January 16, 2008
Could 2008 begin any better for a guy whose life’s work focuses on voting?
The first weeks of the new year have brought: cliff-hanging primaries in an historic battle for president; publication of not one but two of his new books on voting behavior; fresh controversy jumping off the pages of The New York Times over the reliability of new voting technologies; and, for good measure, award of a major grant leading to a thorough checkup of Michigan’s voting system.
Welcome to the world of Michael Traugott, University of Michigan professor of political science and communication studies. Although his is hardly a household name among the candidates, office holders, journalists and voters who parade across his vast electoral Petri dish, Prof. Traugott is among a couple dozen academics recognized nationally and around the world as experts on voting behavior — especially as it is influenced by the constantly changing world of technology.
“Anybody who is involved in academic work is interested in doing things that are not only central to the discipline but also have some public impact or make a difference. For those of us who study campaigns, elections and voting behavior, yes, I think this is about as good as it gets,” Traugott confirms.
On the surface, the act of voting seems so simple. You make up your mind then pull a lever, punch a hole, make a mark or touch a screen for the candidate of your choice. Your vote is tallied along with the votes of hundreds or thousands or millions of your fellow citizens, and the candidate who receives the most votes wins. But Traugott sees below the surface, and he knows just how deep and increasingly murky the voting waters have become.
Mall Lab
For three weeks in the summer of 2005, Traugott and a small crew of graduate assistants set up shop in Ann Arbor’s Briarwood Mall to learn how new voting technologies can affect election outcomes. They recruited some 500 mall shoppers, paying each of them $10 to participate. They brought in five of the most widely used new voting systems (plus a prototype of another being developed), most of them put in place to remedy the hanging-chad fiasco of Florida’s 2000 presidential count — the one that began to draw public attention to some of the dirty little secrets about the fallibility of ballot technology.Traugott, along with colleagues at the University of Maryland and University of Rochester who were doing the same experiment in Baltimore and Rochester, New York, constructed a simulated election in which they knew how the “voters” intended (and were instructed) to vote. The participants voted the same way on each machine, even making and correcting deliberate mistakes, enabling the political scientists to measure how far each system strayed from voters’ intent. Other findings included ease and speed of use, as well as voter satisfaction with each device.
Results from that three-city exercise and other studies led to publication this month by the Brookings Institution in Washington, D.C. of the professors’ Voting Technology: The Not-So-Simple Act of Casting a Ballot. The publisher describes it as “the first book to investigate in a scientific and authoritative manner how voters respond to the new equipment.” The collaborators’ findings revealed the good and bad about the systems, including specific features that contribute to greater clarity as well as those leading to confusion and error. Their book concludes by pulling together best practices to guide voting-system manufacturers, ballot designers, election officials, political analysts and voters.
“We didn’t write about specific brands — eventually something will develop that looks like a Consumers Reports for voting machines — but we do talk about the technology that works better and that people can use more easily and are more satisfied with,” said Traugott. “We also talk about the machines’ verification [paper trail] devices, which are problematic because they’re not yet well-integrated into the voting machines. Many of the manufacturers are putting inexpensive printers, like you find on a cash register, into the devices, and they have a tendency to jam and miss some individuals.”
The flood of new, unproven voting technology is largely the result of the chaotic Florida election, followed by congressional enactment of the Helping America Vote Act (HAVA). Among other things, HAVA mandated elimination of lever machines and punch card ballots and provided the states with millions of federal dollars to purchase new voting equipment. State and local election officials, with little experience or research to guide them (prior to Traugott’s studies), were suddenly beset by equipment sellers.
High Marks
Traugott gives high marks to Michigan election officials for using HAVA to move the state to optical-scan voting. Putting the entire state under what is essentially one type of system was all the more difficult because of the decentralized nature of election administration and the autonomy of local clerks in Michigan.The optical scan choice was a good one, he observes, in light of problems other states have had since adopting touch-screen and other technologies that have produced greater numbers of errors and failed at producing credible paper trails necessary for recounts and resolving disputes. Ohio and parts of California are notable examples of states that recently decertified their new touch-screen systems and are now spending millions more dollars to buy optical-scan or other devices.
Given how much he knows about the shortcomings of new voting technologies, you could forgive Traugott if he had become a cynic. But overall, he’s optimistic that election administration and voters can and will adapt.
“[The system] is not totally screwed up,” he said, “but all of the research suggests that one critical issue that hasn’t been addressed very well yet is the need to educate voters about the new technology and to train poll workers in a different way — and it may also require a different kind of poll worker.” He points to the need to understand and address such issues as the computer proficiency gap between young and old and the “digital divide” separating citizens’ access to computers along economic lines.
