A rigorous experiment in 'deliberative polling' finds civil discourse over 'civil war'
Just as Donald Trump’s fears of impeachment were leading him to tweet threats about "civil war," a real-life microcosm of the American electorate that gathered just outside Dallas, dubbed “America in One Room,” has produced a diametrically opposite picture of highly functional civic dialogue. “By the end of their time there, they referred to each other as friends,” one facilitator, Amy Carl, said of her group of 14 people.
Think “New England town meeting” rather than “civil war.”
“Even with our differences in opinion, we were able to constructively converse, to be polite to one another," said Raymond, an older white man from Colorado. “I really believed our country was divided — I don’t believe that anymore.”
But it wasn’t just a feel-good exercise, although 98.2% found it “valuable.” Suggestions that entertainment would be needed to break things up were politely ignored. Instead, there were reams of briefing materials to digest, informative large-group plenary sessions with experts on exciting subjects like tax policy, intensive deliberations in small groups of 14 or 15 people, like the one Carl facilitated, and rigorous polling before and after, with a control group for comparison.
All the polling, as well as recruitment, was done by NORC at the University of Chicago, one of the oldest public opinion research institutions in America, drawing on its 35,000-household AmeriSpeak Panel, through a multi-step selection process. It’s arguably the best-curated polling panel in America. NORC did preliminary polling to identify the issues covered as well.
In short, it was a rigorous experiment in democracy, testing the soundness of the process called “deliberative polling,” which Stanford political scientist James Fishkin first came up with more than 30 years ago, and which has been used more than 100 times around the world. (See previous Salon story here.) This effort was perhaps the most ambitious ever, made possible with sponsorship from the Helena Foundation. which focuses on identifying solutions for significant societal problems.
“We have a control group that didn't change appreciably at all on the issues," Fishkin told Salon, while the America in One Room participants "changed dramatically." He added, “This is a pilot for a more substantive form of public opinion, and a more substantive form of democracy.”
“There are thousands and thousands of polls that show what the public is thinking when it's not thinking,” Fishkin said. "Because most of the time the public's not thinking very much, and my colleagues wrongly conclude that the public is not capable of dealing with complicated important issues.” But that’s mistaken, he believes.
“They are certainly capable, they just normally don't do it, because they think their voice doesn't matter, and they know everybody's trying to manipulate them. They're ‘rationally ignorant,’ they've got other things to do with their time,” Fishkin explained. ”So if there's so many polls showing what the public is not thinking, why not have some that show what they would think and insert that into the public dialogue and the policy process?”
“We got very considerable movements in each of the five issue areas,” Fishkin noted, even though some skeptics thought that five issues would be too much, that “the public would just be overwhelmed — you'll get absolutely nothing.” Instead, it turned out just as Fishkin’s theory and past experience predicted, even perhaps exceeding his expectations.
On the question of undocumented immigrants being "forced to return to their home countries before applying to legally come to the U.S.," for example, Republican support dropped from 78.7% before deliberations to 40.3% afterwards — a shift so dramatic it surprised even Fishkin.
“My expectation about that was that the support and the opposition was entrenched by this point,” he said. “So to get a big change in Republican support on that issue is just stunning.”
There were also drops on the Democratic side — support for "baby bonds," a key plank in Sen. Cory Booker's presidential campaign, dropped dramatically from 61.6% to 21.3%. Aside from that, none of them matched the largest GOP shifts on immigration-related proposals. On a set of five immigration policies, Republican views shifted an average of just over 30 points, while the average Democratic shift was just half that on a set of five economic issues, which included Medicare for All and free public college for those who can't afford it.
“I think what happened is they listened to the other side,” Fishkin said of the GOP immigration shifts. “Of course there were also children of immigrants and other immigrants in some of the rooms,” he added, “But most importantly, they listened to the other side of the argument, and they thought about the consequences.” They also learned some facts. By far the largest knowledge gain was the increased percentage who could correctly identify the number of undocumented immigrants in the U.S. (roughly 10 million). That percentage rose from 20.5% to 74%, accounting for roughly half the total knowledge gain across seven items.
