Home » Blog » ideas » Mistrust, Efficacy and the New Civics – a whitepaper for the Knight Foundation

Mistrust, Efficacy and the New Civics – a whitepaper for the Knight Foundation

When things get serious in the media space, my friends at the Knight Foundation rally the troops. Last week, I was invited to a workshop Knight held with the Aspen Institute on trust, media and democracy in America. I prepared a whitepaper for the workshop, which I’m publishing here at the suggestion of several of the workshop participants, who found it useful.

The paper I wrote – “Mistrust, efficacy and the new civics: understanding the deep roots of the crisis of faith in journalism” – served two purposes for me. First, it’s a rough outline of the book I’m working on this next year about mistrust and civics, which means I can pretend that I’ve been working on my book this summer. Second, it let me put certain stakes in the ground for my discussion with my friends at Knight. Conversations about mistrust in journalism have a tendency to focus on the uniqueness of the profession and its critical civic role in the US and in other open societies. I wanted to be clear that I think journalism has a great deal in common with other large institutions that are suffering declines in trust. Yes, the press has come under special scrutiny due to President Trump’s decision to demonize and threaten journalists, but I think mistrust in civic institutions is much broader than mistrust in the press.

Because mistrust is broad-based, press-centric solutions to mistrust are likely to fail. This is a broad civic problem, not a problem of fake news, of fact checking or of listening more to our readers. The shape of civics is changing, and while many citizens have lost confidence in existing institutions, others are finding new ways to participate. The path forward for news media is to help readers be effective civic actors. If news organizations can help make citizens feel powerful, like they can make effective civic change, they’ll develop a strength and loyalty they’ve not felt in years.

To my surprise and delight, the workshop participants needed little or no convincing that journalism’s problems were part of a larger anti-institutional moment in America. I brought in Chris Hayes’s idea that “left/right” is no longer as useful a distinction in American politics as “insurrectionist/institutionalist”, and we had a helpful debate about what it would mean to bring insurrectionists – those who believe our institutions are failing and need to be replaced by newer structures – into a room full of institutionalists – people who see our institutions are central to our open society, in need of strengthening and reinforcement, but worth defending and preserving.

What started to emerge was a two-dimensional grid of left/right and insurrectionist/institutionalist to understand a picture of American politics that can include both the Occupy Movement and Hillary Clinton as leftists, and Donald Trump and Paul Ryan on the right. While the US is currently wrestling with right-wing insurrectionism (and this meeting happened even before neo-Nazis marched in Charlottesville, VA and the President assured us that some of them were nice people) it’s important to remember that anti-institutionalism comes in a variety of flavors. (One researcher at the meeting observed that the common ground of insurrectionism may explain the otherwise weird phenomenon of Bernie supporters going on to support Trump.)

Mapping politicians on the twin axes of left/right and institutionalist/insurrectionist

Another idea that came up was that the categories of institutionalism and insurrectionism are fluid and changing, because revolutionaries quickly become institutions. When Google came into the search engine game, it was a revolutionary new player, upending Yahoo!, Lycos, Alta Visa and others. Twenty years later, it’s one of the most powerful corporate and civic actors in the world. Understanding that successful revolutions tend to beget institutions is helpful for understanding insurrectionism as a political pole. Some insurrectionists will win their battles and may find themselves defending the institutions they build. (I think fondly of my friend Joi Ito, who spent years as an enfant terrible in Japanese internet circles before becoming surprisingly acceptable as the director of the MIT Media Lab.) Others will hold onto insurrectionism even when their side wins – my friend Sami ben Gharbia moved seamlessly to being a critic of the Ben Ali government in Tunisia to critiquing the new government the Arab Spring brought to power.

Understanding that revolutions beget new institutions made me think of Pierre Rosanvallon’s work on Counterdemocracy, which isn’t as well known in the US as it should be. Rosanvallon observes that, during the French Revolution, a slew of new institutions sprang up to ensure the new leaders stayed true to their values. This wasn’t always a pretty picture – the Terror came in part from the institutions Rosanvallon explores – but the idea that democracy needs to be counterbalanced and bolstered by forces of oversight, prevention and judgement is one worth considering as we examine modern-day institutions and their shortcomings.

When a disruptive entity like Google or Facebook becomes an institution, it’s incumbent on us to build systems that can monitor their behavior and hold them accountable. It’s rare that existing regulatory structures are well-equipped to serve as counter-democratic institutions to counterbalance the new ways in which they work. As a result, there’s at least two ways look for change as an insurrectionist: you can identify institutions that aren’t working well and strive to replace them with something better, or you can dedicate yourself to monitoring and counterbalancing those institutions, building counterdemocratic institutions in the process.

What does this mean for the fine folks working with Knight on news and trust? The press is a key part of counterdemocracy. It needs to hold power responsible. As democratic institutions of power change, counterdemocratic systems have to change as well – nostalgia for how the press used to work is less helpful than understanding the way in which political and civic power are changing, so that the press can continue to act as an effective counterweight. The good news for me was that the folks at Knight are emphatically not holding onto a nostalgic view of the press, trying to return to a Watergate golden age. The bad news? Just like the rest of us, they don’t know what the shape of this emergent new civics is either.


Here’s a prettier version of this paper, hosted my MIT’s DSpace archive. What follows below is a less pretty, but web-friendlier version.

