Fear and Loathing in the American Electorate: Part 3
Media consumption — Where do voters get their information and what difference does it make?
This is Part 3 of an eight-part exploration of the 2020 American National Election Study, focusing on the motivations and sources of information driving the American electorate, especially the Republican/conservative voters who cast their votes for Donald Trump in 2020.
Note (4/13/22): A recent update of the ANES 2020 dataset corrected some errors in the media usage data reported in previous versions of the dataset. Several graphics in this article have been updated to reflect this corrected data (changed graphics are so indicated in their captions). The corrected data actually gives media bias a larger role in this story than it appeared to have before. Other articles in the report will be updated with corrected data as well. Overall findings remain unchanged.
Here are quick links to the other seven parts.
- Part 1: The most important question facing American democracy today
- Part 2: Demographic predictors of 2020 vote
- Part 4: Seven “deplorable” beliefs that predict 2020 Presidential vote
- Part 5: The role of conspiracies and misinformation in the 2020 election
- Part 6: Partisan animosity
- Part 7: Combining sources of behavior — Path models of the 2020 Presidential vote
- Part 8: Republicans have crossed the Rubicon — Conclusions and implications
Media exposure, the quality of information presented by different media outlets, the role of social media, and people’s levels of trust in the media as a purveyor of accurate information — all these topics generated significant discussion and debate throughout the 2020 campaign, and indeed, throughout the tenure of the Trump Administration.
Trust in the media
The 2020 ANES surveys includes a wealth of information on all these topics, at a very granular level. At the top of this data trove are questions about trust in media. And here — not unexpectedly after five years of Donald Trump decrying the media as “fake news” and “the enemy of the people” — trust vs. distrust in media is revealed to be a major fault line in American public opinion.
Fully 86.7% of Americans who voted for Donald Trump in 2020 reported little or no trust in the news media.
This rather stunning data point has serious implications for how people absorbed and evaluated information about the campaign. It also raises a number of additional questions. What media exactly do Trump voters distrust? All media? Mainstream media? Liberal media? What do they think about conservative media? Do they distrust it as well, or do they not consider it a part of “the media” they distrust? Does distrust imply a refusal to follow media content, or does it imply a willingness to follow media content, but only to dismiss it as “fake news”? Additional questions about media consumption provide some surprising answers to these questions.
Media accessed to follow the campaign
In the pre-election ANES survey, people were asked whether they followed the campaign and, if so, in what media formats: TV programs, newspapers, internet sites, radio shows, or none of the above. Despite the high level of media distrust expressed by Trump voters, their reported levels of media access to follow the campaign were quite comparable to those of Biden voters.
Even though Trump voters expressed high levels of distrust in the media, 85% of them followed the campaign on TV, 69% followed it on internet sites, and 55% followed it on radio news. The only format Trump voters accessed significantly less than Biden voters was newspapers, which only 32% of Trump voters followed, compared to 47% of Biden voters. These findings answer one question about the meaning of “distrust”:
Although Trump voters overwhelmingly claimed to distrust of the media, they remained active consumers of media across a variety of platforms during the 2020 campaign.
Their levels of media access were quite similar to those of Biden voters, who expressed much higher levels of media trust.
Respondents who followed the campaign in any of these formats were asked what specific programs and outlets they accessed at least once a month over the previous year. To complete this task, they were shown a list of over 100 programs, newspapers, and internet sites — from Sean Hannity (both TV and radio shows) to CNN (both TV network and website) to Rachel Maddow. Some of these shows were pure entertainment “fillers”, like American Idol and NCIS. Others could be classified as either conservative, liberal, or mainstream political news sources. There were 66 of these.¹ Here are the top 25 political media outlets mentioned, sorted by (weighted) count of mentions.
This is an interesting list, for several reasons. First, even the most frequently mentioned media source — the CNN website — was visited at least once a month by only 17% of the public. Second, the top five media sources were websites, not TV programs. Third, of the TV programs mentioned, both liberal and conservative outlets enjoyed similar audience sizes (generally between 6–8% of the weighted sample). And finally, the top sources include only a smattering of conservative and liberal-leaning TV programs; for the most part, they include mainstream or middle-of-the-road media outlets, such as the ABC and CBS evening news shows and the BBC and USA Today websites.
A lot of speculation about media usage has revolved around the idea of media bubbles, that is, a tendency for people only to patronize media sources that reinforce and reflect their own biases and beliefs.
The ANES surveys provide substantial support for the idea of media bubbles.
Looking at some of the most popular conservative and liberal TV shows, we find very little overlap in their audiences. Whichever side of the political spectrum a show caters to, its audience remains predominantly on that side. Only a mainstream show like NBC Nightly News appears to attract a more diverse audience.
All of these differences are statistically significant.
When it comes to TV news, it appears that media bubbles among both liberals and conservatives were intact and highly biased during the 2020 Presidential campaign.
Relatively exclusive media bubbles also appear is in the radio and website formats. Here are audience breakdowns by Presidential vote for some popular radio shows and websites, both conservative and liberal-leaning.
As with the TV shows, these partisan radio shows and websites drew audiences predominantly from within their own ideological bubbles. Although generally reaching smaller audiences than TV programs, these outlets appealed to a more selective and receptive audience of fellow believers. This was especially true for conservative radio shows, which Biden voters essentially ignored.
