Americans are sharply divided along partisan lines when it comes to the media outlets they turn to and trust for their political and election-related news, according to a new Pew Research Center study.

Amy Mitchell, director of journalism research
Amy Mitchell, director of journalism research at Pew Research Center

More Democrats and Democratic-leaning independents trust than distrust most of the 30 outlets in the study, but the reverse is true among Republicans and GOP leaners. And while Democrats’ trust in many of these outlets has remained stable or in some cases increased since 2014, Republicans have become more alienated from some of them, widening an already substantial partisan gap.

Amy Mitchell has directed the Center’s journalism research since 2012 and oversaw the new study, which is based on an online survey of more than 12,000 U.S. adults. The study serves as the framework for our new Election News Pathways project. In this Q&A, she answers key questions about how the analysis was done and what it says about Americans’ news habits as the first votes of the 2020 presidential election cycle loom.

The study examines public trust in 30 U.S. media outlets for political and election news. How did you choose these outlets? Why weren’t some well-known outlets included?

A table showing the outlets asked about in this study

With today’s vast and fractured media landscape, our goal was not to do anything like a census. Instead, we wanted to choose a variety of news outlets with substantial audiences across different platform types. To that end, we included major broadcast and cable TV networks, public broadcasters, political radio shows, high-circulation national newspapers, high-traffic digital news outlets and international news sources with a substantial readership in the United States, among other kinds of outlets. Most of the outlets we studied were part of a similar study we published in 2014, which allowed us to track whether partisans’ trust in them changed over time.

One group not included here are wire services like The Associated Press and Reuters. While those organizations certainly produce a great deal of original reporting, our study is not an assessment of news brands, but an analysis of outlets Americans turn to for news and the trust levels of those outlets. Most Americans get news from the wires through another news outlet that carries their syndicated content. We also took into consideration things like web traffic, topic focus and responses to open-ended survey questions about people’s main sources for political news.

Social media sites as a source for political news were asked about separately and will be a part of a future analysis.

How did you measure Americans’ trust and distrust in these outlets?

We first asked our survey respondents whether they had heard of each of the 30 outlets. If so, we asked them if they trusted it for political and election-related news. If they didn’t indicate trust in the source, we asked them if they distrusted it. After all, there’s a difference between simply not expressing trust in an outlet and actively expressing distrust of it. In cases where respondents had heard of a source but didn’t indicate that they trusted or distrusted it, we classified the response as “neither trust nor distrust.”

These questions allowed us to measure things a few different ways. For example, we could examine trust gaps between different groups of people for different news outlets. We found that CNN is trusted by 70% of self-described liberal Democrats, but only 16% of conservative Republicans – a gap of 54 percentage points. Conversely, Fox News is trusted by 75% of conservative Republicans but only 12% of liberal Democrats – a 63-point gap.

Asking about both trust and distrust also allowed us to examine the “ratio” between these two measures for each outlet. Fox News, for instance, is among the sources in this study trusted by the largest percentage of Americans, with 43% of U.S. adults saying they trust it for political news. But it is also among the sources with the largest portion who distrust it – 40% of Americans express this view. Since that’s the case, we classified Fox News as “about equally trusted and distrusted.”

Besides asking our survey respondents about their trust and distrust in different outlets, we also asked them if they had gotten political or election-related news in the past week from any of the sources they told us they had heard of. This allowed us to surface some other interesting findings, like the fact that some Americans use news outlets even if they don’t express trust in those outlets.

How did you measure changes in trust over time?

We examined political polarization in Americans’ media habits in a major 2014 report, and since 20 of the 30 media outlets we examined in our new study were also part of the earlier report, we could assess how trust in these outlets has changed.

One of the biggest changes we saw was increased distrust among Republicans for 14 of the 20 news sources included in both studies, with particularly notable increases in distrust of CNN, The New York Times and The Washington Post – three frequent targets of criticism for President Donald Trump. While there has been far less change on the Democratic side, two exceptions are The Sean Hannity Show and Breitbart News, which are now distrusted by a larger share of Democrats than in 2014.