Mistrust, efficacy and the new civics:
understanding the deep roots of the crisis of faith in journalism

Ethan Zuckerman, Center for Civic Media, MIT Media Lab
August 2017

Executive summary
Current fears over mistrust in journalism have deep roots. Not only has trust in news media been declining since a high point just after Watergate, but American trust in institutions of all sorts is at historic lows. This phenomenon is present to differing degrees in many advanced nations, suggesting that mistrust in institutions is a phenomenon we need to consider as a new reality, not a momentary disruption of existing patterns. Furthermore, it suggests that mistrust in media is less a product of recent technological and political developments, but part of a decades-long pattern that many advanced democracies are experiencing.

Addressing mistrust in media requires that we examine why mistrust in institutions as a whole is rising. One possible explanation is that our existing institutions aren’t working well for many citizens. Citizens who feel they can’t influence the governments that represent them are less likely to participate in civics. Some evidence exists that the shape of civic participation in the US is changing shape, with young people more focused on influencing institutions through markets (boycotts, buycotts and socially responsible businesses), code (technologies that make new behaviors possible, like solar panels or electric cars) and norms (influencing public attitudes) than through law. By understanding and reporting on this new, emergent civics, journalists may be able to increase their relevance to contemporary audiences alienated from traditional civics.

One critical shift that social media has helped accelerate, though not cause, is the fragmentation of a single, coherent public sphere. While scholars have been aware of this problem for decades, we seem to have shifted to a more dramatic divide, in which people who read different media outlets may have entirely different agendas of what’s worth paying attention to. It is unlikely that a single, authoritative entity – whether it is mainstream media or the presidency – will emerge to fill this agenda-setting function. Instead, we face the personal challenge of understanding what issues are important for people from different backgrounds or ideologies.

Addressing the current state of mistrust in journalism will require addressing the broader crisis of trust in institutions. Given the timeline of this crisis, which is unfolding over decades, it is unlikely that digital technologies are the primary actor responsible for the surprises of the past year. While digital technologies may help us address issues, like a disappearing sense of common ground, the underlying issues of mistrust likely require close examination of the changing nature of civics and public attitudes to democracy.

Introduction
The presidency of Donald Trump is a confusing time for journalists and those who see journalism as an integral component of a democratic and open society.

Consider a recent development in the ongoing feud between the President and CNN. On July 2nd, Donald Trump posted a 28 second video clip to his personal Twitter account for the benefit of his 33.4 million followers. The video, a clip from professional wrestling event Wrestlemania 23 (“The Battle of the Billionaires”), shows Trump knocking wrestling executive Vince McMahon to the ground and punching him in the face. In the video, McMahon’s face is replaced with the CNN logo, and the clip ends with an altered logo reading “FNN: Fraud News Network”. It was, by far, Trump’s most popular tweet in the past month, receiving 587,000 favorites and 350,000 retweets, including a retweet from the official presidential account.

CNN responded to the presidential tweet, expressing disappointment that the president would encourage violence against journalists. Then CNN political reporter Andrew Kaczynski tracked down Reddit user “HanAssholeSolo”, who posted the video on the popular Reddit forum, The_Donald. Noting that the Reddit user had apologized for the wrestling video, as well as for a long history of racist and islamophobic posts, and agreed not to post this type of content again, Kaczynski declined to identify the person behind the account. Ominously, he left the door open: “CNN reserves the right to publish his identity should any of that change.” The possibility that the video creator might be identified enraged a group of online Trump supporters, who began a campaign of anti-CNN videos organized under the hashtag #CNNBlackmail, supported by Wikileaks founder Julian Assange, who took to Twitter to speculate on the crimes CNN might have committed in their reportage. By July 6th, Alex Jones’s Infowars.com was offering a $20,000 prize in “The Great CNN Meme War”, a competition to find the best meme in which the President attacked and defeated CNN.

It’s not hard to encounter a story like this one and wonder what precisely has happened to the relationship between the press, the government and the American people. What does it mean for democracy when a sitting president refers to the press as “the opposition party”? How did trust in media drop so low that attacks on a cable news network serve some of a politician’s most popular stances? How did “fake news” become the preferred epithet for reporting one political party or another disagrees with? Where are all these strange internet memes coming from, and do they represent a groundswell of political power? Or just teenagers playing a game of one-upsmanship? And is this really what we want major news outlets, including the Washington Post, the New York Times and CBS, to be covering?)

These are worthwhile questions, and public policy experts, journalists and academics are justified in spending significant time understanding these topics. But given the fascinating and disconcerting details of this wildly shifting media landscape, it is easy to miss the larger social changes that are redefining the civic role of journalism. I believe that three shifts underlie and help explain the confusing and challenging landscape we currently face and may offer direction for those who seek to strengthen the importance of reliable information to an engaged citizenry:

– The decline of trust in journalism is part of a larger collapse of trust in institutions of all kinds
– Low trust in institutions creates a crisis for civics, leaving citizens looking for new ways to be effective in influencing political and social processes
– The search for efficacy is leading citizens into polarized media spaces that have so little overlap that shared consensus on basic civic facts is difficult to achieve

I will unpack these three shifts in turn, arguing that each has a much deeper set of roots than the current political moment. These factors lead me to a set of question for anyone seeking to strengthen the importance of reliable information in our civic culture. Because these shifts are deeper than the introduction of a single new technology or the rise of a specific political figure, these questions focus less on mitigating the impact of recent technological shifts and more on either reversing these larger trends, or creating a healthier civic culture that responds to these changes.

