Stanford’s Jim Fishkin on devising deliberative democracy
The director of the Center for Deliberative Democracy talks about antidotes to affective polarization and deliberative polling.
The following is a transcript of the interview conducted between GovSight Editor in Chief Gillian Brassil and Stanford Center for Deliberative Democracy Director Jim Fishkin, the author of “When the People Speak.” The interview has been edited lightly for clarity.
Gillian Brassil: Could you talk about the Center for Deliberative Democracy at Stanford — what’s the mission and what is deliberative democracy itself?
Jim Fishkin: Deliberative democracy is the ideal of democracy when people are thinking, when they’re weighing competing alternatives on the basis of the best information available. The root of the word deliberation is “weighing,” weighing different sides of an argument.
And the problem that we address with not only the ideal of deliberative democracy, but our particular method for realizing it — which I call deliberative polling — is that most people most of the time are not paying a lot of attention to politics or policy. Or if they are paying attention, they will talk to the people they already tend to agree with. They will get the news sources they already find congenial. And they also may not think their voice matters so much because in a country of millions of people, millions of voters, they’re each subject to what social scientists have called “rational ignorance:” If I have one vote in millions, why should I pay a lot of attention? My individual vote or opinion won’t make much difference. And so people feel disconnected.
As a society, we tend to practice a kind of “audience democracy:” where we are spectators, not participants. We view what we view with a kind of inattention because we don’t think our voice matters. So a number of years ago I asked a couple of hypothetical questions which changed my whole — certainly all my work. It was, “what would the people think, under the best practical conditions for thinking about the issue?”
So to get the people, we have to have a really good scientific sample of the people that’s representative to begin with. And then the best practical conditions, it turns out that, with deliberative polling, it’s fairly commonsensical and straightforward. The amazing thing is that it works — and that it works well. And we can demonstrate that as a matter of social science.
So we have an issue. With America in One Room, we had five big issues, because our survey partner — N.O.R.C. at the University of Chicago — had surveyed what the people really wanted to see discussed in the campaign and they had five big issues: the environment, health care, immigration, foreign policy and the economy — with taxes especially. So then what we did — in order to prepare a deliberation, you need to identify proposals, because the idea is what would the people think about what should be done. So you need some proposals, we specified arguments for and against; we had an elaborate committee and advisory process. You’ll find a briefing book on our website at the Center for Deliberative Democracy, which actually boils down, we think, a good representation of the key issues facing America and the proposals to do something about it. And this is a non-partisan exercise, so we had Democratic leaning proposals, Republican leaning proposals; progressive proposals, some quite conservative proposals — but in each case it was balanced and we had arguments for and against. And the advisory committee wasn’t supposed to tell us what was the best proposal: They were supposed to vet the proposals just to say “this was pretty good information,” as accurate as could be found or if it’s something contested you have competing accounts of one said issue. And this document was then turned into a questionnaire.
So then we went and recruited with our survey partner a national random sample — and the thing is, we actually got the sample to turn up. So we had [526] Americans, registered voters, representative of the entire country, who flew to this big resort in Texas. We picked this resort, a nice place, right next to the Dallas-Fort Worth Airport so we could have the logistics and people could get there conveniently and leave on a Sunday afternoon and get home. We made it a controlled experiment because we also had 800-and-some people in a control group who took the same questionnaire at the recruitment time and at the end, they didn’t do any deliberation and they didn’t come to the event.
So then the people deliberate in randomly assigned small groups and then they identify the key questions that they wanted panels of competing experts — and indeed some politicians — to answer about these proposals. And we altered the small groups at the [primary] sessions and we actually fit in all five issues. The people worked until late at night and they also bonded with each other despite their extraordinary differences. And we collected the opinions again with a questionnaire at the end. The groups did not have to reach consensus, they only had to discover what questions they wanted to ask. They kept their actual opinions to themselves. It was sort of like a secret ballot. Because unlike a jury, there’s social pressure to get the verdict. But if you have a secret ballot or a confidential questionnaire, then we think we are actually getting what the people think.
