Election Reform: Fixing the System
[Sorry, the video for this story has expired, but you can still read the transcript below. ]
MARGARET WARNER: The nation’s voting systems came under sharp scrutiny after last year’s election debacle in Florida. Now, several new studies have concluded that the problems weren’t confined to that state. One report released this week found that while 103 million votes were recorded, another four to six million Americans who wanted to vote were unable to, or had their votes discarded. The report by the Massachusetts Institute of Technology and the California Institute of Technology, blamed these “lost votes” on faulty equipment, mis-marked ballots, polling place confusion and foul-ups with registration or absentee voting.
For more on the problems, and what can be done to fix them, we hear from the authors of two recent nonpartisan reports. Stephen Ansolabehere is a professor of political science at MIT who was involved in the MIT-Cal Tech study. And Norton Garfinkle is chairman of the Institute for Communitarian Policy Studies at George Washington University. The institute finished its election procedures study last month. Welcome, gentlemen.
Professor Ansolabehere, first to you. It was your study that came up with the four to six million lost votes figure; tell us a little bit more about the biggest problems.
STEPHEN ANSOLABEHERE: The biggest problems lie in the area of registration. According to the census, which conducts the current population survey, about 7.5 percent of the 40 million registered people who didn’t vote cited registration problems as their main reason. And those are people who are already registered. So these are people who went to the polls, tried to vote but couldn’t. Fixing these problems seems to us to be paramount. That translates into about three million lost votes.
There are another million-and-a-half presidential votes that were lost because of faulty equipment, confusion on the part of voters of how to use the equipment, failures of the equipment to actually capture the votes and tally the votes properly. In addition, there are many more lost votes down the ballot due to equipment in Senate and governor elections. And then there are problems dealing with polling place locations, hours and operations that lost roughly another million votes: People who tried to vote but couldn’t because of polling place problems.
MARGARET WARNER: Tell us a little bit more about the machine problems. Which machines did you find had the highest error rates versus the lowest?
STEPHEN ANSOLABEHERE: We studied equipment over the last 12 years. Equipment is used at the county level and purchased at that level in almost every state, so there was a lot of variation across the country and over time in which counties were using which kinds of equipment. Surprisingly to us one of the more reliable systems is hand-counted paper ballots, which is the oldest technology used around the country right now. Another very reliable technology, it appears, is what’s called optical scanning. That’s where the voter is given a paper ballot, they mark the paper ballot and then that paper ballot is counted through an optical reader, a little like maybe a bar code reader at a grocery store.
There are other kinds of equipment. Punch cards proved to be pretty unreliable. Most of the viewers have probably seen punch cards in the news in the wake of the Florida debacle. About 10 percent of the voters used electronic equipment. Some of the electronic equipment doesn’t seem to be performing up to the standards that we’d expected of it. That was another surprise in our study.
MARGARET WARNER: And just let me interrupt for a second. By electronic you mean the computerized screens, the touch screens?
STEPHEN ANSOLABEHERE: Well there’s a variety of electronic machines being used. The most common variety is a big panel that has a set of push buttons corresponding to all offices. It’s called a full faced touch panel DRE. There are… The newer DREs are touch screen computers that are not unlike an ATM machine. There are newer models that are coming on the market right now that are pretty much untested. Lever machines did pretty well in the presidential vote. They did the worst in the Senate and governor elections where about 7.5 percent of the votes over the last election cycle has been lost.
MARGARET WARNER: Professor Garfinkle, what would you add to that sort of catalogue of the problems and where the worst problems lie?
NORTON GARFINKLE: Well, I think it’s clear that the worst problems lie with the punch card systems which became so notorious in the Florida recount area. It is true that the lever systems are quite accurate but they’re obsolete. It’s hard to buy new lever machines. And, therefore they don’t afford a really good option. The standard paper ballot systems are used only in 2 percent of the country, and they don’t afford a real reliable option. And as my colleague at MIT has said, the ATM-type system really doesn’t work very well.
That leaves us with a relatively modern paper ballot, fill-in-the-dot optical scan system that is already deployed in 10 to 20 percent of the country and seems to be the clear winner among the available alternatives. It also turns out to be a very affordable system. And it has really three particular benefits. It’s easier for the voter to use. The voter sends the ballot through the scanner. The scanner tells the voter instantly if he or she made a mistake by voting for more than one candidate. And if that’s the case, the voter can ask for another ballot to correct the mistake. The bottom line is this system helps the voter cast a ballot that will be counted accurately. Second, if an election is contested because of a close vote, the fill-in-the-dot system gives us a paper trail so we can verify ballots easily and finally, the fill-in-the-dot system is cost effective. So it’s clear that there is one winner that can be deployed easily and readily, surprisingly easily, surprisingly readily, and cost effectively throughout the country.
