Addressing the usual throng of press reporters earlier this month, Ted Cruz said, “The mainstream media loves talking about Donald Trump” – a statement that was instantly relayed by the media, complete with a loud reference to Trump in the headline. The irony doesn’t end there. In the past few weeks, countless articles have wondered aloud whether too many articles are being written about Donald Trump.
Is the media to blame for Donald Trump? The risks of navel-gazing and delusions of grandeur are high when the media itself takes on that question. Nonetheless, this data journalist’s analysis suggests there are several strong reasons to take on the question, and several strong reasons to raise doubt about the conclusion.
The stakes in the question are high (if the media truly is to blame then perhaps we could also shape who becomes the next US president). And a majority of Americans agree that the media is indeed very powerful. Earlier this month 68% of Americans told YouGov that they believed the mainstream media bore at least some responsibility for violence at Trump events.
Although the question is important, it’s still problematic. Asking whether journalists are to blame for Trump assumes that there’s something wrong with Donald Trump, his success and/or his supporters. And the question also assumes that the media is in a position to fairly judge the media. I’ll try to minimize my own subjective bias in what follows by sticking as closely as possible to the data at hand.
Trump had the highest name recognition to begin with
Media coverage, good or bad, can be helpful for someone looking to enter the White House. There’s a pretty clear correlation between support for a candidate and simple name recognition. That’s especially important early on in presidential campaigns when many poll respondents might be wondering “Marco who?”. As Clifford Zukin, a member of the executive council of the American Association for Public Opinion Research, explained to NPR in August last year “I think [early polls] measure name recognition – and that’s something, but is name recognition really the way you want to choose a presidential candidate?”.
But on this measure Trump had a considerable lead on Ted Cruz, Marco Rubio or any of his other rivals right from the outset. If Americans hadn’t already heard of Trump Tower, Trump Hotel or Trump Entertainment Resorts, they might well have seen Donald Trump appear on their screens while watching World Wrestling Entertainment (he was inducted into the WWE celebrity hall of fame in 2013) or been one of the 28.1 million Americans to have watched Trump host the hit NBC reality show The Apprentice.
The New York billionaire was already an American cultural reference point about New York and about billionaires before he became a reference point in American politics.
When Trump announced he would be running for the presidency, the media coverage he received may well have boosted his recognition to some poll respondents. But the fact is, most of them already knew him so it’s hard to argue that Trump gained a great deal from his initial press coverage.
Avoiding coverage may not mean a better outcome
If the mere fact of mentioning Donald Trump could unfairly contribute to the candidate’s success, is the alternative to not mention him? Anyone ready to make such a suggestion would do well to look to Sweden’s recent history.
From the mid 1990s, the far-right party Sweden Democrats began to grow, but the media were reluctant to take note of what many felt was an unpalatable force in the country’s politics. Party members were excluded from political debates and banned from publishing party advertising in the media. As one Swedish journalist Niklas Orrenius who has studied the movement for years explained, “Most journalists detest them, so they don’t write about them seriously”.
That media neglect helped the party foster a narrative that they were the underdogs, willing to say the things that a “politically correct elite” and liberally biased journalists would not or could not say. The press realized that their strategy had been counter-productive, but the party was already mainstream. Socialist magazine Jacobin explains:
Now, media outlets across the country make sure to include the party and its platform in opinion polling, and take pains to avoid describing it as racist or xenophobic.
The audience demand for coverage of Trump is high
Whatever your stance on how journalists should come up with stories, in practice, demand influences supply. News outlets tend to publish more stories that they know people – lots of people – want to read. Google searches are one of the best available indications of the public’s demand for information. On Trump specifically, searches for the candidate were initially higher than the number of media articles produced about him as shown in the chart below from FiveThirtyEight.