What happened to trust?

Since 1958, the National Election Study and other pollsters have asked a sample of Americans the following question: “Do you trust the government in Washington to do the right thing all or most of the time?” Trust peaked during the Johnson administration in 1964, at 77%. It declined precipitously under Nixon, Ford and Carter, recovered somewhat under Reagan, and nose-dived under George HW Bush. Trust rose through Clinton’s presidency and peaked just after George W. Bush led the country into war in Iraq and Afghanistan, collapsing throughout his presidency to the sub-25% levels that characterized Obama’s years in office. Between Johnson and Obama, American attitudes towards Washington reversed themselves – in the mid 1960s, it was as difficult to find someone with low trust in the federal government as it is difficult today to find someone who deeply trusts the government.

Data from Gallup, derived largely from the National Election Survey

Declining trust in government, especially in Congress – the least trusted branch of our tripartite system – is an old story, and generations of politicians have run against Washington, taking advantage of the tendency for Americans to re-elect their representatives while condemning Congress as a whole. What’s more surprising is the slide in confidence in institutions of all sorts. Trust in public schools has dropped from 62% in 1975 to 31% now, while confidence in the medical system has fallen from 80% to 37% in the same time period. We see significant decreases in confidence in organized religion, banks, organized labor, the criminal justice system and in big business. The only institutions that have increased in trust in Gallup’s surveys are the military, which faced Vietnam-era skepticism when Gallup began its questioning, and small business, which is less a conventional institution than the invitation to imagine an individual businessperson. With the exception of the military, Americans show themselves to be increasingly skeptical of large or bureaucratic institutions, from courts to churches.

Data on the left from Gallup. On the right are my calculations of drops in trust, based on Gallup data.

American media institutions have experienced the same decades-long fall in trust. Newspapers were trusted by 51% of American survey respondents in 1979, compared to 20% in 2016. Trust in broadcast television peaked at 46% in 1993 and now sits at 21%. Trust in mass media as a whole peaked at 72% in 1976, in the wake of the press’s role in exposing the Watergate scandal. Four decades later, that figure is now 32%, less than half of its peak. And while Republicans now show a very sharp drop in trust in mainstream media – from 32% in 2015 to 14% in 2016, trust in mass media has dropped steadily for Democrats and independents as well.

In other words, the internet and social media has not destroyed trust in media – trust was dropping even before cable TV became popular. Nor is the internet becoming a more trusted medium than newspapers or television – in 2014, 19% of survey respondents said they put a great deal of trust in internet news. Instead, trust in media has fallen steadily since the 1980s and 1990s, now resting at roughly half the level it enjoyed 30 years ago, much like other indicators of American trust in institutions.

It’s not only Americans who are skeptical of institutions, and of media in particular. Edelman, a US-based PR firm, conducts an annual, global survey of trust called Eurobarometer, which compares levels of trust in institutions similar to those Gallup asks about. The 2017 Eurobarometer survey identifies the US as “neutral”, between a small number of high trust countries and a large set of mistrustful countries. (Only one of the five countries Eurobarometer lists as highly trusting are open societies, rated as “free” by Freedom House: India. The other four – China, Indonesia, Singapore and the United Arab Emirates, are partly free or not free. Depressingly, there is a discernable, if weak, correlation (R2=0.162) between more open societies and low scores on Edelman’s trust metric.) As in the US, trust in media plumbed new depths in Eurobarometer countries, reaching all time lows in 17 of the 28 countries surveyed and leaving media contending with government as the least trusted set of institutions (business and NGOs rate significantly higher, though trust in all institutions is dropping year on year.)

So what happened to trust?

By recognizing that the decrease in trust in media is part of a larger trend of reduced trust in institutions, and understanding that shift as a trend that’s unfolded over at least 4 decades, we can dismiss some overly simplistic explanations for the current moment. The decline of trust in journalism precedes Donald Trump. While it’s likely that trust in media will fall farther under a government that presents journalists as the opposition party, Trump’s choice of the press as enemy is shrewd recognition of a trend already underway. Similarly, we can reject the facile argument that the internet has destroyed trust in media and other institutions. Even if we date broad public influence of the internet to 2000, when only 52% of the US population was online, the decline in trust in journalism began at least 20 years earlier. If we accept the current moment as part of a larger trend, we need a more systemic explanation for the collapse of trust.

Scholars have studied interpersonal trust – the question of how much you can trust other individuals in society – for decades, finding the broader world shows fairly stable interpersonal trust. Yet a decrease of trust in institutions is widespread globally, as seen both in the Eurobarometer data and in Gallup OECD data. It’s not just that we trust each other less – people around the world appear to trust institutions less.

It’s also possible that reduced confidence in institutions could relate to economic stress. As numerous scholars, notably Thomas Piketty, have observed, economic inequality is reaching heights in the US not seen since the Gilded Age. The decrease of confidence in institutions roughly correlates with the increase Piketty sees in inequality, which is stable through the 50’s, 60’s and mid-70’s, rising sharply from there.

