Election Day has arrived. The final stage of the contest has featured two unique presidential candidates: a former first lady who might become the first female president of the United States, and a billionaire-turned-reality TV star who defied conventional logic time and again on the campaign trail. The news media have deployed abundant resources to cover an electoral process like no other in recent American history. But how influential has this coverage been?

In my Nieman Lab prediction for journalism in 2016, I wrote:

As the United States gears into a presidential election year, we can expect that during 2016, leading news organizations will spend considerable effort and resources covering the political process. They will hope that their solid coverage will matter greatly, and therefore will act accordingly with their usual fanfare. But this coverage will likely play a derivative role to the more central, mundane communication practices that will connect political leaders with the citizenry in social media.

Over the past several months, most of the traditional, leading news organizations in the United States have consistently and almost uniformly engaged in negative coverage of Donald Trump. Detailed reporting and elaborate opinion pieces have dug into every conceivable aspect of the Republican nominee’s sexist, racist, and xenophobic statements. The leading news outlets have also scrutinized Trump’s business operations, from the management of his casinos and real estate projects to his tax-filing practices, and reported profusely on the less-than-positive aspects of these operations.

The negative coverage of Donald Trump has come not only from traditionally liberal sources, but from across a relatively broad portion of the ideological spectrum — even though some critics on the left have argued that sometimes the media created a false equivalence between the two presidential candidates in the pursuit of objective reporting. A clear indicator of the breadth of negative coverage of Trump is the distribution of newspaper and magazine endorsements of both candidates. Hillary Clinton was endorsed by 229 dailies and 131 weeklies, including news organizations that historically have not been identified with either party and others clearly representing a conservative ideology normally linked to Republican candidates. By contrast, Trump received the endorsement of 9 dailies and 4 weeklies. That’s a 27-to-1 difference.

Although counterfactuals are always difficult to evaluate, I think it’s reasonable to argue that, for instance, during the second half of the twentieth century, even a fraction of this negative coverage would have been enough to seriously damage the chances of a presidential candidate, and maybe even derail the candidacy altogether. By contrast, as Americans finish casting their ballots, the RealClearPolitics average of major polls shows Trump only 3.2 points out of the popular vote lead.

To understand how this is possible, we must look at the dynamics of political communication on social media. During the 2008 and 2012 electoral cycles, the campaigns of Barack Obama vastly outperformed those of John McCain and Mitt Romney on social media. The current electoral cycle has seen a dramatic reversal of this situation: Donald Trump’s social media operation has clearly surpassed that of Hillary Clinton. As of late last week, Trump’s Facebook page had accumulated 11.9 million likes; his Twitter account had 12.9 million followers. Clinton’s numbers were 7.8 million and 10.1 million, respectively. (There are other reasons why candidates have a social media presence, including to elicit donations and collect email addresses from potential volunteers, but this is beyond the scope of this analysis.) In an era of click farms and bots, not all likes and followers are genuine, but there’s no reason to believe that this can account for a major portion of these divergent figures between the two campaigns.

In light of Trump’s sometimes outlandish communication practices, it’s possible that a portion of his social media public has followed him to check on the latest outrageous post rather than because they support his candidacy. However, engagement metrics also show the much greater success achieved by Trump’s strategy.

For instance, as of Friday morning, a post about a campaign appearance made 14 hours before on the Facebook page of Trump had 92,000 likes, 40,000 loves, and 29,000 shares — and the video included in it had 2.1 million views. By way of comparison, a post made 12 hours earlier on Clinton’s Facebook page, also about a campaign appearance, had 14,000 likes, 1,300 loves, and 1,965 shares — and the video included in it had 218,000 views. Not only has Trump’s social media messaging elicited higher engagement than Clinton’s, but the discrepancy in intensity has been striking —witness the “loves” gap between those two posts. Similar comparisons on Twitter, such as the number of retweets and likes, show similar patterns.

