To what degree did outsize media coverage of Donald Trump contribute to his unexpected victory in the race for the Republican presidential nomination? A great deal, according to some vocal political scientists. Not so much, retort many journalists, who insist they report the news rather than shape it.
The disagreement has generated a lengthy Twitter dustup between advocates for each side. But who’s right? The answer depends in part on how one defines media influence.
Political scientists certainly do not believe that journalists formed a cabal that sought to tilt the Republican nomination race toward Trump. Instead, drawing on several decades of media research, they argue that although journalists might not have told voters whom to choose in the Republican contest, they almost certainly influenced how that choice was made, both by the quantity of news coverage they allocated to Trump and its tone.
Journalists, however, defend their work. As Molly Ball, a journalist with the Atlantic, recently argued, voluminous coverage "is the effect … not the cause" of Trump’s popularity. For the media to have scaled back Trump coverage out of some sense of civic responsibility would have amounted to "deliberately put[ting] its collective thumb on the scale" — a serious violation of the journalism credo.
Still, although it is not airtight, the political scientists (disclosure: I am one) have the stronger case.
Early studies were fueled by the rise of mass media
To better understand their claim, it’s worth briefly touching on some of the more important findings in the academic literature regarding media effects. Dating back to the nation’s founding and the era of the partisan press, journalists have routinely been accused of acting on behalf of a political party.
However, the rise of the electronic media during the 20th century — first radio, then television — raised concerns that this new medium might allow those controlling its message to shape public opinion on a whole new level. (Witness the panic reportedly induced by Orson Welles’s "War of the Worlds" radio broadcast.) Those fears sparked efforts by political scientists to determine whether, and to what degree, media coverage influenced voters.
Given the extensive media coverage devoted to presidential campaigns, it’s not surprising that researchers looked there first for evidence of media effects. In two classic studies of the 1940 and 1948 presidential campaigns, Columbia University scholars Bernard Berelson, Paul Lazarsfeld, and associates conducted repeat interviews of voters in two communities to gauge media influence on voters’ choices, asking them about their consumption of radio and newspaper coverage and their political preferences.
They found very little evidence that the media exerted an independent influence on whom voters decided to back. They found that the voters’ choices owed much more to their partisan predispositions and socioeconomic standing, influences that tended to get reinforced by their immediate social environment — such as friends and family — than to media coverage. However, the media did serve an informational purpose by helping voters decide what the election was about — what the major issues were and where the candidates stood on them.
The most likely time for media effects? Primaries and caucuses.
Subsequent research that focused more heavily on television coverage of presidential elections came to similar conclusions. However, as Vanderbilt political scientist Larry Bartels’s research suggests, a general election campaign for the presidency is probably not the best place to find evidence of media influence. By that stage in the contest for the presidency, most voters already have strong views toward the two major candidates, and party identification plays a powerful role in driving their vote.
It’s not surprising, then, that in the 2012 presidential election nearly 70 percent of voters made up their minds on whom to vote for vote prior to September. Lacking new and pervasive information from the media that contradicts their views, voters aren’t likely to change such preferences after that. Moreover, the professional norms that drive how journalists cover elections tend to produce somewhat uniform and balanced campaign coverage focused on the horse race and on candidates’ personality traits. That’s also unlikely to change minds.
But nominating contests are different. In those, there are more candidates, party cues play less of a role, and voters start with lower information levels and hence greater uncertainty regarding whom to support. Under these conditions, one might expect to see stronger media effects.
Agenda setting, priming, and framing
But what kind of effects? In an important study, political scientists Shanto Iyengar and Don Kinder used doctored television stories and identified two specific types. The first, which they called "agenda setting," refers to how story placement influences the importance placed on it by viewers. Stories that lead the news, or that are shown more frequently, are deemed by viewers to be more significant (particularly when they corroborate viewers’ personal experiences).
Thus, when viewers were exposed to a doctored broadcast that began with a segment on civil rights, they indicated in follow-up questionnaires that the topic was of greater importance to them than did a control group of viewers who were not exposed to the story. (Black viewers who saw the story were even more likely than white ones to be influenced by the civil rights coverage.)
"Priming," the second media effect Iyengar and Kinder identified, refers to the way the press can influence the standards by which audiences evaluate a candidate. Thus, when the researchers exposed viewers to programs focused on national defense, and then asked them to evaluate a president’s performance, viewers were more likely to judge the president by how well they thought he had provided for the nation’s security (as opposed to, say, domestic issues).