We had some very highly contested issues, so the big surprise was that both the Republicans and the Democrats changed on the most contested, most extreme positions — they changed dramatically. In fact, we had 49 policy proposals covering the five areas and we found that in 27 of those proposals, we had people — significant numbers of people — taking the most extreme positions. We had a zero-to-10 scale and at least 15% of the sample took either the “zero position” in terms of “oppose as strongly as possible” or the “10 position:” support as strongly as possible. And in many cases, you had Republicans on one side and Democrats on the other; it was usually a division by party. But we can see that in the data.
So of those 27 highly polarized issues, it turns out on 26 of them, the Republicans and Democrats who took the most extreme positions moved significantly closer together. So I’ll give you an example of the most extreme positions and the movements overall which just astonished us — and by the way astonished the New York Times, which not only published them and featured them, but then went on to publish the photographs of every one of our [526] participants — and so to say this is what America really looks like. And we are very lucky to have this extremely credible survey partner N.O.R.C. of the University of Chicago, which is famous for being systematic and meticulous. So it was one of the best samples that you could find.
But to give you some examples, before the deliberations — and I’ll divide this by Republicans and Democrats — before deliberation, 79% of the Republicans supported forcing undocumented immigrants to return to their home countries before applying to work in the United States: 79%. After deliberation, that dropped 39 points to 40%. You never see a 39 point drop in actual public opinion, almost never.
Similarly, support for reducing the number of refugees allowed to resettle in the U.S. went from 66% to 34%. And support for actually increasing the number of visas for low-skilled workers went up 31 to 66%. Support for D.A.C.A. — this is among Republicans only — went up from 50 to 72%. And you know how contested D.A.C.A. has been on the Republican side.
But on the Democratic side, you also had tremendous changes. For example — you may remember this proposal, which is a very thoughtful and clever proposal, in terms of what it would do — that Cory Booker had of a government-funded “baby bond,” which is a bond that every child would get when they were born and it would grow in value and then support once somebody once they turned 18, supported college education, support a down payment on a house or other worthy purposes. Before deliberation, 62% supported that; 21% afterward supported that. Also a fall of 40 points. Also you had other proposals on the Democratic side — basically on the Democratic side the things that were very expensive and looked like they would just be kind of a giveaway — went down dramatically as things that the country couldn’t really afford.
On the Republican side, the things that were very conservative proposals directed at immigration went down, but also in other very conservative proposals you had big changes like support for in international relations/foreign policy area, support for the Iran Nuclear Accord or the Trans-Pacific Partnership — basically a lot of these issues where either the Republicans or the Democrats, either the conservatives or the progressives, had mostly been talking to themselves in their own filter bubbles. When they actually opened up to each other, a random sample randomly assigned is kind of a random sample. The small groups were all diverse; they were diverse demographically, they were diverse politically. And people got along, they listened, they opened up and they listened to the arguments on the other side.
And we didn’t explicitly connect the arguments to presidential candidates or political parties. People knew, people knew generally where the Republicans were advocating. They knew generally what the Democrats were advocating. But we had them focus on the merits of the issue. So the reason why I call my book “Democracy When the People are Thinking” is so that we can show that the public can really think about these issues instead of just exercising its partisan reflexes. They can open up to each other. Indeed some of the media coverage — there’s a wonderful CNN report on immigration discussions where the people were castigating each other and yelling at each other at the beginning. It was following one group saying, “send them all home,” and “you people don’t belong in the country etc.” And at the end, they’re hugging each other. They are opening up to each other, they are listening to each other. They have realized: these are people too. They have children. They have families. They have real concerns.
I think deliberation requires weighing the intellectual arguments. I think it’s facilitated by empathy. We think this project overall and that deliberation is the antidote to extreme polarization. Extreme polarization is destroying our democracy, creating deadlock. It’s substantive in terms of policies, and it’s “affective.” That is, the members of the two parties hate each other or dislike each other much more.