We might think of an explanation in which citizens, frustrated by their decreasing share of the pie, punish the societal institutions responsible for their plight. But with this explanation, we would expect to see rising inequality accompanied by a steady drop in consumer confidence. We don’t – consumer confidence in the US and in the OECD more broadly is roughly as high now as it was in the 1960s, despite sharp drops during moments of economic stress and a rise during the “long boom” of the ’90s and 2000s. It’s possible that citizens should be punishing governments, banks and businesses for rising inequality, but consumer behavior and confidence doesn’t corroborate the story.

I favor a third theory, put forward by Kenneth Newton and Pippa Norris, called the institutional performance model. Simply put, when institutions perform poorly, people lose trust in them: “It is primarily governmental performance that determines the level of citizens’ confidence in public institutions.” That trust in institutions, easily lost, takes a long time to regain. We might understand the collapse of confidence in US institutions as a set of high visibility crises: Vietnam and Watergate as eroding confidence in the federal government, the Catholic Church sex scandal destroying trust in that institution, the 2007 financial collapse damaging faith in banks and big business.

Newton and Norris developed their theories in the mid-1990s, noting that confidence in public institutions was plumbing new depths. In retrospect, their concerns seem well-founded, as the trends they observed have simply increased over time. In the mid 1990s, Newton and Norris were comfortable positing a relationship between society-wide interpersonal trust and trust in institutions – that relationship is less clear now, because interpersonal trust has remained fairly constant while trust in institutions has decreased. One explanation for the decrease in institutional trust is that institutions have performed poorly, and that citizens are increasingly aware of their shortcomings.

Cultural and technological shifts may have made it easier for institutions to lose trust and harder to regain it. Watergate returned the US press to its progressive-era muckraking roots and ended a period of deference in which indiscretions by figures of authority were sometimes ignored. (It’s interesting to imagine the Clinton-era press covering JFK’s personal life.) An explosion in news availability, through cable television’s 24-hour news cycle and the internet, has ensured a steady stream of negative news, which engages audiences through fear and outrage. The rise of social media fuels the fire, allowing individuals to report institutional failures (police shootings, for example) and spread their dismay to friends and broader audiences. Accompanying the evolution of media technologies is education: in 1971, 12% of Americans had graduated from college, and 57% from high school. By 2012, 31% had college degrees, and 88% had high school diplomas. The citizens of 2017 are better positioned to be critical of institutions than those of 1964.

If we accept any of these explanations for a decrease in trust in institutions, the obvious question emerges: How do we reverse this trend? How do we restore public trust?

It’s worth noting that those most concerned with restoring public trust tend to be elites, those for whom existing institutions are often working quite well. Eurobarometer’s 2017 report focuses on a widening trust gap between a well-informed 15% of the population and a less informed 85%. The well-informed minority scores 60 on Edelman’s trust index, while the less-informed majority is 15 points lower, at 45. The gap between elites and the majority is largest in the US – 22 points separate the groups.

One approach to institutional mistrust is to try and educate this disenchanted majority, helping them understand why our institutions are not as broken as we sometimes imagine. Any approach is unlikely to reach all citizens – some will remain frustrated and alienated, due to disinterest, misinformation, a healthy distaste for being told what to think, or due to the fact that their mistrust may be justified.

TV commentator Chris Hayes encourages us to recognize that those frustrated with institutions constitute a large and powerful segment of society. He suggests that dividing Americans into institutionalists, who want to strengthen and preserve our existing social institutions, and insurrectionists, who see a need to overhaul, overthrow, replace or abandon existing institutions, is at least as useful as dividing the population into liberals and conservatives. Insurrectionists include progressives (Bernie Sanders), libertarians (Rand Paul) and nationalists (Donald Trump), while both Republicans and Democrats are well represented within the institutionalist camp.

The defeat of a consummate institutionalist – Hillary Clinton – by an insurrectionist outsider suggests a need to take rising insurrectionism seriously. What if our citizens now include a large plurality unlikely to be persuaded to regain trust in our central civic institutions?

How mistrust reshapes civics

Assume for the moment that a large group of citizens is mistrustful of existing institutions. How do these citizens participate in civic life?

Low participation in congressional elections is often offered as evidence of the decline in American civic life. But in 2012, only 35 of 435 congressional seats were considered “swing” districts, where voting margins were within 5% of the national popular vote margin – the remaining 92% of districts strongly favor either a sitting Democrat or Republican. The safety of these districts leads to an extremely high rate of incumbent re-election, 95.9%. Combine the very low chance of making a difference in a Congressional election with extremely low trust in Congress (9% in 2016) and it’s easy to understand why many citizens – including some institutionalists – would sit an election out.

When we teach young people how to have a civic voice, we tend to emphasize the importance of voting as a baseline civic responsibility – as the bumper sticker says, “If you don’t vote, you can’t complain.” But at high levels of mistrust, voting doesn’t work very well. If we see Congress, the Senate or the presidency as dysfunctional institutions, either unlikely to accomplish much or to represent our interests, voting for representatives or encouraging them to advance or support legislation doesn’t feel like a powerful way to influence civic processes.

High levels of mistrust present a challenge for protest as well. Unless the goal of a protest – a march, a sit-in, an occupation – is the fall of a regime (as it was with the protests of the Arab Spring), then a protest is designed to show widespread support for a political position and influence leaders. The March on Washington, likely the most remembered event of the civil rights movement as it culminated in Martin Luther King Jr.’s “I Have a Dream” speech, was, after all, a march on Washington. It sought to pressure President Kennedy and Congress to take action on civil rights legislation and is credited with creating the momentum for LBJ to act quickly on civil rights after Kennedy’s assassination.