It’s important to point out that because of differences in methodology and survey design, the new study is not comparable in all ways with the 2014 study. For example, the new survey is representative of the total U.S. adult population, while the older survey was based only on web-using U.S. adults. And the questions asked, while similar, are not identical in all cases. But there are more points of continuity than differences, and we feel confident in the broad changes in trust and distrust that we’ve documented in the two studies.

The study refers to the audiences of different news outlets as “left-leaning,” “right-leaning” or “mixed.” How did you make these determinations? And are you saying that some of these outlets themselves are “left-leaning” or “right-leaning”?

This study doesn’t make any determination about where news outlets themselves fall on the ideological spectrum based on either the content of their reporting, their self-identification or the views of their editorial boards.

I’ll answer the last question first because it’s a crucial point to understand. This study doesn’t make any determination about where news outlets themselves fall on the ideological spectrum based on either the content of their reporting, their self-identification or the views of their editorial boards. This project wasn’t designed to evaluate outlets themselves or the content they produce.

Instead, we used an approach that grouped each outlet according to the ideological composition of its audience, based on where our respondents told us they get political and election-related news and how they describe themselves ideologically – liberal Democrat (including independents who lean Democratic) or conservative Republican (again including leaners). We’ve used this same system in past studies about the media.

In the new report, outlets classified as having “left-leaning” audiences are those with at least two-thirds more liberal Democrats in their audience than conservative Republicans, and outlets with “right-leaning” audiences are those with at least two-thirds more conservative Republicans than liberal Democrats. Those whose audience does not fall into either of these are classified as having more mixed audiences.

Here are three real-world examples that show this system in action. The share of Business Insider readers who identify as liberal Democrats is at least two-thirds higher than the share who identify as conservative Republicans (40% vs. 20%, or 100% larger), so we classified Business Insider as having a “left-leaning” audience. On the other end of the spectrum, the share of Washington Examiner readers who identify as conservative Republicans is at least two-thirds higher than the share who identify as liberal Democrats (44% vs. 14%, or 214% larger), so we categorized the Examiner as having a “right-leaning” audience. Then there’s the middle: The Wall Street Journal is defined as having a more “mixed” audience because the share of its readers who identify as liberal Democrats is not two-thirds higher than the share who identify as conservative Republicans (31% vs. 24%, or just 29% larger).

It’s important to keep in mind that just because we classified a news outlet as having a “left-leaning” or “right-leaning” audience does not mean that majority of its audience identifies as either liberal Democrat or conservative Republican. In fact, relatively few of the outlets we studied have audiences that consist mostly of liberal Democrats or conservative Republicans.

Why does the study include more outlets with left-leaning audiences than right-leaning audiences?

We selected these outlets based on a number of factors, including their audience size and platform type, but not based on the ideological orientation of their audiences, which we didn’t measure until later in the research process. Using this method, we ended up with 17 outlets whose audiences are left-leaning, six outlets whose audiences are right-leaning and seven outlets with mixed audiences.

One factor that may be at play here is that Republicans have a more compact media ecosystem. They rely to a large degree on a small number of outlets and view many established brands as not trustworthy. Democrats, on the other hand, rely on a wider number of outlets.

What do you hope readers will take away from this study?

It’s often tempting to use studies like this one to “rank” media outlets against one another in terms of trust or distrust, but that wasn’t the purpose of this research. Instead, we wanted to offer insight into the news sources partisans rely on for political news, and the degree to which there is common ground or division. That’s especially important in an election year like this one. (Throughout the campaign, in fact, we’ll be applying this research to do additional analyses as part of our Election News Pathways project, which will let users do their own analyses with an interactive data tool.)

Overall, these findings reveal sharp divides in the use and trust of political news sources. They don’t reveal completely separate media bubbles. There are some news sources that both Democrats and Republicans turn to, but even those areas of overlap can be hard to fully gauge since using a news source doesn’t always mean people trust it.

John Gramlich  is an associate director at Pew Research Center.