MARGARET WARNER: Professor Ansolabehere, I know your report also recommended the use of this, that is, where you fill in a little box and then feed it into the machine yourself; is that right?
STEPHEN ANSOLABEHERE: That’s right.
MARGARET WARNER: Now, what did you recommend with the registration problems, people who found they weren’t registered or somehow couldn’t vote even though they showed up? How do you fix that?
STEPHEN ANSOLABEHERE: Well there’s one simple solution that could be deployed more widely and that’s called provisional balloting. About 19 states use it now. That’s where if you walk into the wrong polling place or there’s a question about your registration, you’re allowed to vote on some elections and then that ballot is stored and your registration is checked afterward. In Los Angeles County, for example, where this system is used, about two-thirds, a little less than two-thirds of all provisional ballots cast turned out to be valid votes. A large part of the registration problems could probably be fixed quite simply with the adoption of provisional ballots.
MARGARET WARNER: Now, you also found though that voters said, you know, if they showed up at the polls and there’s this laborious process of checking against the voter registration rolls and so on. How do you fix that?
STEPHEN ANSOLABEHERE: Registration creates some of the problem of lines because if there’s a question with your registration in most places a precinct poll worker has to get on the phone and call the central office. And that takes that poll worker away from doing other tasks. That only lengthens the lines. So some of the line problems could probably be fixed by fixing some of the registration problems.
MARGARET WARNER: So Professor Garfinkle, how do you… Given that most of these elections are run at the county level — and I think there’s something like 3,000 counties in the country — how does the government go about fixing this? Who pays for it? How do you manage it?
NORTON GARFINKLE: Well, in our calculations, we found that the total cost of installing the new equipment throughout the country would be $1.2 billion. That’s less than 1 tenth of 1 percent of one year’s annual budget, federal budget of $2 trillion. Now, the easiest way to do this, the most straightforward way to do this, the thing that could accomplish the task within the next three months is if the Congress simply allocated a $600 million matching grant program to enable the states then to match the amount of money and then acquire the new equipment that’s needed. It really surprised us that the problem is so easily manageable. This is one of those problems that essentially is now highly visible but small enough to solve.
MARGARET WARNER: Professor Ansolabehere, would you agree that it’s easily manageable and that it can be done at about that cost?
STEPHEN ANSOLABEHERE: Those numbers are about right. That’s about what we calculate. More interesting are some… Is the experience of Rhode Island, which actually leased its equipment and came in with a total lease cost that was less than what they were paying to pay for their lever machines, so it’s even possible within state budgets to pay for the upgrading.
MARGARET WARNER: And finally, you found… Talk to us or explain to us what you found about two other things that voters seemed to like, more absentee voting and also there’s been a lot of excitement about Internet voting.
STEPHEN ANSOLABEHERE: They’re related problems with Internet and electronic voting. Internet voting suffers from problems of disruptions of service. If your computer crashes, if some kid somewhere hacks into the system and brings it down, that could disrupt an election. That’s a somewhat separate concern than the problem that Internet voting shares with absentee voting, which is that Internet voting from home or what’s called remote voting could be subject to coercion or fraud. People could buy your votes.
Absentee voting on demand might also suffer from the same problem. The most sensational court cases involving voter fraud recently have come in the area of absentee voting. Absentee voting over the last 20 years has grown from 5 percent of ballots cast to 15 percent of ballots cast. And we’re not talking about absentee voting for cause, which is if you’re disabled or out of the country, we’re talking about absentee voting on demand or for convenience. And it’s possible to reduce that by adopting systems such as early voting where voting machines are set up so that people can vote before election day.
MARGARET WARNER: Professor Garfinkle, do you agree with that?
NORTON GARFINKLE: The biggest problem with absentee voting is it is the one… The one system that is most subject to fraud, the one that is most subject to influencing voters the wrong way. So while there’s been an increase in its use because of the convenience for the actual voter, the convenience comes at a very high price in inaccuracy of voting and voting counting.
So I think that the notion that absentee voting somehow is going to be the correct wave of the future is probably completely wrong, and therefore, we have to settle in at least for the foreseeable future, for the next decade or so, with the available options that now exist. Again what’s nice about that is we do have an available option that has been tested, that has been approved by the National Association of Election Systems and therefore can afford an easy and straightforward solution that will restore the sense of legitimacy to our voting system and to the governments that are elected in our voting system.
MARGARET WARNER: All right. Well, Mr. Garfinkle and Professor, thank you both very much.