“The most polarizing proposals, whether from the left or the right, generally lost support, and a number of more centrist proposals moved to the foreground,” as the executive summary of the event noted. But the shifts were somewhat asymmetrical, as noted above, and the more centrist proposals tended center-left. Support for Medicare for All dropped by 7.6% overall (13.5% among Democrats) to 39.9%, while support for repealing the Affordable Care Act dropped by 9.2% (20.9% among Republicans) to 25.6%. Support for a Medicare buy-in rose almost 4 points (similarly among Democrats) from 67.3% to 71.2%.
Similarly, support for taxing income above $2 million at a higher tax rate rose slightly from 70.4% to 71.8%. On the statement, “Cooperation between the countries of the world is necessary to fight climate change,” agreement rose from 80.4% to 92.9% overall (which is pretty much “mom and apple pie” territory), and from 65.2 to 89.4% among Republicans.
Those numbers highlight something else: There was bipartisan support for some “far left” positions as Fox News would have it, even before deliberations began. And not just on those that were on the official menu.
In the very first group discussion, which was just supposed to be a meet and greet occasion, gun control came up. Carl joked, "I had a moment of, ‘Oh my God! I don't want to do this! I do not want to have this conversation with these people I do not know, that I am trying to establish rapport with.’" But to her "delight and surprise," she said, "the group went around the table and immediately came to a consensus that guns should be should be regulated like cars." In this case, the preconceived division lines didn't need to be bridged, since they didn't exist in the first place.
As noted above, Fishkin has been promoting deliberative polling for over 30 years, and has a long string of success stories he can point to. But this approach has yet to receive the kind of broad acceptance that its track record so clearly deserves. The reasons may reflect a deeper sociopolitical reality: Fishkin’s idea was first floated in the late 1980s, roughly coinciding with two other proposals I highlighted in my earlier story that share in common a concern for giving common citizens a greater voice. The timing of all three roughly corresponded with an early wave of mass discontent, reflect in the popular independent candidacy of Ross Perot in 1992. But that discontent was largely not shared by elites — a situation that has now changed.
Elites as much as anyone else now seem to realize there are profound problems with our democracy. This awareness lag can be explained in terms of Peter Turchin’s "structural demographic theory" (Salon story here), which explains why elite distress tends to lag behind mass immiseration by about a generation or so across civilizations and centuries. Solving the problems we face is far from a given, but an end to denial is a start.
Still, elite intransigence, and the tendency to blame-shift onto the mass public remain serious impediments. One such strain of thought is reflected in the idea that “human brains aren’t built for self-rule,” as Politico recently summarized the argument of a controversial paper that nonetheless echoes a widespread preoccupation with the failures of democracy.
But the problem isn’t individual citizens, Fishkin argues, it’s the lack of proper settings in which democracy can flourish. America in One Room vividly shows what he means, and the results it produced should influence the 2020 campaign — that’s one reason it was held when it was, months ahead of the Iowa caucuses. But this was less an answer in and of itself than a high-profile proof of concept, meant to inspire emulation and repetition, both throughout the election cycle and beyond.
“It's very obvious there are problems with liberal democracy in the world,” said Henry Elkus, Helena’s founder and CEO. “There is not in our discourse, politically, a space where the will of the American people can actually be ascertained or understood.”
“From my standpoint as a democratic theorist, the whole point of democracy is to connect the will of the people with what is actually done,” Fishkin said. But that's easier said than done.
“On a debate stage there's only 30 seconds or 40 seconds per candidate for them to get their point across," Elkus noted. "They aren't able to have a nuanced discussion about the issues, because there's not enough time." With a nod to Fishkin’s “rational ignorance,” he made a similar point about citizens. “If you’re a coal miner in West Virginia, you simply don't have the time to sit with a strategic briefing document and understand with nuance what your beliefs are on NAFTA. You have to put food on the table,” Elkus said. “There are simply not platforms for folks to come together and understand what they actually believe about the critical issues,. This event is a direct antidote to that.”
Helena is a new model foundation, not focused on any particular subject area, Elkus explained. "Our purpose is to do individual, one-by-one projects. The one goal of the organization for these projects is to identify societal problems we feel should be solved and have feasible solutions, find those solutions, turn them into these projects, and then do whatever is necessary to make the project successful."