We measured that with a “feeling thermometer” about each party. We had a big effect on the affective polarization. They liked each other a lot more at the end. They respected each other, in terms of being willing to grant that someone you strongly disagree with has reasonable arguments. Maybe you disagree, but there are arguments on the other side.
We think deliberation is the secret to making democracy work. It’s the secret to formulating, I think, the missing ingredient to modern democracy. You see, modern democracies are based on party competition. Parties will do most anything to win. But that includes misleading people — interest groups and independent expenditure groups also — they will do whatever it takes to win. That includes bamboozling the people, misleading the people, propaganda. Now, that’s not a good way to assess the will of the people.
But ultimately democracy is about connecting the will of the people to what’s actually done. This tool of deliberative polling is not just a change in polling, it’s a tool for facilitating the expression and measurement of the will of the people. It can be used to make actual decisions and has been. Back in Texas, we used it in eight projects that led to Texas being the number-one state in wind power. An important policy area; not so contested at the time, more so contested now. It went from last to first in the 50 states thanks to our projects, this is described in my book.
Mongolia actually liked it so much — we did this in 110 cases in more than 30 countries. In Africa, in Asia, in Latin America, all over Europe: Europe-wide for all the members of the European Union. We’ve done it all over the world basically. In many cases, it’s actually been used to advise on important decisions. Mongolia actually instituted a law, requiring it, before they could change their constitution. Now they have changed their constitution, having done [deliberative polling], the amendment has passed.
Iceland has recently employed it for constitutional amendments and we are waiting for parliament to approve the amendments. It was commissioned by the prime minister and a cross party group in the parliament. I was in Chile recently, before the great crisis, and we announced with the senate in Chile to reform both the pension system and the health-care system which are big issues down there. I’m sure this will be delayed now with the virus, but we still hope to do it. It’s an example of how the people can be convened, to think together productively and with mutual respect.
I have collegues in political science in the United States who make really good careers talking about how the public is incompetant, the public is stupid. In fact, the public is not stupid, the public is very smart. You just need to engage them in a position where they think their voice matters and give them access to good information on both sides of an argument — and they turn out to have very thoughtful views. Those views deserve to be listened to in any system that calls itself a democracy because that is what democracy is about: listening to the people. Listening to the people under conditions where the people are giving more than just their vague impression of some sound-bite or headline and they actually engage with the arguments.
Democracy needs some retooling, deliberative polling is our contribution to that.
We can also do this successfully online with video-based group discussions. We recently in Britain did a national one online where we had several hundred participants and a control group of several hundred. The participants spent a whole weekend — Saturday and Sunday — I think everybody who participated in the Saturday came back for the Sunday. This is a scientific sample recruited by NatSen, which is the best academic survey organization in the United Kingdom. It was a rigorous study, representative of the entire country, engaged where the small groups could see each other the whole time. Lets say, a dozen people in each small group, adding up to several hundred. [There were] moderated discussions and they operated exactly as if they were in a face-to-face deliberative poll — except it was more cost effective. We didn’t have to pay for the hotel and the food and the transportation. We had coffee breaks, but the people found their own coffee. They could see each other — and because they could see each other, they could interact with each other just as they would have in a face-to-face environment.
We’re hoping and planning during this crisis to do some online deliberative polls. Our Japanese collaborators have just completed one online, I just got news of it today. We’re planning one in South Korea as well. We’ve had some very good success in South Korea with face-to-face ones, the face-to-face one was used by the government to officially decide whether or not to continue the construction of two nuclear reactors, because the government leaned anti-nuclear in its election manifesto, but they had these two nuclear reactors that were half built. Were they just gonna leave them half built? They’d have to import fossil fuels to make up the difference, there are implications for climate change. There’s a real deliberation — and it significantly moved to say, complete the reactors, but don’t build any more. And that’s what the government has accepted. It can be used to make actual decisions.