What happens when protesters no longer trust that institutions they might influence can make necessary social changes? The Occupy movement was widely criticized for failing to put forward a legislative agenda that representatives could choose to pass. Occupiers, in part, were expressing their lack of confidence in the federal government and didn’t put forth these proposals because their goal was to demonstrate other forms of community decision-making. Whether or not Occupy succeeded in demonstrating the viability of consensus-based governance, the resistance of Occupiers to turning into a political party or advocacy organization shows a deep insurrectionist distrust of existing institutions and an unwillingness to operate within them.

The danger is that insurrectionists will drop out of civic life altogether, or be manipulated by demagogues who promise to obviate the complexities of mistrusted institutions through the force of their personal character and will. The hope is that insurrectionists can become powerful, engaged citizens who participate in civic life despite their skepticism of existing institutions. To make this possible, we need to broaden our understanding of what it means to be a good citizen.

There is a tendency to assume that the actions that constitute good citizenship are stable over time. Good citizens inform themselves about issues, vote in elections, contact representatives about issues they care about and, if they fail to be heard, protest peacefully and non-violently. Michael Schudson argues that this model of citizenship is only one of several that has held sway in the US at different moments in our nation’s history. Early in the American republic, “good citizens” would be expected to send the most prominent and wealthy member of their community to Washington to represent them, independent of agreement with his ideology. Later, good citizens supported a political party they affiliated with based on geography, ethnicity or occupation. The expectation that voters would inform themselves on issues before voting, vote on split tickets making decisions about individual candidates or vote directly on legislation in a referendum was the result of a set of progressive era reforms that ushered in what Schudson calls “the informed citizen”.

We tend to see the informed citizen as the correct and admirable model for citizenship a hundred years after its introduction, but we miss some of the weaknesses of the paradigm. Informed citizenship places very high demands on citizens, expecting knowledge about all the candidates and issues at stake in an election – it’s a paradigm deeply favored by journalists, as it places the role of the news as informing and empowering citizens at the center of the political process. Unfortunately, it’s also a model plagued with very low participation rates – Schudson observes that the voting was cut nearly in half once progressive political reforms came into effect. And while we often discuss civics and participation in terms of the informed citizen mode, he argues that America has moved on to other dominant models of citizenship, the rights-based citizenship model that centers on the courts, as during the civil rights movement, and monitorial citizenship, where citizens realize they cannot follow all the details of all political processes and monitor media for a few, specific issues where they are especially passionate and feel well-positioned to take action.

Young people in particular are looking for ways they can be most effective in making change around issues they care about. Effective citizenship, in which individuals make rational, self-interested decisions about how they most effectively participate in civic life, can look very different from the informed citizenship we’ve come to expect. Joe Kahne and Cathy Cohen surveyed thousands of youth in California and discovered that while participation in “institutional” politics (rallies, traditional political organizing, volunteering to work with a candidate) is low, there is strong engagement with they are volunteering at a much higher rate than previous generations, looking for direct, tangible ways they can participate in their communities.

We are beginning to see new forms of civic participation that appeal to those alienated from traditional political processes. One way to understand these methods is as levers of change. When people feel like they are unlikely to move formal, institutional levers of change through voting or influencing representatives, they look for other levers to make movement on the issues they care about.

In his 1999 book, Code and Other Laws of Cyberspace, Lawrence Lessig argues that there are four primary ways societies regulate themselves. We use laws to make behaviors legal or illegal. We use markets to make desirable behaviors cheap and dangerous ones expensive. We use social norms to sanction undesirable behaviors and reward exemplary ones. And code and other technical architectures make undesirable actions difficult to do and encourage other actions. Each of the regulatory forces Lessig identifies can be turned into a lever of change, and in an age of high mistrust in institutions, engaged citizens are getting deeply creative in using the three non-legal levers.

In the wake of Edward Snowden’s revelations of widespread NSA surveillance of communications, many citizens expressed fear and frustration. The Obama administration’s review of the NSA’s programs made few significant changes to domestic spying policies. Unable to make change through formal government processes, digital activists have been hard at work building powerful, user-friendly tools to encrypt digital communications like Signal, Center for Civic Media.)

Effective citizenship means that people look for the methods of social change they see as most effective. Young people often look for norms-based theories of change, taking advantage of their skills in building and disseminating media. Insurrectionists frustrated with legal institutions or with the behaviors of corporate America look for change through new technology and new ventures.

This shift in citizenship is still emerging. Media often hasn’t caught up with the idea that effective civic engagement happens outside the courts, the voting booth and Congress. This understandable overfocus on law-based theories of change leaves those frustrated with institutions frustrated with media as well. For insurrectionists who see Washington institutions as ineffective and untrustworthy, a strong media focus on these institutions can look like an attempt to maintain their legitimacy and centrality.

One of journalism’s key roles in an open society is to help citizens participate effectively. From close scrutiny of those in elected office to analysis of legislative proposals to editorial endorsements of candidates for office, news outlets help their customers make civic decisions. If mistrust in institutions is changing how people participate in civics, news organizations may need to change as well. We can recommit ourselves to explaining the importance and centrality of our institutions, but we run the risk of being insufficiently skeptical and critical, and the danger that we lose even more trust from our alienated and insurrectionist readers. Or we could rethink our role as journalists as helping people navigate this emergent civic landscape and find the places where they, individually and collectively, can be the most effective and powerful.

