paper ballot was essentially the same kind of paper ballot we still use today, with all the contest options listed and a circle, box, or arrow to fill in or complete in order to make your selection.
With the introduction of Australian ballots, one old form of fraud (casting multiple votes, or “ballot box stuffing”) became harder to pull off, so corrupt politicians and their political machines devised a new form of fraud – namely, questioning the way in which votes were actually counted.
This problem of subjective interpretation of ballot markings – which can be exploited to commit electoral fraud – continues to pose potential problems, not just with paper ballots but even with hanging or dimpled chad on punch cards, as we know too well from Florida in 2000.
Coincidentally, it was around the same time the Australian ballot became adopted in the U.S. that the first mechanical lever machine was first used, which essentially automated the vote tabulation process. The Myers Automatic Booth lever voting machine was first used in 1892 in Lockport, NY. By the 1930s, mechanical lever machines were installed in almost every major city, and by the 1960s well over half of all votes were cast on these machines.
Mechanical Lever Machine (source: Smithsonian)
Even as recently as the 1996 general election, 21.2% of registered voters cast their votes on mechanical lever machines, and 6.7% of registered voters will do so in the 2008 election (thanks to New York still using many of these older machines). Despite their durability, the problem with mechanical lever machines is that they do not store independent records of each voter’s ballot (they use a counter similar to an odometer on a car) which essentially negates the possibility of performing recounts.
This is where the punch card system came in. Originally used for controlling textile looms then later used for data entry, storage, and computing, punch cards only really took off as an election technology after IBM bought Harris Votomatic, Inc. in 1965. Although punch card technology had some known problems and high-profile critics (most notably, Roy Saltman of the National Bureau of Standards in the mid 1980s) arguing for their abandonment, these voting systems were used by more registered voters than any other type of voting technology for every election from 1986 to 1998 without any major controversy… and then, of course, Florida happened in 2000. But, let’s not get into that…
Punch Card System (source: Smithsonian)
Datavote systems were very similar to punch card systems, except that Datavote systems used a cutting tool and vacuum to clean away material from unperforated cards while punch card systems typically required the voter to punch out a pre-perforated rectangle (“chad”) from the card using a stylus. Because of this difference, Datavote systems tended to be more accurate than other punch card systems. However, Datavote systems were historically used by approximately 8-10 times fewer registered voters than the pre-perforated punch card systems.
After 2000, legislators scrambled quickly to try to prevent another Florida-like debacle from happening, and the resulting 2002 Help America Vote Act (HAVA) represented the first time the federal government truly became involved in election administration – which is a state’s responsibility – by establishing some minimum standards. It also provided the money and impetus for upgrading voting systems, which as a result are now predominantly one of the two major types described next.
Optical mark-sense scanners are a ballot tabulation technology whose origins are in the world of standardized testing (like the SAT and ACT). You probably knew these better as “Scantron” sheets. When first used for counting ballots in 1962 in Kern City, California, they were tremendously unwieldy systems (that one in particular weighed 15,000 pounds). In the late 1960s and 1970s they shrunk down in size and became easier to operate. By the 1996 general election, optical mark-sense scan systems were used by more than 1/5th of registered voters. Their popularity continues to grow today – in fact, even more so in 2008 due to recent backlash against DRE systems. Most absentee or mail-in ballots are of this type.
Proposed Optical Scan Ballot (source: Design for Democracy)
Direct recording electronic (DRE) voting machines – what we typically think of today as electronic or computer voting systems, usually with touch screens – were first used in real elections in 1975 in Streamwood and Woodstock, Illinois. Used by only 12.6% of registered voters in the 2000 general election, they quickly became more widely adopted after the issues in Florida and as soon as funding became available through HAVA. That legislation gave states incentives in the amount of $3.9 billion to replace punch card and lever machines with newer technology. The main vendors of voting machines at that time were aggressively marketing their touch screen machines, pitching their better legibility, fast tabulation, and improved accessibility for the blind, paralyzed, and foreign language speakers. So, without much guidance or a lot of resources at their disposal, state and local governments and election officials began buying them.
DRE Voting Machine (source: Smithsonian)
Unfortunately, these types of DRE machines have witnessed a number of well-publicized problems – particularly during the elections of 2004 and 2006 – including high costs, poor usability, high error rates, lack of testing, vulnerability to hacking, lack of transparency, and often a lack of an auditable paper trail. Consequently, there has been a national swing back toward paper-based voting in 2008, as can be seen in the figure below.
Type of Voting Equipment Available to Registered Voters
Conclusion
A whole lot of change has occurred in voting technology in the past two centuries, and especially since 2000. The predominance of paper ballots for optical scanning machines and the recent problems with DRE machines suggests that the public (as well as election officials wishing to avoid embarrassment) feels most comfortable at this time with the tangibility of paper ballots, which are generally intuitive and more familiar to the average voter than touch screen DREs.
Regardless of the type of voting technology used, no system is perfect, and errors will occur. Sure, some technologies are probably more reliable than others, but we think the key point here is that the voting technology used for marking and casting a ballot represents only one small piece of the voter experience in the larger voting process. A lot of excellent work has been done already on ballot design and usability guidelines for that part of the experience occurring on or around Election Day.
Unfortunately, not enough of that guidance has been heeded to date. With today’s DRE voting machines, we continue to see some of the common design mistakes that cause people to have bad experiences with software:
- Not speaking the users language
- Lack of appropriate feedback
- Lack of consistency
- Not providing the user enough control
- Bad error messages
- Not being efficient to use
However, equally (if not more) important to the functioning of elections are the people, processes, and infrastructure that also comprise the “election ecosystem.” We’ll talk more about that, as well as some of the relevant social and communication technologies that are changing the voter’s experience before and on Election Day, in our next installment.
Stay tuned for “Part 2: User Needs, Usability, and the Larger Voting Process”
(Special thanks to Douglas Jones’s “A Brief Illustrated History of Voting” and the Smithsonian’s “Vote: The Machinery of Democracy.”)