Testing Tech for Consensus in a Purple Town
How a radical experiment in participatory democracy came to Bowling Green, Kentucky.
Over the past year, town halls across America have occasionally erupted over hot button issues like health care reform. Rep. Tom MacArthur was shouted down during a five-hour meeting in New Jersey last May; later that summer, a Californian said they hoped Rep. Doug LaMalfa would “die in pain”; and Rep. Ron Blum was called a liar by a prescreened audience in Iowa. A town hall in Bowling Green, Kentucky, last month had none of the shouting or vitriol that made those events national news, but it was the site of something even more elusive: the search for consensus in an increasingly divided nation.
Yes, a radical experiment in participatory democracy brought me from New York City, the towering headquarters of the liberal media elite, to the land of bourbon and barbecue—to Bowling Green: home of the Corvette, Fruit of the Loom, and Kellyanne Conway’s infamous massacre.
On the drive up to the college town, 70 miles north of the Nashville airport, Joe Karaganis described part of their goal as “trying to reconstitute a world of shared facts.” Karaganis is the vice president of The American Assembly, a think tank affiliated with Columbia University, and one of the organizers of the Bowling Green town hall, the penultimate event in a three-part civic engagement process.
For locals, the Bowling Green Civic Assembly was a rare opportunity to publicly raise concerns and air grievances. More than 2,000 people voted at least once in a virtual town hall on Polis, an innovative new survey platform that allows participants to generate their own statements in addition to responding to statements by others. The Bowling Green Daily News estimated 250 people attended the February 22 town hall, and more than 1,000 people tuned in to watch on Facebook Live.
The hunger to engage was there, but I also heard a healthy amount of skepticism about whether anything would come of it.
For Polis and The American Assembly, the process had the potential to be something even greater: a model for bottom-up civic engagement that could be replicated in cities and towns across the nation to find hitherto-unknown areas of agreement, and fuel productive conversations between residents and local government. But although the organizations celebrated the Bowling Green pilot as a qualified success, the model still needs work—and lots of follow up—in order to reap the hoped-for benefits.
The Bowling Green Civic Assembly can trace its roots, through The American Assembly, to Dwight D. Eisenhower and his “fanatic” belief in American democracy, as historian Travis Beal Jacobs has characterized it. During Eisenhower’s tenure as president of Columbia University—before becoming the 34th president of the United States—he came up with the idea for “a great cultural center where business, professional, and government leaders could meet from time to time to discuss and reach conclusions concerning problems of a social and political nature.” The result of such meetings would be, Eisenhower hoped, “some clearly agreed upon truths and observations.”
In 1950, the descendants of a railroad magnate gave Columbia a castle-like estate outside of New York City called Arden House, enabling Eisenhower to realize his vision. He founded The American Assembly that year, and over the next half century it became an influential driver of public policy.
During that time, The American Assembly developed a rigorous and formulaic methodology to discuss issues ranging from inflation and world hunger to nuclear power and alcoholism. As Eisenhower envisioned, experts from business, government, and academia would travel to Arden House for several days of discussion and debate.
Participants were given a syllabus of background reading to do in advance, and during the assembly, they heard talks and panel discussions by the top experts in attendance. Discussion leaders would draft a document summarizing the group’s conclusions, and send it around for comment. The final report would then be published by The American Assembly, with no restrictions placed on republishing the work in part or in its entirety.
“It was a process designed for a kind of policy deliberation that really reflected its era,” Karaganis said.
Times were different during the latter half of the 20th century—bipartisanship wasn’t a punch line; the internet had not yet upended the media and our democratic processes. However, by the 1990s, the relevance of The American Assembly was being undermined by new technologies and shifting norms. For most of its existence, The American Assembly had been one of just a handful of organizations that held public policy deliberations in the country; then suddenly, Karaganis said, “it was one of half a dozen on the Columbia campus.” In 2007, The American Assembly suffered another blow when Columbia University sold off the upstate mansion.
For the past six years, Karaganis has been working to bring The American Assembly into the 21st century—essentially, to help it become relevant again.