Our agenda is to convene, facilitate the expression of and make impactful the thoughtful and informed voice of the people.
GB: Could you expand more on the actual process for deliberative polling for participants and how it might change? Especially given that obviously we are in an internet age where we have access to broadband and are able to connect — like right now as we go through social distancing — online. So could you describe how that process might look like as well as how participants can be active?
JF: If you’re part of a scientific sample in a deliberative poll, whether it’s online or face to face, we try to keep the two versions as close as we can. We used to do the online process over a scheduled period of weeks. One hour a week or two hours a week. But we found that it didn’t have the compactness. My collaborator and deputy director Alice Siu said, “well, why don’t we just do it on a weekend?” Some of our collaborators said, “people won’t look at a computer screen for a whole day.”
In fact, they did. And they loved it. It’s the same thing. Let’s say you have 400 people and you randomly assign to small groups of 10 or 12 and you use software. We developed our own software with computer scientists at Stanford. It works great with Zoom. Zoom is what we used in the U.K.
Whether it is online or face to face, you randomly assign — let’s say — 300 people to small groups. You have trained moderators who will manage the discussion without giving any hint of their own position; we train the moderators carefully. Then they discuss the proposals on the topic and they try to discuss the competing sides and bring everybody into the discussion. Then they come up with key questions they want to ask with panel sessions of experts who bring different points of view. They don’t give speeches, they are only allowed to answer questions in the panel. Then you alternate that process for a weekend: small group plenaries, small group plenaries, until you cover the whole agenda. And then they take the same questionnaire at the end as the one that they took when they were first recruited, usually several weeks before. As with America in One Room, we have a control group. The opinions are changing because of the deliberations, not because of something happening in the wider world. So it’s very much the same.
One of the additions we have — because of one of our computer scientists here at Stanford — we have an automated version of the moderator. It controls the queue, allows everybody to get the discussion process and asks the group whether the competing sides of the argument, of the agenda, have been covered. Are they ready to move to the next proposal? Then it actually facilitates writing out and revising on the screen — that everybody can see — until they get to an agreed version of the key question or they vote on the top two or three questions. We actually used the automated moderator in the Japanese deliberative poll. We are planning to pilot it also in Finland with our collaborators there. In the U.S., we were going to pilot a big meeting in Colorado two weeks ago, but that was canceled because of the virus.
The actual reason for the automated moderator was not for deliberative polls — now we see it works quite well with that — but it was meant to spread the dialogue. I envision having social media partners or partner organizations and having large numbers of people go through the same experiences as in the deliberative poll.
We have an algorithm that will assign diverse members to a given small group. Republicans, Democrats, Independents; old, young, different parts of the country, different race or ethnicity etc., so that each of the groups is a small microcosm. The idea of that is not to have a representative example to collect the data, but rather to spread the experience of “moderated deliberation with diverse others.” Moderated deliberation with diverse others is an excellent way for everybody to come to their considered judgements and have the experience of deliberating.
Imagine what our politics would be like if massive numbers of people deliberated with each other. I think all the rancour that we see — the spread of disinformation, the spitefulness — it would be calmly resolved for most people. Some people might be difficult and there will always be a few problems, but basically deliberation as we showed with America in One Room is an antidote to extreme polarization. If we could spread it broadly, we would do that with the automated moderator. And we will do that.
In the meantime, deliberative polling, with a carefully recruited scientific random sample of the public, provides a representation of what the public would think under good conditions. What really weighs with them when they think about it? In other words: what are their conclusions and why, and what are the factors that weigh in when it comes to those conclusions? That could be an input into the policy debate — and it could be an input into policy making.
GB: Definitely. So this is obviously one great antidote to fighting polarization, but do you think that polarization has gotten worse over recent years? And what some of the factors might be — and how we can get past that?
JF: I do think so. I think polarization has gotten worse. I’ve got a colleague, Morris Fiorina, who thinks it’s gotten worse among the policy elites and not so much among the public.