One thing certain above all else, Traugott says, is that the voting experience will continue to change for voters. “When you buy into the best available or most commonly available technology it means you are essentially signing on for continued change, because there will be updates all the time. So [voter and poll worker] training will have to be modified as the devices are modified.”
The new realities are a far cry from the stability of past voting systems, chiefly paper ballots and mechanical lever machines. “One of the reasons people liked the old lever machine is because it was very straightforward. You could see how you were casting votes and the physical action of pulling the lever down had a direct connection with the notion of making a preference. The problem with them is that they hadn’t been manufactured since the early 1980s and there were no spare parts for them, so it was a disaster waiting to happen. They provided a great deal of stability for a lot of voters across the U.S. over the past 30 years because voters knew exactly what they were going to see when they went to the polls. And now all that is changing.”
Research into these issues, he points out, is hardly an academic exercise. A lot is riding on the answers he and his colleagues are uncovering, no less than the functioning of American democracy. “A big concern with all of this new stuff and the problems of 2000 and 2004 is whether or not people will lose confidence in the electoral system and, as a result, possibly stay home.”
Polling
Traugott’s other new book, The Voter’s Guide to Election Polls, deals with his other major area of expertise: polling and its effect on voting behavior. He wrote it with Paul Lavrakas, an independent consultant, former chief researcher at Nielsen Media Research and former professor at Northwestern and Ohio State universities.The book is written in Q & A format based on the many questions they have received when speaking to groups about election polling. They wrote the first edition in 1996 and have revised the material every four years. Traugott calls the new edition a substantial rewrite to include current issues related to the Internet, computer-generated “robo calls,” cell phones and other evolving technologies.
“We’ve had the rise of Internet polls, which I think are the wave of the future because they will give pollsters certain interesting opportunities they don’t have over the telephone, such as showing video clips. But Internet polls have problems currently in their sampling, because there is no frame or list of people who have access to the web. Various firms are trying different techniques to work with that…but as the penetration of the Internet increases and we can get beyond issues such as the digital divide, then I think this is where the future of polling lies.”
Other noteworthy trends involve the use of cell phones. In Europe, for example, where only the caller pays for a cell phone call, pollsters are conducting surveys of cell phone users. The most interesting of these surveys, Traugott says, involve contacting people and asking them what they are doing at that moment, in order to better understand that group and produce more accurate results. “Think about telephone surveys in the U.S., where we have to operate on the assumption that everyone is at home. That’s one of the major problems of non-response, because Americans are active and mobile and frequently not at home.”
The other big change in election-related polling, both in pre-election and exit polling, is that more Americans are voting before election day. In the 2004 presidential election it was about one in five across the U.S. — and more than a third of California voters. To ensure they are getting an accurate sample of voters, pre-election pollsters must now find out if respondents have already voted. And exit pollsters are having to supplement their in-person interviews with phone surveys.
How does the public view the continued onslaught of polls?
“In general, the public is interested in polls,” Traugott says. “Citizens see them as useful both in communicating their views to other citizens...and to elected officials and government leaders. So overall they are quite interested in polls. At the same time, they have almost no knowledge of the science of polling and there are parts they are skeptical about — they don't understand sampling, they have disbelief that 1,000 people can represent the views of an entire population.”
“At a different level,” he adds, “it’s pretty clear in the U.S. and many other countries that people don’t like to see a lot of polling data about elections near the end of the campaign. They like to be left alone. This also goes for early projections from exit polls on election night.
“One of the things that distinguishes the U.S., however, from all these other countries is the U.S. Constitution and the First Amendment. While a lot of countries have laws that limit the publication and dissemination of polling data at the end of a campaign, that would never pass a court test in the U.S. ”
In order to help journalists and the public better understand political polling, Traugott has become involved in another new national project. His work with the American Association for Public Opinion Research and the Poynter Foundation in Florida has led to the development of a webinar to help train journalists how to better understand and report more effectively on polls for this year’s elections.
No Junkie
Despite being deeply involved in the world of politics, Traugott says he doesn’t consider himself a political junkie by any means.“I’ve seen political junkies and I don’t think of myself as one of them in the way that I understand the term. I’m very interested in politics, but I don’t live and breathe politics in excruciating detail. I don’t really attend to very many blogs — I don’t follow politics constantly on the web. I have a very strong interest in public opinion, so I do visit a blog run by a guy named Mark Blumenthal — it’s called Mystery Pollster — who comments on methodological issues and their consequence for data. But I don’t go there every day.”
