The argument of this chapter is straightforward. During the 2016 American presidential election, the elite journalism practiced at top-tier media outlets was better than it has ever been—the most nuanced, accurate, and fact-oriented journalism of the modern media age. This journalism, however, mattered less to the conduct of American politics than ever before, for reasons only partly attributable to structural transformations in the ecosystem of news production or audience “filter bubbles.” The reasons are also partly cultural: the 2016 election witnessed the clash of a journalistic tribe increasingly driven by commitments to facticity and a nuanced form of objectivity, and a populace increasingly prone to seeing data-driven objectivity as an elitist form of cultural discourse. It is not even that many readers failed to be persuaded by journalistic truth but that they found the aesthetics of that reporting to be alienating and disempowering. This chapter focuses on various forms of quantitative journalism (“data,” “computational,” or “interactive” journalism) as a lens through which to understand these media dynamics, particularly insofar as they relate to American political culture more broadly.
The chapter begins by briefly recapitulating the general factors that may have been at work in determining the manner by which data and quantitative journalism played out in the 2016 election: filter bubbles, the decline of the so-called “mainstream media,” and general notions of our “post-truth” culture. The chapter then traces the historical emergence of data-driven forms of objectivity and discusses the manner by which data journalism came to be an “elite” form of journalistic discourse. It concludes by discussing the “audience” or “reception” side of the equation, speculating about how information and information display might be interpreted as an aesthetic style and “taste signifier,” a style with deeper implications for the American public and political practice.
One final note: it should be clear that all of these explanations for the 2016 election outcome are highly mediacentric, and we should not lose sight that political press coverage does not exist in a vacuum. Alongside media narratives about candidates, actual candidates also exist, candidates with campaign staffs, organizational infrastructures, political skills, and professional competencies. Society may or not be undergoing what Hepp and Coultry (2017) have called “deep mediatization,” but personalities and world events continue to exert an effect on political outcomes, even those buried deep within our media funhouse hall of mirrors.
It is important to keep in mind from the outset that observers have meant many different things when they’ve talked about journalistic “failures” during the 2016 campaign. Some are talking specifically about polling failures, or the fact that predictive polling outlets like the FiveThirtyEight and The Upshot gave Donald Trump low chances of winning. Others are referring to the fact that reporters somehow failed to understand the mindsets and attitudes of a large swath of white working-class voters who defected from Barack Obama to Donald Trump, giving him the narrow margin of votes that propelled him to an Electoral College victory. In this piece I refer less to any of these specifics than to the fact that many journalists (a) feel like they somehow “missed the story” of the election (a grave journalistic sin according to the canons and codes of the profession) and (b) produced fact-driven reporting that somehow failed to have a the anticipated impact on its audience. Many journalists might disagree with my assessment of the situation; nevertheless, I think the core of my argument lies in my claim that there was a disconnect between journalism and the American electorate, that this disconnect had real public consequences, and that it is partially based on “empirical tribalism.”
What other explanations have been offered for this disconnect between journalists and their public? There have been, primarily, three: Facebook and other social media platforms have created digital “filter bubbles”; there has been a decline in the economic health and vitality of the more traditional, centrist, and geographically dispersed news media; and we live in a “post-truth” or “post-fact” culture.
The question of filter bubbles has received a great deal of public attention in the past six years or so, spurred on by Eli Pariser’s book of the same name. According to this argument, because of the algorithmic and financial incentives of media platforms and search engines, audience members are increasingly exposed only to points of view they already agree with (Pariser 2012). This selective exposure then leads to a hardening of extreme political attitudes and a clustering of people with the same beliefs who also reinforce each other’s opinions.
A second point of view, most recently detailed by Politico media critic Jack Shafer (Shafer and Doherty 2017) but also discussed widely by scholars analyzing transformations in the American news industry, points to the “hollowing out” of the American news industry. For these critics, the economic and professional decline of newspapers has led to both a diminution of critical political coverage in places like Pennsylvania, Wisconsin, and Michigan, but also to a corresponding “clustering” of news companies in cities like San Francisco, New York, Los Angeles, and Washington, DC.