Social media accessed during the campaign
Another aspect of media usage that is often cited as a source of conservative thinking and behavior is social media usage. A popular notion is that social media platforms like Facebook and Twitter are breeding grounds for conservative conspiracies and misinformation dissemination. Implied in this characterization is an assumption that social media exposure is more a phenomenon of the right than the left. An extensive battery of questions on social media usage in the 2020 ANES surveys yields a somewhat different perspective on this issue.
In the ANES 2020 post-election survey, respondents were asked which social media sites they visited in the last year. The results confirmed extensive social media usage, but not just by conservatives. In fact, as with political TV programs, Biden voters accessed most of these sites at rates equal to or greater than Trump voters.
Social media sites are as much a playground for liberal voters as conservative voters.
Respondents were also asked how frequently they visited selected social media sites,² on a scale ranging from “many times every day” to “less than once a month”. Converting those answers into a rough indicator of “days per month” accessing social media, I found more evidence that social media usage occurs at similar rates for Biden, Trump, and ‘other’ voters, with Biden voters actually accessing social media slightly more frequently than Trump voters.
Looking at these social media visit frequencies compared to liberal-conservative self-identification provides another snapshot into the partisan distribution of social media usage.
Social media fixation is definitely not an affliction exclusive to the conservative side of the American political divide.
If anything, the more liberal a person’s self-identification, the more time they are likely to spend on social media.
Attributing a preference for Donald Trump to greater social media involvement by his followers does not appear to be supported by the ANES data, at least not in terms of the frequency of social media access.
One more aspect of social media usage covered in the ANES 2020 surveys is posting behavior. Perhaps visiting social media sites occurs at similar rates on both sides of the political spectrum, but what about posting? It turns out that posting is also fairly equally balanced between eventual Biden voters than Trump voters. Reported rates of posting on Facebook are statistically equivalent, at just under 30% for both groups. Biden voters are more than twice as likely as Trump voters to post on Twitter. Reddit, the well-documented bastion of some of the most extreme Trump supporters, barely registers as a posting platform compared to Facebook and Twitter.
Media bias
Media bias is often cited as a source of extreme political beliefs and attitudes. This can be tested with the ANES media exposure data. First, all the media sources mentioned can be classified into three categories: conservative, liberal, and mainstream. Conservative and liberal media are marketed and promoted as providers of opinions that support liberal or conservative political agendas. Mainstream media, in contrast, are marketed and promoted as providers of information that is meant to be objective and factual (although they do not always meet that goal). Of the 66 political news sources covered in the ANES survey, it is pretty easy to classify them: 22 had a conservative agenda, 8 had a liberal agenda, and 36 could be classified as mainstream media sources.
Beginning with these raw counts, I created three media usage metrics: cmedia_scaled, lmedia_scaled, and mmedia_scaled. Each is a count of media sources mentioned, then standardized to a 0-to-1 range, so scores are comparable even though there are a different number of sources in each category. Comparing these measures for Biden and Trump voters yields an interesting pattern.
As expected, eventual Trump voters accessed much more conservative media and much less liberal media than eventual Biden voters. An additional big difference between the two groups was in their exposure to mainstream media. Biden voters accessed over 3 times as many mainstream media outlets as Trump voters. This may provide a clue as to the meaning of “distrust of media” for Republican voters.
What Republican voters distrust most, compared to Democratic voters, appears to be mainstream media sources, not liberal media sources.
To explore the effects of media bias on political attitudes, voting, and other political behavior, I created another metric called mbias. For each respondent, I first subtracted their lmedia_scaled score from their cmedia_scaled score, producing a measure that is negative for respondents with a liberal media bias, positive for a conservative media bias, and zero for those whose media usage is perfectly balanced or completely absent. The magnitude of this measure captures the strength of one’s media bias in either direction. Second, I converted this measure into a z-score to produce a dimensionless metric normalized around its mean and standard deviation. The result is a measure of media exposure that trends positive for Trump voters and negative for Biden voters, as would be expected.
Media bias (mbias) joins mainstream media exposure (mmedia_scaled) as a significant predictor of Presidential vote.
There is also a fair amount of overlap in mbias scores between Biden and Trump voters, with some Trump voters having a liberal media bias and some Biden voters had a conservative media bias. Vertical grey lines in the comparative histograms below show the mean value for each group. For mbias, means are -.51 for Biden voters, +.70 for Trump voters. The other media consumption predictor of Presidential vote, mmedia_scaled, shows greater differentiation between Biden and Trump voters, with Biden voters on average accessing twice the proportion of mainstream media outlets as Trump voters (19.8% for Biden voters vs. 6.6% for Trump voters).
A natural question is whether media bias and mainstream media exposure are associated with trust in the media. They both are.
People who have some or a lot of trust in the media tend to have a more liberal media bias and greater mainstream media exposure than people who have not much trust in the media (a combination of none and a little). This tendency for media bias and mainstream media exposure to correlate with ideological orientation extends all along the liberal-conservative spectrum, where each step closer to the conservative extreme is associated with a (usually significant) increase in media bias and a decrease in mainstream media exposure.
Continue to Part 4: Seven “deplorable” beliefs that predict 2020 Presidential vote
References
- A number of Spanish language programs were also included in the survey. I did not include those in the present analysis.
- Facebook, Twitter, and Reddit.