Dueling spheres of consensus

Shortly after the 2016 elections, a friend asked me to lunch. A Trump supporter, he knew we had voted differently in the election, and we both wanted to talk about the future of the country under the new administration. But he invited me specifically because he was angered by an article I’d written that grouped Breitbart founder Steve Bannon with alt-right leader Richard Spencer.

My friend explained that he read Breitbart religiously, not because he supports white supremacy, but because he supports net-zero immigration to the US as a strategy for raising the incomes of white and non-white Americans. Breitbart was the only major media outlet he found seriously discussing that policy stance. “If Bannon is beyond the pale, and Breitbart’s beyond the pale, does it mean that my views on immigration are beyond the pale? And what about the millions of Americans who agree with me?”

Research that Yochai Benkler and our team at MIT and the Berkman Center confirmed my friend’s assertion that Breitbart covered matters of immigration much more closely than other media outlets leading up to the 2016 election, focusing on the issue more than 3x as often as right-leaning outlets Fox News and the Wall Street Journal. Thanks to the strong influence of Breitbart, we speculate, immigration became the most-reported on policy issue in the 2016 election, despite GOP efforts to soften the party’s stance on immigration to reach Latino voters.

The move of immigration from the fringe of the news agenda to a central topic is a phenomenon addressed by media scholar Daniel Hallin in his 1985 book, The Uncensored War: The Media and Vietnam. Hallin argues that we should think of potential news stories as fitting into one of three spheres. In the sphere of consensus, there is widespread agreement on an issue or a position (democracy is the best form of government; capitalism is a good way to build an economy) and therefore it’s not worth our time to discuss. In the sphere of deviance, there is widespread agreement that a stance is beyond the pale (sexual relationships between adults and minors are natural and should be legal; collective ownership of all goods is the best way to end economic inequality) and also not worthy of discussion. The (sometimes very narrow) sphere of legitimate controversy includes the standard political debates within a society, and journalists are expected to show themselves as neutral on those topics legitimate to debate (tax cuts for the wealthy will lead to economic growth; for-profit insurers will only survive with federally mandated medical insurance).

Lobbyists, activists and PR professionals have used Hallin’s spheres to shape what’s at stake in public policy debates. Health insurance companies have worked hard to push the idea of single payer healthcare into the sphere of deviance, rebranding the idea as socialized medicine to associate it with a disfavored economic idea. By citing the small number of scientists who do not see evidence that humans are contributing to climate change, advocates have kept the phenomenon of global warming within the sphere of legitimate debate.

While Hallin’s Spheres are related to the Overton window – the idea that certain policy prescriptions are so radical that a politician could not embrace them without compromising her own electability – being consigned to Hallin’s sphere of deviance has psychological implications that falling outside the Overton window lacks. Advance a policy suggestion that is outside the Overton window and you suffer the disappointment that your idea is discarded as impractical. Stray outside the sphere of legitimate debate into the sphere of deviance, and your position becomes invisible to mainstream media dialog. Journalism scholar Jay Rosen observes, “Anyone whose views lie within the sphere of deviance — as defined by journalists — will experience the press as an opponent in the struggle for recognition. If you don’t think separation of church and state is such a good idea; if you do think a single payer system is the way to go… chances are you will never find your views reflected in the news. It’s not that there’s a one-sided debate; there’s no debate.”

The growth in media diversity brought about by the rise of the internet and social media means that if your ideas are outside the sphere of legitimate debate, you can simply find a media sphere where you’re no longer in the sphere of deviance. My friend, frustrated that he could not find media debating his ideas on immigration, began reading Breitbart, where his deviant ideas are within the sphere of consensus, and the legitimate debate is about the specific mechanisms that should be used to limit immigration. He is not alone. While less popular than during the 2016 election, Breitbart is the 61st most popular website in the US, close in popularity to the Washington Post. In our data set, which examines how websites are shared on Twitter or Facebook, Breitbart is the fourth-most influential media outlet, behind CNN, The New York Times and politics site The Hill.

The ability to find a set of media outlets compatible with your political views is not new. Even in the days of political pamphlets and early newspapers, it was possible to experience a Federalist or Anti-Federalist echo chamber. The rise of large-circulation newspapers and broadcast media, which needed to avoid alienating large swaths of the population to maintain fiscal viability, led us into a long age where partisan journalism was less common. Even as cable news made partisan news viable again, broadcast news networks and major newspapers maintained aspirations of fairness and balance, attempting to serve the broader public.

Those economic models make little sense in a digital age. As purveyors of wholly manufactured fake news (like the Macedonian teens who targeted content at Trump supporters) know, there is a near-insatiable appetite for news that supports our ideological preconceptions. But it’s important to consider that people seek out ideological compatible media not just out of intellectual laziness, but out of a sense of efficacy. If you are a committed Black Lives Matter supporter working on strategies for citizen review of the police, it’s exhausting to be caught in endless debates over whether racism in America is over. If you’re working on counseling women away from abortion towards adoption, understanding how to be effective in your own movement is likely to be a higher priority for you than dialog with pro-choice activists.