“I was hired and have been involved in a variety of attempts to rethink what the mission of the organization is,” Karaganis said. “If we try and preserve the values of the organization in a context in which the methods have to change, what do those methods look like?”
The tentative answer, at least in the case of the Civic Assembly model, is three-fold: To narrow the focus of American Assembly’s work from the global or national to the local or hyperlocal; to supplement in-person meetings with virtual conversation on Polis; and to reverse the top-down paradigm.
“It’s not a handful of experts setting the agenda for a group,” explained Karaganis, “but rather an agenda that emerges out of a much broader conversation.”
Bowling Green is a small, rural city, with fewer than 70,000 residents, but has a more diverse population than most of Kentucky. It’s a college town and—like many college towns—students and faculty are not always on the same political page as the rest of the region. The city has a sizable Muslim population because the International Center of Kentucky has resettled more than 10,000 refugees in the area, including more than 2,000 Bosnian refugees starting in 1993. In 2012, the Center estimated there were more than 5,000 Bosnian-Americans living in Bowling Green.
Last year, Bowling Green was chosen as the site for a study on political polarization and media consumption by Andrea Wenzel and Sam Ford, both fellows with the Tow Center for Digital Journalism at Columbia University. “We selected this area of Kentucky because we wanted to understand what polarization looked like in rural communities and smaller cities where pockets of blue and purple were woven into a sea of red,” they wrote in the Columbia Journalism Review.
They found that national politics had divided the area politically in sometimes surprising ways: “Rachel was a registered Democrat who happened to vote for Trump,” they write. “Her experience was one of many that illustrate the messiness belied by color-coding of states and counties as ‘red’ or ‘blue.’ While the area leaned Republican and conservative, almost everyone knew someone from the ‘other side.’ Numerous participants reported having difficult interactions from social media bleed into their offline lives in ways that had not been the case prior to the 2016 election.”
It was at a workshop led by Wenzel and Ford that Karaganis and his colleague Stephanie Sung met Joe Imel, the director of media operations at the Bowling Green Daily News, and proposed trying Polis out in Bowling Green. It was a platform that The American Assembly team wanted to experiment with in a small American city with a local partner, and Imel was enthusiastic about doing something to engage readers in civic life.
“I’m always looking for ways to engage our community,” Imel told Civicist. “The newspaper, I think over the years, with the disruption that’s kind of blown up the media landscape, I wanted to get back to our roots of really letting people know what’s going on. These days more people can tell you…more about national government than they can about their local government.”
The Bowling Green Civic Assembly was an opportunity to promote a conversation about local issues, and rise above divisive national politics.
The Bowling Green Civic Assembly opened on February 4 with a 10-day virtual town hall that asked participants: “What do you believe should change in Bowling Green/Warren County in order to make it a better place to live, work and spend time?” Participants were encouraged to both submit suggestions of their own, and to vote “agree” or “disagree” on statements submitted by others. The goal was for the survey to resemble a conversation more than a poll.
“Normally surveys talk at you and you don’t get to talk back,” Darshana Narayanan, a Composites Collective consultant working with Polis, explained to the 250 people at the in-person town hall. “With Polis, you hear and are heard.” (Disclosure: Narayanan has written for Civicist in the past, and will be working on programming for the Civic Hall-affiliated event, Personal Democracy Forum 2018.)
The Polis platform responds to participants. If a particular statement was frequently passed on without a response, it was shown less often, under the assumption that it must be irrelevant, poorly articulated, or otherwise inscrutable. For example, the statement “I still have to drive out of town to eat Golden Corral” was passed on by 15 percent of Bowling Green participants, and only voted on by seven percent, and in the end received less engagement than other statements submitted at approximately the same time.
Just over 400 people contributed a statement to the Bowling Green virtual town hall, while 2,025 people voted, casting 225,450 votes in all. Participation picked up towards the end of the 10-day period, leading organizers to wonder if they should have let it run a bit longer to allow more time for information about the poll to spread by word-of-mouth.
After the polling period closed, the Polis team analyzed the results to identify majority and minority positions, points of consensus, and points of divisiveness, and then shared the results online so that both the creators of the poll, as well as the participants, could glean meaningful and actionable insights from those results.