For Helena's first project, “We identified and helped launch the first-ever technology that can suck carbon dioxide out of the air, and then sell it as a product in order to fight climate change,” Elkus explained. The second project was to protect the national electric grid from multiple forms of threat. Helena sees America in One Room as a way to significantly improve democracy both immediately, in this election cycle, and long-term at every level of government.
Not only can the deliberative polling results influence political campaigns and the media, the questions asked by participants can help as well — along the lines of the “public agenda” model for campaign coverage, long championed by NYU’s Jay Rosen as a substance-based alternative to the media’s horse-race model. The full model calls for media organizations to investigate what questions their audience wants candidates to answer, and to ground their coverage in that. Doing it properly calls for an intense front-end investment, but America in One Room provides a virtually no-cost approximation.
NORC’s preliminary poll identified the issue areas of greatest public concern. Then, in the process itself, the small groups came up with questions they posed to experts in the plenary sessions.
Those questions were "astoundingly good,” said Helena’s COO and executive director, Sam Feinburg. They were “very nuanced, very detailed, very on-the-nose, not ‘gotcha’ questions. Almost every single one of our speakers remarked on how high-quality the questions were.”
Helena is already talking to CNN and the New York Times about using those questions for the next round of Democratic primary debates, since “a representative sample of the country came together to decide that these are the most important things we the American people want to hear the answers to.”
Beyond that, this model is producing data on something "far more important than the short-term, ‘Who's going to be president?’-type data,” Elkus said. “It's producing data on ‘What are the fundamental beliefs of our country?’ And that data doesn't really spoil like milk. Whether a representative sample of the entire country believes America should be interventionist as a foreign policy power — whether we should be the global police force or not — is permanently important data. It's going to be permanently important data next November, but also the November after that.”
Elkus and Feinburg are both in their early 20s. “We are about to go through a lifetime of critical changes in the world, and our country is going to be at the helm of those changes one way or the other, whether negative or positive," Elkus said. "The data that this event produces will be very important to how America is going to tackle or not tackle climate change. How it’s going to handle climate refugees, how it deals with China, and ongoing shifts in global power in the decades to come. These are the critical issues of my generation. We're doing this event to understand what the American public believes about those critical points.”
At the same time, Helena isn’t looking to this one event for all the answers. “There should be a deliberative democracy poll of the entire country every quarter, four times a year," Feinburg said. "The American people should be advised on what the American people believe on the issues all the time. We should be as a country acting upon the best information we can, and every other country should have this kind of process as well.”
In the meantime, Fishkin has also been involved in developing a streamlined version. “A cheaper way of doing it is to do the deliberate poll online and we have done some deliverables online using video-based software,” Fishkin said. “You have 300 people in the plenary session and you have 20 small groups of 15 people, and you see everybody in your small groups, and we go through one, two or three days in the same way as we do face-to-face.”
While balanced, representative samples are essential for polling purposes, the deliberative process is actually somewhat distinct. There was no way for a 15-person small group to be representative of the whole country. So the essentials required for deliberation to work were less rigorous.
“The experimental treatment if you will, is a moderated discussion with diverse others in a context where you have good information and can get your questions answered,” Fishkin said. “That's a mouthful, but that doesn't require a random sample, it just requires diversity… . So with that insight, we've been talking about spreading the deliberation to much larger numbers in order to allow the deliberation to change opinion right now.”
Working with engineering colleagues at Stanford, they’ve developed special software with an automated moderator function. “So we could do thousands or tens of thousands or hundreds of thousands of these groups,” Fishkin said. “We hope to scale this sometime during this presidential election on the issues, maybe using some of the same briefing materials, and use social media to recruit diverse others,” using an algorithm to create politically and demographically diverse small groups.
In the meantime, Amy Carl reflected on perhaps the most basic takeaway from the experience. “We are not as woefully divided and beyond hope as we are made out to be," she said. "I really feel that in my bones. I think if people are forced to step out of their bubbles, their comfort zones, people would be amazed at how fundamentally not different we are.”