The divergence between news media and social media happens in a historical period in which there has been a dramatic contraction of the former and a phenomenal expansion of the latter. For instance, according to the most recent annual report that The New York Times Co. — not only the crown jewel of American journalism but most importantly a corporation betting its future on all things digital — filed to the SEC, total revenues had been in the $1.5 billion range every year between 2011 and 2015, with net income in excess of $63 million for 2015. Contrast with Facebook: revenue growth from $3.7 billion in 2011 to $17.9 billion in 2015, with a net income of $3.6 billion in 2015. While revenues at The New York Times have remained stagnant, with a profit margin of 4 percent last year, Facebook’s revenues quadrupled and reached a profit margin of 20 percent.

Despite the differences in their strategies and products, all media — including the news and social media — play in the same market competing for the attention of the public. The stark discrepancy in the performances of The New York Times and Facebook is an indicator of the distribution of people’s media attention in contemporary society. Advertisement expenditures follow this distribution, and a look at where ad dollars went last year tells a conclusive story: 65 percent of digital display ad dollars spent in America went to five companies, none of which are in the news-creation business. Facebook was the top performer, receiving 30 cents of every dollar — a number that goes up to 38 cents for mobile display ads.

These figures about financial and advertising matters signal the growing importance of social media relative to news media. It also begins to account for why a stronger presence in social media might have allowed the Trump campaign to counter its overwhelmingly inferior standing among the news media. But we also need to look at the character of people’s practices to get a fuller picture of the potential relative influence of political communication in different media.

Informal evidence suggests that people, on average, visit social media sites more than a dozen times a day. So it’s not surprising that according to one report, social media consumption has become a top non-work, non-sleep time activity, second only to preparing and eating meals and significantly more prevalent than physical activity for the average person. Spending time on social media has become so pervasive in people’s daily lives that it’s engulfed their news consumption habits.

Preliminary findings of an ongoing research project I’ve been conducting with colleagues about the dynamics of news and entertainment consumption show that most people get the news primarily as part of their social media use, usually using their smartphones. That is, people increasingly learn about current events not as an activity that is the center of their attention, but as an incidental outcome of consuming Facebook, Twitter, or Snapchat on the small screens of their mobile devices. Most of the time, people concentrate only on the headline and the lede; sometimes they might click on a story, but if they do, they very rarely read it in its entirety. The attention spent on news stories is usually ephemeral and occurs in the midst of focusing on a wide spectrum of information items — from pictures of a friend’s new pet to reports of a relative’s exotic trip.

This reduction of most stories to a handful of words or short video segments, along with their blending into a stream of information about daily life, contributes to a decrease in the potential influence of the news within the typical information practices of most people. In addition, the commercial priorities of a company like Facebook shapes the algorithmic logic of its News Feed: The happier we are, the more likely the ads shown to us will be effective, so the algorithm prioritizes information items that are consistent with our viewpoints. So even if we were presented with a large number of news stories and paid significant attention to them, the likelihood of obtaining information that exposed us to alternative viewpoints and helped us learn something new would be relatively low. This algorithmic logic further insulates people from the influence of news media stories that could potentially alter preexistent political preferences.

Needless to say, the media are only one factor shaping electoral preferences, and certainly not the most important one. Furthermore, Trump is a fairly unique candidate in modern political history, in part due to his strong presence in show business. But even after taking all of this into account, I believe that this election will prove to be a turning point in the nexus that connects the news media with political campaigns.

The stark contrast between editorial dynamics and electoral preferences might lead to two trends directly affecting the news media in the short-term future. First, a decrease in expenditures in election ads on news media and a parallel increase in expenditures on social media. This could have a major negative effect on broadcast news, in particular local television stations. This, in turn, would diminish the resources available for journalistic coverage, further reinforcing the downward trend in news media influence.

Second, we might also see an adaptation of the presentation of news stories to match the character of social media consumption practices in mobile devices: a stronger focus on the headline and the lede, often conveyed with a sensational tone, and a decrease in extensive treatment of the news report. This would further intensify the trend towards soundbites and move political discourse away from complex conversations — a dynamic for which the 140-character ethos of social media is better suited than the news media.

If they materialized, these trends would not bode well for the future of the news media. But sometimes, as this year’s winner of the Nobel Prize in Literature wrote in the now very distant sixties, “You don’t need a weatherman / To know which way the wind blows.”