Although these particular effects have been corroborated by additional studies, it is important to note some potential limits when applying the experimental findings to the real world, in which people have independent means of assessing what the press tells them, and in which they are often exposed to multiple news sources carrying competing messages.
When discussing government-backed health care, for instance, proponents will frame it as a fundamental right, while opponents will "counter-frame" it as socialized medicine. Research indicates that media framing, in isolation, can help to shape opinions. But counter-frames may mitigate the initial framing effects, particularly for people with weakly held opinions on a particular topic. (For those with strong opinions, counter-frames may actually lead them to double down on their initial beliefs.)
Some studies also suggest that media effects may rapidly decay over time. These caveats notwithstanding, decades of research into the media and public opinion indicates that media coverage doesn’t have to change voters’ underlying attitudes toward candidates to influence how they decide to vote.
A billion dollars in free media surely makes a difference
How might these media effects have contributed to Trump’s victory? The most obvious effect came about through the sheer number of news stories about Trump’s candidacy, many of them appearing in agenda-setting positions on front pages and at the top of news shows. Numerous studies have documented that Trump received far more media coverage than did his Republican rivals, even before any votes were cast. For example, one study finds that Trump received almost a billion dollars’ worth of free media coverage through February — an amount dwarfing that of any of his rivals.
Political scientists believe this outsize coverage likely produced several effects. First, it signaled that Trump’s candidacy was something to take seriously, rather than a novelty act that viewers might dismiss. Moreover, the disproportionate coverage of Trump’s views on issues like trade and immigration made these issues more salient to voters, meaning they were more likely to consider them when choosing a candidate.
Conversely, Trump’s media domination meant that his 16 Republicans rivals, and the issues they might like to see highlighted, were not getting beneficial exposure. Molly Ball’s defense of the media notwithstanding, it appears that by focusing coverage so heavily on Trump, journalists did, however inadvertently, put their thumb on the scale.
What about the argument that the media coverage followed Trump’s popularity rather than inspiring it? Importantly, the heavy media focus on Trump began before his rise in the polls and in the absence of other traditional indicators of candidate strength, such as campaign fundraising prowess. That’s suggestive of media influence, but research also backs up the idea that causality runs in that direction.
For both the 2012 and 2016 Republican presidential races, political scientists Kevin Reuning and Nick Dietrich analyzed daily data from the start of election polling up to the Iowa caucus. They looked at public interest in candidates (gauged by online searches), polling support for candidates, and media coverage on major cable news stations. They found that increased media coverage influenced the polls — not vice versa.
In politics, it's good to be the lead horse
But Trump benefited from more than disproportionate media coverage. He was also helped by the horse race frame most of the press adopted (which again, works differently in the primaries than in a two-party contest).
Harvard’s Thomas E. Patterson found that more than half of the coverage of Trump focused on some aspect of his standing in the race, including his position in the polls, his debate performance, or the size of his crowds. In contrast, only 12 percent focused on his issue stances or political beliefs, including his incendiary comments. "Trump is ahead" gets interpreted by voters as a positive frame, the research shows.
And even when the media did focus on Trump’s position on the issues or his inflammatory comments, the potential negative effect of covering Trump’s policies was weakened by the media’s pursuit of balance by quoting supporters of those stances.
In sum, it is true, as Ball argues, that voters "are the ones who decide — not the media." But the research shows that voters’ decisions are predicated on information about candidates that is partly gleaned via media campaign coverage.
Those academic studies suggest that the volume and tone of information provided by the media influenced the way voters viewed Trump’s candidacy, and contributed to his surge in the polls and his eventual victory. (Though surely Patterson takes things a step too far by calling Trump "arguably the first bona fide media-created presidential nominee.")
What should we expect next?
Looking ahead to the general election campaign, press coverage is less likely to benefit Trump. As usual, a majority of voters will already have made up their minds by the end of the political conventions. In a two-person race, Trump will likely no longer receive a major preponderance of news coverage, and with polls showing him trailing Hillary Clinton he is unlikely to benefit from the media’s tendency to prime voters to view the election as a horse race.
Moreover, his messages on issues like immigration, trade, and economic inequality are likely to be subject to aggressive counter-framing by the Clinton campaign. The media will cover this counter-framing out of an interest in balance, and this will likely undercut his appeal with undecided voters.
The media almost certainly contributed to Trump’s unexpected victory in the race for the Republican presidential nomination. But the general election is likely to present another media story — one less favorable to Trump’s chances.
Matthew Dickinson is a professor of political science at Middlebury College and author of the Presidential Power blog.