Traugott was born and raised in Providence, Rhode Island. His father was a local businessman who worked in sales. Neither of his parents had attended college (although his mother earned a degree later in life). “They had a minor interest in politics and they voted, but nothing much beyond that.”
He chose Princeton for undergraduate studies in engineering. In his sophomore year, calculus convinced him otherwise. Looking for something else, he became involved with computers. It was the early 1960s, and the political behavior movement was gaining hold in academia. “People were collecting more data, especially survey data, for election studies.” The American Voter, a landmark book by University of Michigan political scientist Angus Campbell and others, had been recently published. Traugott was comfortable with computers and quantitative analysis at a time when many students weren’t, and he fit right into the new emphasis on analyzing political data. He became a political science major.
His Princeton years also put him in touch with George Gallup, the national polling pioneer located in Princeton. “I was earning my way through school, so one day I responded to a note on the bulletin board to be an interviewer at the Gallup Organization. By the end of my undergraduate career, I worked as a research assistant for Dr. Gallup.”
Like many of his classmates, Traugott considered going to law school. But his professor advised him to go to graduate school in political science. Gallup told him the same thing. The Ann Arbor campus in the mid 1960s was not only extremely politically active, “it was ‘The Place’ to learn about campaigns, elections, voting behavior and public opinion research from a methodological and substantive perspective,” Traugott recalls. “And so right at the very end of my senior year I got turned around and went to Michigan to study at the Survey Research Center.”
One of his colleagues had a contract to help ABC television with its election night coverage, so over the next decade or more, Traugott also became involved as “a participant observer watching broadcast journalism capture election night coverage. It was at the same time that these big combines of networks and major metropolitan dailies started their own polling operations, and that was of interest to me from a survey methodology standpoint. And then I got asked to help out The Detroit News with campaign coverage, so I wrote a column and did surveys and polls for the paper for several years.”
With his expertise, Traugott undoubtedly could have set up shop in D.C. and made a small fortune running his own polling firm or advising political campaigns. He admits to having considered it, but adds: “one of the great things about the work that I do, and I think the thing that attracts all academics, is that we generally get to control our own research agenda. There’s a fundamental difference between this and working for a client and being real interested in the campaign years and then having to fill in during the non-campaign years doing anything you can. That difference made this life more attractive to me.”
Traugott, who earned his master’s and doctorate degrees at Michigan, has run up a long list of accomplishments over the five decades he has worked with the university’s Institute for Social Research, Center for Political Studies, Survey Research Center and Department of Communication Studies. His single-spaced Curriculum Vitae runs 18 pages (and needs updated), 14 of which recite the books, chapters, articles and papers he has authored. He has served as president of the American Association for Public Opinion Research and is currently president of the World Association for Public Opinion Research.
Maintaining a scientist’s perspective is important to Traugott. “There are different ways that different political scientists approach these issues. Some think it’s very important to be active in a campaign...I have a different thought. I believe it’s very important not to be active in a campaign and to remain objective. Generally, that approach has worked very well for me.”
Michigan Project
His concerns over maintaining public confidence in election systems has led him to team up with the State of Michigan, The Pew Charitable Trusts and the New York-based JEHT Foundation in yet another major 2008 initiative, this one called Making Voting Work. Together with small research teams in Utah, New Mexico and Maryland, Traugott is helping devise a system for conducting the first-of-their-kind post-election audits that will bring to light errors and issues of concern over election administration and voting systems.His earlier conversations with Michigan elections officials led to the foundations’ award of a $125,000 grant to UM for the work. “Michigan Secretary of State Teri Lynn Land has a plan for upgrading Michigan elections in which she talks about audits. We discussed a research project to support her interest in preparing legislation.”
It was only after the Michigan project began to take shape that Traugott learned the foundations decided to fund projects in the three other states, broadening the effort and number of researchers grappling with how to perform effective post mortems.
Not surprisingly, one of the first issues the committee is grappling with is what is meant by the term “audit.” While to some that means conducting recounts, Traugott views it in a broader sense of being able to determine how accurately and uniformly the system operated. That could include such issues as ballot errors, use of provisional ballots, number and location of machines, and many other factors. “Our main goal will be to design a system of indicators that can be assembled and widely disseminated through the media to maintain or increase confidence in the electoral system among citizens.”
Their work, he said, will culminate in a report to be given to Secretary Land for her use in seeking the necessary implementing legislation to begin audits with the 2010 elections.
Looking at all of the changes taking place in his field — and the many more coming in the future — Traugott reacts as any good researcher would, by embracing change and the need to study it.
“It means there’s always work to be done, always interesting problems and interesting studies to perform. I find it all very gratifying work.”
Insert your email (required)
Subject
Comments
Insert email of your friends (required)