Partisan isolationism is not just purely a function of homophily. The structure of internet media platforms contributes to ideological isolation. While Pariser and others trace these structural effects to Facebook and other highly targeted social media, I argued in Rewire that three different generations of internet media have made it possible to self-select the topics and points of views we are most interested in. The pre-Google web allowed us to self select points of view much as a magazine rack does: we choose the National Review over the Nation, or their respective websites. Unlike broadcast media, which lends itself towards centrist points of view to attract a wide range of ad dollars, narrowcast media like websites and magazines allow more stark, partisan divisions. With the rise of search, interest-based navigation often led us to ideological segregation, either through the topics we select or the language we choose to pursue them – the vegan cooking website is unlikely place to meet conservatives, much as searching for progressive voices on a hunting site can be frustrating. And the language we use to describe an issue – climate change, global warming or scientific fraud – can be thoroughly ideologically isolating in terms of the information we retrieve.

What’s different about social media is not that we can choose the points of view we encounter, but that we are often unaware that we are making these choices. Many people joined Facebook expecting the service would help them remain connected with family and friends, not that it would become a primary source of news. As of 2016, 62% of American adults reported getting some news via social media, and 18% reported often getting news through platforms like Facebook. These numbers are more dramatic for young adults, and likely increased during the 2016 presidential election. Because Facebook’s newsfeed algorithm presents content to you based on content you’ve liked and clicked on in the past, it has a tendency to reinforce your existing preconceptions, both because your friends are likely to share those points of view, and because your behavior online indicates to Facebook what content you are most interested in. Eli Pariser calls this problem “the filter bubble”, building on earlier work done by Cass Sunstein, which recognized the tendency to create “echo chambers” online by selecting media that fits our politics. Pariser argues (controversially) that algorithms used by Facebook and others increase this tendency.

It’s worth noting that the filter bubble problem isn’t inherent to social media. Twitter has pointedly not filtered their timeline, which avoids the filter bubble, but leaves responsibility for escaping echo chambers to the user. While you can decide to follow a different group of people on Twitter, research from Nathan Matias suggests that even highly motivated people are unlikely to make major changes in their online behavior in order to combat biases and prejudices.

Our team at the MIT Media Lab is working on Gobo, a new tool that allows you to filter your Facebook and Twitter feeds differently, using natural language processing and machine learning to build filters that can increase or decrease the political content of your news feed, give you more or fewer female authors, or consciously choose to encounter more news outside of your echo chamber. One of the key questions we seek to answer in buiding the tool is whether people will actually choose to use these filters. One hypothesis we hope to disprove is that, despite complaining about filter bubbles, many people seem to enjoy ideological isolation and may choose settings similar to what they encounter online now.

General interest media, like broadcast television and national newspapers, traditionally saw themselves as having a responsibility to provide ideological balance, global perspectives and diversity in their coverage. (Whether they succeeded is another question – I’ve heard many reports from people of color that they felt invisible in those “good old days” and far more visible in contemporary, fragmented media.) As that business model becomes less viable, because readers gravitate towards ideologically compatible material, it’s worth asking whether platforms like Facebook have an appetite for this work.

Thus far, the answer seems to be no. Facebook has assiduously avoided being labeled a publisher, trying to ensure both an escape from legal liability for content it hosts under the Safe Harbor provisions of US internet law, and to prevent itself from being criticized about exercising poor editorial judgement. The problems Facebook is confronted with are serious. Demands that the platform block “fake news” are challenging, given that most of what’s called “fake news” is not obviously fraudulent. If Facebook begins blocking platforms like Breitbart, it will be accused of censorship of political content, and rightly so.

One possible escape for Facebook is to eliminate algorithmic curation of newsfeeds, moving back to a Twitter-like world in which social media is a spray of information from anyone you’ve chosen to pay attention to. Another is to adopt a solution like the one we are proposing with Gobo, and put control of filters into the user’s hands. It’s an open question whether Facebook would choose a path forward that gives its users more control over their experience of the service.

In considering how platforms enable online discourse, we need to consider the idea that sharing content is a form of civic participation. Part of our emergent civics is the practice of making and disseminating media designed to strengthen ties within an identity group and to distinguish that group from groups that oppose it. Consider the meme-makers competing for $20,000 from Infowars. Many involved don’t believe that CNN is ISIS, as one popular meme allegesas Judith Donath explains, “News is shared not just to inform or even to persuade. It is used as a marker of identity, a way to proclaim your affinity with a particular community.”

Donath’s insight helps explain why factchecking, blocking fake news or urging people to support diverse, fact-based news is unlikely to check the spread of highly partisan news. Not only is partisan news comfortable and enjoyable (I find it reassuring to watch Trevor Noah or Samantha Bee and assume that friends on the right feel the same watching Fox News commentators), spreading this information has powerful social rewards and gives a sense of shared efficacy, the feeling (real or imagined) that you are making norms-based social change by shaping the information environment.