The third argument is more general and cultural: we live in a post-truth or post-fact era, epitomized by the Oxford English Dictionary choosing “post-truth” as its word of the year for 2016. “Post-truth,” they wrote in a blog post, “is an adjective defined as ‘relating to or denoting circumstances in which objective facts are less influential in shaping public opinion than appeals to emotion and personal belief’” (Oxford English Dictionary 2016). According to many observers of politics in 2016, traditional and enlightenment-indebted journalists made a mistake in thinking that their appeals to reasoned, empirical evidence would mean much to an electorate steeped in emotionalism and irrational belief.
All of these explanations have their merits, and, as will be seen, my own diagnosis draws on elements of all of them. To understand the genealogy of journalism’s empirical failures, however, I argue that it is worth taking a trip back in time to understand how some of journalism’s most quantitative, data-driven forms of news came to be. In the history of this data-driven journalism, I argue, lie many of the roots of our current difficulties.
Journalists have always used numbers, statistics, and other forms of quantitative information in news reports. Indeed, for much of the early history of printing, journalism was largely about numbers, serving as it did of a conveyor of business, shipping, and trade records along with political gossip and rumors from the royal court. It is only with the emergence of what we know today as “modern reporting” that journalism became less material and more oral, less about numbers than words, less about documents than interviews, and transformed, in Hazel Dicken Garcia’s terms, from a “record” to a “report.” This is why, for me, any American history of data journalism must begin not with the first use of infographics or statistics, but rather in the early twentieth century when a more general transformation of social knowledge began to take place. This transformation was propelled forward by the dual impulse of progressive political goals (collecting empirical knowledge as a mechanism for generating political reform) and the more general professionalization of knowledge disciplines (sociology, political science, economics, and journalism all assumed their current professional forms in the early decades of the twentieth century). In my book Apostles of Certainty: Data Journalism and the Politics of Doubt (Anderson 2018), I make the argument that we can best understand the sociomaterial roots of journalistic epistemology by looking at it historically and also comparatively, by examining how it intersected with sociology, political science, and other knowledge disciplines.
For the purposes of this chapter, it is important to keep in mind the following general line of development. The earliest sociology in the 1910s was often indistinguishable from muckraking journalism, particularly in its focus on mobilizing quantitative information for the purposes of social reform. Nevertheless, the vast majority of journalism in the progressive era rejected the insertion of quantitative data as a source of evidence for news stories, owing to its presentist focus and its primary reliance on oral forms of evidence. As sociology professionalized, it rejected its reformist, journalistic, visualizing tendencies and focused on the statistical establishment of quantitative certainty. Ironically, at just the same time, journalism began to act much as sociology had a few decades earlier, embracing context, interpretation of structural events, and data visualization. In other words, both sociology and journalism professionalized in the 1930s, but in ways that pulled them apart and changed what it meant to be properly “scientific.” When the data journalism pioneer Philip Meyer initiated his crusade to make journalism more scientific he would do so on the terrain established by sociology in the 1930s. Journalism would try to become more like social science, rather than sociology trying to be like journalism, as was the case in the 1910s.
This entire process also called into question the status of “the public” and its relationship to social science and to the deployment of quantitative information. The early reformist sociologists believed data could literally “activate” the public, that viewing or consuming data could spur people and groups to political action. Modern sociology abandoned much of this pretense, concerned far more with its own internal field dynamics and, at most, with the relationship between its experts and the policy elite. Professional journalism, for its part, largely understood the role of reporting to be the provision of information to citizens for the purposes of self government, and the crux of the debate between Meyer and other journalists in the 1960s concerned the degree to which journalism should provide a new form of information (“social science in a hurry”) to the public in order to better facilitate democratic goals. Meyer’s critics contended that his “precision journalism” would be inaccessible to many readers and impossible for many journalists to produce; Meyer countered that computers would facilitate reportorial capacity, and moreover, that policy makers and interested citizens would use this new approach for democratic ends.