In the interests of transparency, I asked our own audience research team about the behavior of Guardian readers. In March, Guardian data shows that there were 30m global page views to articles tagged with “Donald Trump” compared with 15m views on articles tagged with “Hillary Clinton” and 12m views about “Bernie Sanders”. Unfortunately, we don’t have numbers that control for the volume of stories about each of those candidates so understanding where those readers came from is important.
Overall, 34% of traffic to articles about Trump came from internet searches and an additional 46% was as a result of readers clicking on Trump articles from a Guardian homepage. In other words, readers are seeking out content about Donald Trump and one could argue that the media has a responsibility to provide it.
Perhaps, then, it’s not a question of whether the media is to blame for covering Trump’s bid for the presidency, but whether the media is to blame for how they covered that bid?
Some voters are highly resistant to fact checks – and mainstream media generally
One argument is that the media neglected to sufficiently hold Trump to account by fact-checking his various claims. “When he says something that’s patently false, start at ‘you’re either lying or you’re completely uninformed’ and work your way back,” explains Chez Pazienza in the Daily Banter this week. “Hit him back. That is, objectively, the right thing to do because you’re dealing in fact and he isn’t.”
The problem is, there have been a large number of online articles that have fact-checked Donald Trump on a whole range of issues, but they don’t seem to have slowed down the groundswell in support for the candidate. TV journalism may have asked fewer tough questions of Trump (a supposition that’s hard to measure given that so few TV scripts are published online), but the speed of 24-hour news means it’s also practically tough for those journalists to conduct rigorous reality-checking research.
Back in 2008, rumors began to circulate that Barack Obama was a Muslim – a claim that was often used in the context of discrediting the presidential candidate (and thereby exposed the hostility of some Americans towards Muslims). Despite being factually disproven, it’s a claim that many still believe. A survey released on Tuesday by Public Policy Polling found that 52% of Republicans still think Barack Obama is Muslim. Among Trump supporters, that number climbed to 62%. Fact-checking articles have arguably done little, if anything, to change those numbers.
In statistics, we try to be watchful for “confirmation bias” – that’s the human tendency to interpret numbers in a way that conveniently confirms your world view rather than challenging it. Unfortunately confirmation bias also guides our online clicks and we’re all guilty of it (I’m ill-positioned to criticize Trump supporters who do not read my fact-checking articles if I refuse to be persuaded to click on their blogs claiming that Obama is a Muslim).
American confidence in the media is at a historic low, and that’s a fact that’s even more stark among the country’s Republicans. In 2015, Gallup found that just 40% of Americans said they had a “great deal” or “fair amount” of trust in the mass media. Just 32% of Republicans said the same.
Fact-checking articles may prove impotent in the face of such distrust. Worse, they may actually heighten those attitudes. In December 2015, Rasmussen found that 47% of likely voters think most reporters are biased against Trump – only 23% felt that Clinton was the subject of unfair coverage.
This election has provided further evidence that media assumptions about voters are wrong
Not everything can be quantified. Those unknowns are important in understanding the real connection between the media and Trump’s ascension. We know that Trump supporters feel ignored and frustrated by the mainstream establishment. But we don’t know how the initial media coverage, which ridiculed them, might have fueled those feelings. We know that some white Americans feel that while they suffered hardship in the financial crisis, African Americans were unfairly protected by Obama. But we don’t know how the media portrayal of Trump supporters as racist (a label many of them might reject) has further enforced the view that their grievances are not heard. And it’s simply hard to measure how the fact that just 7% of reporters identify as Republican might have affected truly objective reporting about Donald Trump.
The loaded question asked this week, “is the media to blame for Trump?” demonstrates the dangers of the shaky assumptions that have dominated this election – assuming that Trump would probably never make it past 20% support, assuming that Trump is one social phenomenon (rather than a symptom of many) and assuming that Trump supporters are simply misguided in their choice for the next American president.
The question also assumes that American opinions are flexible and that the media is playing a large role in changing those minds. We might need to wait until this election is over and the votes are counted to see how reliable those assumptions are.