The way Polis organizes results allows you to see the differences between different groups, and within the groups themselves, on every last statement. The results of the Bowling Green Polis could be roughly divided into two groups, one that tended to be more liberal, and one that tended to be more conservative. If you were highly likely to vote in favor of a fairness ordinance, a law that would prohibit discrimination on the basis of sexual orientation, then you were also more likely to vote to treat drug addiction like a health problem rather than a criminal justice problem, and you were subsequently identified as group “A.” However, there were still divisions within the “liberal” and “conservative” ideological groups; the “B” group, although almost universally against instituting a fairness ordinance, was divided on issues like the legalization of marijuana, aid for the homeless, harsher sentencing for opioid users, and school funding.
Polis also arranged the results visually in several different graphs. One showed the statements, depicted as dots, along a line from least to most divisive. With one glance it is apparent that the vast majority of statements are on the far left, indicating more agreement than disagreement, and that truly divisive statements are far more rare.
“We set out to build a communication system that would handle ‘big’ and stay coherent,” cofounder Colin Megill explains in Polis’ mission statement. “We wanted people to feel safe & listened to, and we felt it was of the highest importance that minority opinions be preserved rather than ‘outvoted’.”
The most established use of Polis is vTaiwan, the collaborative, deliberative, online-offline process used to craft cyberpolicy in Taiwan, which Liz Barry covered for Civicist in 2016. However, the platform has also been used by a newspaper in New Zealand to discuss the obesity epidemic with their readers, and by MarchOn, the Women’s March spin-off organization, to solicit ideas on how to “convert the moment of the women’s march into a movement.” The first Polis pilot in a U.S. city took place last year in Fairhope, Alabama.
However, unlike the process in Fairhope, which was led by the city mayor’s office, the Bowling Green Civic Assembly was led by a local, family-owned newspaper, the Daily News. Working with a media partner had several advantages: a built-in user base; established communication channels to get the word out; and impartiality—a paper is less likely to want to censor comments critical of local government than an elected official might be.
The Town Hall
The car dealership-turned-events-venue was still half-dressed up for a party when the doors opened for the town hall. Cream-colored drapes hung from the industrial ceiling, and a heart-shaped balloon was caught in the rafters. Many of the earliest guests were wearing badges that identified them as candidates for office. One woman came in wearing a sequined American-flag vest.
The first hour was an informal meet-and-greet, to give people time to review the results of the online poll, several copies of which were taped to the back walls. In addition to the results, and in the spirit of transparency, the organizers had also printed out all of the comments that had been submitted but deleted by moderators, either because they were duplicates of statements that had already been posted or personal attacks on individuals or businesses (names were omitted). Out front, a local food truck called the Cheese Wagon peddled grilled cheeses to hungry attendees.
Imel and Sam Ford—the Tow Fellow and Kentucky native had been hired as a consultant by The American Assembly—had lined up 13 local officials to address the audience on the three broad topics that had emerged in the online town hall: quality of life, including transportation, property maintenance, housing, and development; crime, policing, and drug enforcement and addiction; and education and workforce development.
After Imel introduced the organizers of the event, Ford called the officials up one by one. They each had a few minutes to introduce themselves and address the issues raised during the virtual town hall. After all of the officials addressing quality of life issues finished, Ford opened the floor to more questions or statements from the audience. The officials could choose to respond, but were under no pressure to do so, because the popcorn-style questioning was not directed at them. Then Ford repeated this process with the other two groups of officials. Over coffee and sandwiches the following day, Ford said it was a deliberate choice to not have the officials take questions as a panel, to avoid confrontation, although Karaganis, Sung, and Narayanan wondered whether it stymied conversation as well.
People after the town hall were generally excited and happy with the event. At a restaurant later that evening, a young couple came over to thank the Polis and American Assembly team for helping to bring the event to the town. “We really appreciate what y’all are doing,” the woman said.
At his office the next day, Gary Fields, the superintendent of the Bowling Green school system, said he was impressed that the organizers of the town hall were able to get so many people from city government in the same room at the same time. He had been among the group addressing education issues the night before, and spent a good portion of his time reassuring the crowd that they were not planning on making cuts to arts programming.