The research Benkler and our Media Cloud team conducted shows how rapidly these partisan ecosystems can come into being. Examining 1.25 million media stories and 25,000 media sources, we gave each media source a partisanship score based on whether people who shared tweets from the Democratic or Republican candidates also shared a story from a source. Stories from the New York Times were more often shared by people who’d retweeted Hillary Clinton than those who’d retweeted Donald Trump, but the effect was much more pronounced with Breitbart: Breitbart was amplified almost exclusively by Trump supporters. Our research shows a tightly clustered set of sites read only by the nationalist right. The vast majority of these sites are very new, most founded during the Obama administration. This community of interest has very little overlap with traditional conservative sources like the Wall Street Journal or the National Review. In our study, those publications are both low in influence and linked to by both the left and right, while the Breitbart-centered cluster functions as an echo chamber.

The emergence of echo chambers like the one around Breitbart further complicates fact-checking. danah boyd explains that in teaching students not to rely on Wikipedia, we’ve encouraged them to triangulate their way to truth from Google search results. On topics covered heavily in the Breitbartosphere but not addressed in the broader media universe, this leads to a perverse effect. Search for information on Pizzagate as the story was being developed on sites like Infowars and you would likely find links to other far-right sites promoting the story. By the time sites like the New York Times became aware of the story and began debunking it, many interested in the faux-scandal had persuaded themselves of its truth through repetition within a subset of closely related websites, to the point where an unstable individual took up arms to “self-investigate” the controversy.

Hallin’s spheres suggests we question whether we are encouraged to discuss a wide enough range of topics within the sphere of legitimate controversy. The problem we face now is one in which dialog is challenging, if not impossible, because one party’s sphere of consensus is the other’s sphere of deviance and vice versa. Our debates are complicated not only because we cannot agree on a set of shared facts, but because we cannot agree what’s worth talking about in the first place. When one camp sees Hillary Clinton’s controversial email server as evidence of her lawbreaking and deviance (sphere of consensus for many on the right) or as a needless distraction from more relevant issues (sphere of deviance for many on the left), we cannot agree to disagree, as we cannot agree that the conversation is worth having in the first place.

Much as there is no obvious, easy solution to countering mistrust in institutions, I have no panaceas for polarization and echo chambers. Still, it’s worth identifying these phenomena – and acknowledging their deep roots – as we seek solutions to these pressing problems. It is worth noting that the research Benkler’s and my team carried out suggests the phenomenon of asymmetric polarization – in our analysis, those on the far right are more isolated in terms of viewpoints they encounter than those on the far left. There’s nothing in our research that suggests the right is inherently more prone to ideological isolation. By understanding how extreme polarization has developed recently, it might be possible to stop the left from developing a similar echo chamber. Our research also suggests that the center right has a productive role to play in building media that appeals to an insurrectionist and alienated right-leading audience, which keeps those important viewpoints in dialog with existing communities in the left, center and right.

Fundamentally, I believe that the polarization of dialog in the media is a result both of new media technologies and of the deeper changes of trust in institutions and in how civics is practiced. The Breitbartosphere is possible not just because it’s easier than ever to create a media outlet and share viewpoints with the like-minded. It’s possible because low trust in government leads people to seek new ways of being engaged and effective, and low trust in media leads people to seek out different sources. Making and disseminating media feels like one of the most effective ways to engage in civics in a low-trust world, and the 2016 elections suggest that this civic media is a powerful force we are only now starting to understand.

Closing questions

I want to acknowledge that this paper may stray far from the immediate challenges that face us around issues of information quality, in the service of seeking for their deeper roots. My questions follow in the same spirit. For the most part, these are questions to which I don’t have a good answer. Some are active research questions for my lab. My fear is that we may have to address some of these underlying questions before tackling tactical questions of how we should best respond to immediate challenges to faith in journalism.

Trust:
– How long does it take to recover trust in an institution that has failed? What are examples of a mistrusted institution regaining public trust?
– Is the fall in institutional trust an independent or a joint phenomenon – i.e., does losing trust in Congress lessen our trust in the Supreme Court or the medical system
– Is trust in news media higher or lower in countries with strong public/taxpayer supported media? Does trust correlate positively or negatively to ad support? Privacy-invading tracking and targeting?
– If people don’t trust institutions, who or what do they trust? How do those patterns differ for more trusting elites and for the broader population?

Participation:
– What forms of participation (from the traditional, like voting, to the non-traditional, like making CNN-bashing memes) are indicators of future civic engagement? Should we be encouraging and celebrating a broader range of civic participation amongst youth? Amongst groups that see themselves alienated from conventional politics?
– Should media attempt to explain and engage audiences more deeply in institutional politics? Will acknowledging the limits of existing institutional politics restore trust in journalism, or damage trust in government?
– Should media celebrate and promote new forms of civic engagement? Will this further decrease trust in institutions? Increase a sense of citizen efficacy?
– What would media designed for increased public participation look like? Are there models in the advocacy journalism space, or in solutions journalism, constructive journalism or other movements?

Polarization:
– Is it reasonable to expect Americans to rely on a single, or small set, of professional media sources that report a relatively value-neutral set of stories? Or is this goal of journalistic non-partisanship no longer a realistic ideal?
– Could taxpayer-sponsored media serve a function of anchoring discourse around a single set of facts? Or will public media be inherently untrustworthy to some portion of American voters? Why does public media seem to work well in other low-trust nations but not in the US?
– Is there a role for high-quality, factual but partisan media that might reach audiences alienated from mainstream media?
– Should media outlets learn from what’s consensus, debatable and deviant in other media spheres and modify coverage to intersect with reader’s spheres? Is shifting the boundaries of these spheres part of how civics is conducted today?