I think the weight of the evidence from an awful lot of research is that it’s gotten worse in policy terms and it’s gotten worse in this rise of “affective polarization” where the two parties actually dislike each other. I think the internet has something to do with it. We are in our own filter bubbles in terms of the media we consult and in terms of the people we talk to. Party competition always had the problem of partisan-based propaganda in communication. The theorist of competitive democracy, Joseph Schumpeter, said, “it was a delusion to think about the will of the people. You could only get synthetic preferences.” Just as corporations will advertise and create demand for their product, political actors will advertise and use the same methods to create demand for their parties and candidates. That’s part of the competition. And that will include indoctrination.
It’s not only indoctrination, it’s sometimes outright falsehood — and people have crossed the line. Parties want to win, interest groups want to win and the media has been greatly weakened. Ironically by the rise of technologies that allow people to get exactly the messages that they select. It’s an increase in the opportunities to exercise your freedom that have led to the very opposite of what they were supposed to. John Stewart Milton thought that liberty would lead to an exposure of diversity, so that people could really think for themselves. But now liberty leads to exposure to sameness — and people are more subject to manipulation because they may never hear the other side of the argument. The public sphere has been decomposed into many sorts of droplets. Falsehoods are not answered. Misleading arguments are not answered. The cable news networks, talk radio and a lot of internet-based sources that are not really practicing the art of journalism, they’re practicing the art of propaganda. I’m not talking about the Russians. We can bring in the Russians and other sources of misinformation and make it even worse. The sheer force of competition has led to this impasse.
We need an effort to bring deliberation into the public dialogue. That’s what we are trying to contribute. Yes, we are more polarized substantively — and in terms of affective polarization as well. There is definitely an increase in the affect. We are in a sorry state.
Democracy is under threat in another way: It seems to have produced such deadlock and division that some people have turned to authoritarian solutions or quasi-authoritarian solutions, where you have the trappings of democracy but you don’t really have real elections. Or you have some kind of benevolent authoritarianism where technocrats or experts make the decisions for the people and try to provide good government. But who’s to keep those technocrats accountable? Singapore, in what is sometimes called the “China model,” lays out the aspiration.
I think that democracy should be cherished. We should not have all our decisions made for us. What kind of citizens would we be? We should collectively figure out ways that we can take responsibility for the decisions that affect our lives. Democracy can be recalibrated to do that.
GB: Definitely. This is my last question for you. In times like this, we are all kind of trapped at home and are trying to flatten the curve as people are trying to stay in their houses or have shelter-in-place policies. How can people become more involved, more informed and open themselves up to learning about the other side in this way of becoming better citizens?
JF: I think that you need to open up sincere dialogue with people you disagree with. You need to consult news sources that hear what the other side is saying, if you’re on one side. You need to join with others on issues that actually engage your deepest commitments.
After this crisis resolves itself, there will be an opportunity for a tremendous awakening of civic capacities as we deal with the situation that we will find ourselves in. There will be big tradeoffs that have to be considered. What sacrifices will the public make? What will they be willing to live with? There will be opportunities for all kinds of deliberative democracy efforts once we are allowed to go outside — or even when we are inside, if we can mount some online things. There are lots of organizations that exercise some civic capacities for dialogue; I would recommend pursuing those.
If anybody has ideas for what our center can collaborate on, let us know. You can find us on the web: Center for Deliberative Democracy at Stanford University. But we are just a catalyst. We are very tiny, but we play well with others: We have partners all over the world.
Now is not a time to be despondent: It’s a time to dream big and prepare for the awakening.
Questions? Ask us at contact@govsight.co.
Like what you read but prefer to learn with your ears? Listen to The Insight Podcast by GovSight on Apple Podcasts, Spotify or Podbean every Monday.
Follow GovSight on Twitter @GovSight1, Instagram @govsight and Facebook @GovSight. Go to govsight.com to see how GovSight is making “Citizenship. Simplified.”