On the other hand, Fields said he was disappointed by the lack of diversity in the room, which was overwhelmingly white. He said that Hispanic students had recently begun to outnumber black students, but that he hadn’t seen much representation from either group at the town hall. He pointed out that the people of color in the room were mostly from the university community, not the “other side of the tracks”—a division Fields said exists quite literally in Bowling Green, gesturing in the appropriate direction.
“They don’t feel like they have a voice in the community, so why would they feel like they have a voice last night?” Fields asked. He added that it’s his job to consider the needs of everyone in the city, not just the most vocal groups.
Down the street at the County Courthouse, Amy Milliken, the Warren County Attorney, said she wished she had more of an opportunity to respond to specific concerns at the town hall. Milliken was the only woman to speak in an official role the night before. She used her floor time to educate people about who she is and what she does, but the next day Milliken said that she wished she had more of an opportunity to respond to specific issues—to explain how to go about petitioning to allow the sale of liquor in Warren County, for example. (Under current laws, selling alcohol is legal in Bowling Green but not in the rest of the county, earning Warren County the designation of “moist.”) She wished there had been an opportunity to go over issues in more depth and tell people how to accomplish their goals.
The Community Meeting
As with the previous two parts of the Civic Assembly process, nobody knew what to expect at the community workshop on Friday afternoon, including the organizers, but nearly 50 residents showed up. It was the middle of the workday, so the attendees skewed older, but the young couple from the night before was there.
Imel, Ford, and others had rearranged the chairs into a circle. The man to my left introduced himself as Gary Mezsaros, a retired professor and current owner of a local chocolate store, and said he was really looking forward to digging into the issues. “I’m more excited about the community meeting than about last night,” he said.
That had been the promise of the event—an opportunity to go deep where the night before had been broad and shallow—but the event got off to a sluggish start when Ford spent nearly 45 minutes explaining the origins and purpose of the Civic Assembly, mentioning several times how very new it was, and how the organizers were learning as they went, just like the participants.
It showed. Journalists are not organizers, nor are academics, and that was readily apparent during the three-hour meeting.
Ford finally opened up the floor to questions about the Civic Assembly process. I saw people shift in their seats, looking somewhat confused; they thought they had come to talk about how to begin addressing some of the many issues that had been raised the night before.
The conversation meandered. One of the larger themes to come up was the future of the community, and the need for a shared vision. Several attendees were concerned about the city’s growth.
One man bluntly raised the issue of funding, and asked how much the Civic Assembly had cost. He got an answer, in part: Imel laid out the cost of renting the events space and printing the reports out for everyone, but The American Assembly team did not explain how much they had spent, and on what.
Some of the officials in attendance were not afraid of pushing back on residents’ complaints. Ben Peterson, the executive director of the Planning Commission of Warren County, said people don’t show up at other kinds of public meetings unless something is going on that they don’t like. Illustrating that point, Joe Plunk, an engineer with the Kentucky Transportation Cabinet, pointed out that a specific improvement that had majority support in the Polis survey was something the town had tried to do some years before but had been foiled by a vocal minority.
The meeting wasn’t all stern rebukes. At one point a woman praised the addition of a roundabout and said she would just about kiss the person responsible for that—and the man sitting next to her raised his hand and said that was him. Everyone burst into laughter.
“In many ways I was especially pleased with the community meeting because we had no agenda, and we wanted the conversation to go in a way that citizens and officials could take it, and they did,” Ford said afterwards. “If you told me at the beginning of it that it would have ended with a concerned citizen almost locking lips with the bureaucrat sitting next to her, well…that aspect of it was nice.”
While it may have been nice, fun, funny—I can’t say it seemed very productive. On the other hand, if you’re playing the long game, maybe a little goodwill goes a long way.
Ford said that they initially envisioned having break-out discussions during the community meeting. “We had labeled it as an unconference, but people outside the organizers didn’t really come into it with what they should use that time for,” he said.
I asked why attendees should be expected to come in with an agenda, without any instruction from the organizers, and Ford conceded that it “wasn’t a realistic use of that time.”