Data journalism, then, was always pitched as a more elitist practice meant for more discerning readers, and this epistemological tendency was reinforced by the way it institutionalized in the 1980s and 1990s. One path forward for data journalism would have been for it to find its institutional home within newsrooms themselves, with every cub reporter gaining a certain level of statistical literacy in the same way that journalists learned to write a nut graf or conduct an interview. This was not, however, what took place. Instead, data journalists began gathering under the banner of “computer-assisted reporting,” a cross-newsroom group of journalists familiar with computers and statistical techniques, who largely practiced forms of the “social science in a hurry” recommended by Meyer. These computer-assisted reporters formed the group NICAR (the National Institute for Computer Assisted Reporting) that later affiliated with a second group, IRE (Investigative Reporters and Editors). This, in turn, tied data journalism even more closely with forms of investigative reporting—itself one of the most respected and elite forms of newswork. This cluster of underpinnings—the link between data journalism and investigative reporting, the fact that CAR stories often won journalistic prizes, the unfortunate reality that a large number of reporters were numerically illiterate, and the fact that data journalism was always pitched to a more policy-focused audience—reinforced the elitist tendencies of data reporting.
Industry-wide changes in journalism have only exacerbated these divisions between ordinary, superficial, tabloid, emotionalist, or partisan reporting and more high-level data journalism work. As Rasmus Kleis Nielsen puts it, alongside the growth of superficial, impressionistic, and transient news items distributed through smart watches and social media, the twenty-first century has also seen
[t]he parallel and simultaneous growth in forms of digital news that are far closer to the “knowledge-about” end of [William] James’s spectrum, forms of long-form, explanatory, data-enriched journalism that offers mediated, publicly available, forms of news very much concerned with questions of causality and teleology, with the relations between events, and that offers this in a far more accessible and timely fashion than other forms of “knowledge about” current affairs. (Nielsen 2017)
Clearly, the type of journalism advocated by Meyer and his followers has played a large part in this journalistic efflorescence. Digital sites that dominated the public conversation about the 2016 American presidential election—Nate Silver’s FiveThirtyEight, the New York Times’s Upshot, ProPublica, the Marshall Project, and other forms of high-level data-driven political reporting—can be seen as the heirs to Philip Meyer as well. But how have these professionalized forms of objective communication been received by audiences? My answer is: not well, and the reasons for this problematic uptake go beyond simply providing information itself. Simply providing highly factualized discourse would not in and of itself be a problem unless there were not deeper fractures and fissures within the American polity, and if citizens were not as prone to thinking about politics in terms of identity as they were in terms of factual knowledge (see Kreiss, this volume). Filter bubbles and business model failures play a role in this, but the primary culprit, I argue, is the aesthetic style of “intelligence” and the populist reaction to it (Perlstein 2017).
In other words: in the partisan and polarized American political environment, professional journalistic claims to facticity have become simply another tribal marker—the tribal marker of “smartness”—and the quantitative, visually oriented forms of data news serve to alienate certain audience members as much as they convince anyone to think about politics or political claims more skeptically. The problem of the public discussed a few pages earlier has re-emerged under conditions of digitization. Data journalism and other forms of quantitative newswork mark the further extensions of journalistic skill and professionalism we have yet seen in the news business. Unfortunately, what is good for the journalism industry is not always good for democracy. Quantitative news, partly because of its own institutional history and partly because of political polarization in the United States, is an aesthetic style that alienates some and confuses others; it is not “value-neutral,” whatever the journalists who practice it might hope for.
Journalism theorists and critics, looking at the current state of American politics and the press, have generally offered two suggestions for how the press might produce a journalism that fairly represents more of the public and has a greater impact on political outcomes. The first is what I call the “discursive” solution: journalists ought to do a better job including the voices, perspectives, and thoughts of so-called “ordinary Americans” in their coverage. Understanding the perspectives of so-called “middle America,” the theory goes, will lead to a fairer, more accurate journalism. A second solution might be called the “emotionalist” solution—rather than doubling down on facts, mainstream journalists should take a page from the Fox News playbook and, as President Donald Trump might put it, aim for the gut.
I’m skeptical that either of these solutions will be successful. The discursive solution seems to place too great an emphasis on dialogue as a process, with objective information as its input that then leads to positive democratic outputs. I also find it hard to imagine journalism ever de-professionalizing to such a degree that it is willing to be as emotional and colorful as might be required to match the right-wing media machine and its obvious disdain for facts. Whatever it does in the future, however, it seems clear that journalism’s confidence in its own professional certitude has been shaken, and that new paths forward are be required if the media is to fulfill its tremendously important democratic functions in the years ahead.