The aimlessness exhibited towards the end of the process seemed to stem from the local organizers’ inexperience in leading events of this kind, but also from the determination of The American Assembly and Polis to let the local partners be in the driver’s seat. Those teams probably had more facilitation experience—certainly The American Assembly, which has led a half-century’s worth of facilitated discussions—but they wanted to let the Daily News and Ford take point, to not overrun local expertise with outsider’s incompetence.
Peterson summed up the sentiment in the room well when he said, towards the end of the community meeting, “This has been a great brainstorming process, but if we don’t do more, that’s all it is.”
The American Assembly and Polis went to Bowling Green with few expectations. The entire process was brand new and a bit ad-hoc. Similarly, the Bowling Green Daily News wanted to engage and serve readers and the broader community in a new way, but didn’t know what they were looking for beyond that.
The biggest success according to all parties was participation. They worried nobody would contribute to the virtual town hall or show up in person, and that was hardly the case. People wanted to show up; people wanted to engage, including a large number of local officials and bureaucrats—the kind of people who make Bowling Green run.
However, perhaps the failure of imagination and of leadership was in not having a really good idea of what they wanted the people to show up for. In a town with few opportunities to be civically engaged, residents can’t be expected to know how to make the most of big public meetings like that. That responsibility has to be on the instigators.
Polis and American Assembly have acknowledged that they could have done more to analyze and synthesize the results into an even more detailed report, instead of having attendees try to make sense of nearly 900 statements presented without much context. That would go a long way toward helping structure and guide follow-up conversations as well. Although the town hall was loosely guided by the Polis results, those results were almost entirely irrelevant to the community meeting, which became more about what the handful of attendees wanted to discuss in detail than the issues raised during the virtual town hall. Unlike the vTaiwan process, where the points of consensus unearthed by Polis are used to channel the decision-making process forward into policy drafting, here there was little concrete connection between the results of the online survey and the offline conversation.
Increasing participation, particularly of the minority groups within Bowling Green, is another priority of Imel and Ford for future events.
Imel’s immediate concern, he said, is turning dialogue into content. He is working on figuring out how to translate the issues that arose during the Civic Assembly into stories for the Daily News, which will then further the conversation. He wants to do more solutions-oriented journalism at the paper, and thinks the Civic Assembly can help push that forward. He wanted the Civic Assembly to open a two-way conversation with readers, and it did. Imel just needs to figure how to talk back.
Progress on that front may take time, because Imel is currently occupied with election season. The Daily News is moderating two political forums starting in May. Only after those wrap up can he think about hosting additional topic-driven forums, à la Civic Assembly.
The possibility of creating citizen-led task forces to tackle specific issues came up during the community meeting, but as Ford pointed out, the Daily News might not be the best organization to lead such an effort. Ford also said he’d like to see a Civic Assembly process around the idea of vision-planning for Bowling Green’s future.
The American Assembly is working on a report on the Bowling Green Civic Assembly. Their interest is in finding a model that can be replicated elsewhere, so some of the questions raised here about moderation and funding will be on their list of things to figure out and refine in future iterations. The ultimate goal is a sort of how-to guide that other cities and organizations could pick up and adapt to their own purposes.
A bigger question that needs to be answered is what the goal of Civic Assembly really is. What does it mean to find consensus, and what do you do with it? When people were asked what could make Bowling Green and Warren County better—a better place to live, work, and spend time—a lot of people just wanted to talk about traffic. Roads and traffic was far and away the most popular topic, with 59-related statements, and that’s not even counting statements about parking, mass transit, biking, walking, or trains. Roads and traffic outnumbered statements about immigration, poverty, homelessness, and the opioid crisis. The most agreed-upon statement, out of 895, was about the importance of studying the impact on traffic flow of new development.
The American Assembly was created to tackle the most pressing challenges of the 20th century; this new version should aspire to do the same for the 21st. Perhaps consensus is built one road at a time. But a looming challenge for this coalition is to figure out how to use Polis and the Civic Assembly model to tackle issues that aren’t so simple or so easily agreed-upon, or else the process may only be used to solve problems of least resistance.