Politics & Trust in the COVID-19 Pandemic
Updated: Jun 3, 2020
A survey of how political affiliation shapes attitudes and behaviors regarding coronavirus
By: Marilyn Salzman, Lori Eronimous, and Bill Skeet
Have you wondered whether political viewpoints influence a person's diet of COVID-19 news and information? Our survey of 429 people across the US suggests political affiliation plays a role in the COVID-19 information people view as important, which sources they trust, and how they navigate the coronavirus information ecosystem. Specifically:
Political segments have different views on what’s important.
Different political affiliations rely on different trusted sources.
Attitudes about what’s important and trusted shapes how people navigate the web to find answers.
Perceived flaws with the COVID-19 ecosystem align with what’s viewed as important.
Attitudes and behaviors concerning COVID-19 vary by political affiliation.
Few consumed a balanced diet of information, trusting only one domain of information (news, medical, or government) instead of drawing across all three.
This article is part of a series in which we discuss lessons learned from our COVID-19 research. Additional details about our methodology are in the About this study section of this post.
Politics & what’s important
As discussed in Trusted News and Information in a Pandemic: A Survey of COVID-19 News Practices and Perceptions, we asked people to “Briefly describe what information about COVID-19 is most important to you?” Several specific themes emerged:
Statistics (numbers of cases, deaths, and recoveries)
Prevention, safety, and spread
Symptoms, testing, and treatment
Local or information “near me”
Economic or business impact
Other factors such as when it will be over
Let’s take a closer look at what people in each political segment (Democrat, Independent, Republican, and Unregistered/other) said was important.
Democrats stood apart from other parties in their focus on statistics.
“Fatality rate, transmission rate, how many people are infected, how many recovered, how many died.” — Democrat
"Now if I want to get more specific, this is one of the things I check as well (searches for COVID19 numbers) the numbers--seeing if they’re rising, if the curve is flattened. I also make sure I check specifically for my state (Michigan)." — Democrat looking for statistics
In this video, a participant shows how she quickly navigates the Johns Hopkins site to get statistics.
Republicans were unique in their mentions of the economy or business impact and when the pandemic will end.
“Solutions for working with COVID. We need telework, delivery/pickup/cyber school option to become part of the new primary way, not a special circumstance of these weeks.” — Republican
"Another site I go to a lot is Yahoo News. I like to see what’s going on, what celebrities and other folks are saying about it instead of just the Fox News broadcast channel ...and I like to see business issues related to coronavirus." — Republican looking for business issues
Independents cited facts, science, and truth notably more than other parties.
“The truth. It is important to know exactly what is going on without political bias.” — Independent
“Scientific facts.” — Independent
Unregistered/other voters seemed to care most about their own safety, mentioning prevention, safety, and spread, local or “near me”, and social distancing more than their peers.
“Local infections, deaths, and recoveries. Local testing and testing state and nationwide.” — Unregistered/other voter
“How we can stay healthy, how much we should social distance.” — Unregistered/other voter
Politics & trusted sources
When we asked “Which sources do you trust to provide COVID-19 information? (List up to 5 in order of trust.)” here is what participants told us:
News was the most trusted.
Medical and health organizations were also critical.
Government ranked lower as a trusted source than the above.
Many people turned to search, data/news aggregators, or social media, and even friends, family, or self for COVID-19 information.
Few consumed a balanced diet of information, trusting only one domain of information (news, medical, or government) instead of drawing across all three.
Here, we take a closer look at how trusted sources varied by political affiliation.
Across the political segments, media sources were the most trusted and trusted sources tended to align with what each segment viewed as most important. Democrats, who said they most valued statistics, tended to list media and medical sources as the sources they trust. Independents, who wanted facts, are going to medical sources and search and data aggregators, and they seemed to be less trusting of the media than the others. Republicans, who are focused on the economy, turn to the US government more than the others. Unregistered/other voters were more oriented to local media, and family and friends than their peers.
Let’s take a closer look at the specific sources cited. Figure 3 shows word clouds representing how often each source was mentioned. They illustrate how different each political segment’s mental model is when it comes to trusted sources. Word clouds were generated on wordclouds.com.
Democrats said they trusted the CDC, CNN, and other news sources such as the NY Times, NPR, and MSNBC. Governors and Fauci also emerged as highly trusted sources. Very few Democrats cited Trump or Fox News. Independents, like Democrats, trusted the CDC and CNN, but they also were more apt to cite local and general news sources. Individuals such as Trump and Fauci were more commonly cited. Republicans trusted Fox News and the CDC, and quite a few also listed CNN. Trump and the Federal government were trusted more by Republicans than any other party. Unregistered/other voters were the least specific about their trusted sources, with many of them simply citing news, local news, or the internet. Unregistered voters were also unique in their focus family and friends as trusted sources of COVID information.
The table below summarizes the frequencies for commonly cited sources.
As discussed in our earlier article, we also looked beyond specific sources to look at trust profiles, which examine the aggregate set of sources for each individual. Our analysis to date reveals a few interesting trust profile tendencies:
About 12% of the participants followed only one source. This was slightly higher for Independents and Unregistered/other voters (18% and 15% respectively) than for Democrats or Republicans (8% and 11% respectively).
A much higher percentage of participants (41%) followed five or more sources, the maximum number of sources we allowed them to report. Slightly more Democrats (46%) fell into this category than their cohorts, for whom only 36% to 38% listed five sources.
While most people (85%) reported more than one source, very few (10%) consumed a balanced diet that triangulated information from news, media, and government. Nearly one in three (31%) cited sources from only one of these venues, and this was consistent across the political segments.
Where we are currently seeing the biggest differences in trust profiles is in the specific sources people follow. As shown in Table 1, Fox News is dominant for Republicans, with one-third (33%) following them for COVID-19 news. CNN is primary for Democrats, but good percentage of them also seek news from other major news outlets (New York Times, NPR, ABC, NBC, etc.). It is also noteworthy that CNN was frequently listed (by 17 to 18%) as a trusted reference for the other segments as well.
Politics & navigating COVID-19 information on the web
In our task-based sessions using TryMyUI, 30 participants took us on a tour of the information sources they were using to monitor COVID-19 information. Immediately, it became evident that their political viewpoints influenced how they navigated the news ecosystem.
In this video clip, a Democrat participant described the news sources she avoids.
In contrast to the Democrat above, this Republican describes how she relies on Fox News for coronavirus information.
In this video clip, an Independent participant trusts the information he finds on Reddit rather than the information he finds on Google.
Here, a participant describes why he avoids government websites.
Politics, attitudes, and behaviors
In addition to asking people about the information they value and sources they trust, we asked them to rate their experience with COVID-19 information along several dimensions:
Thinking about the websites/apps you have used, how easy or difficult was it for you to find the COVID-19 information you were looking for? (1 = very easy; 5 = very difficult)
Again, thinking about the websites/apps you have used, how confident are you in the accuracy of the information you found? (1 = not at all confident; 5 = extremely confident)
Thinking beyond COVID-19, are you more or less concerned about the quality of news from websites/apps today as compared to six months ago? (1 = much less concerned; 5 = extremely concerned)
How worried or not worried are you about the impact of COVID-19 on the following factors: your own health, your family and friends’ health, your financial well-being, the US economy? (1 = not at all worried; 5 = extremely worried)
Political affiliation did not appear to play a significant role in ratings of ease of use, confidence in the accuracy of information, or concerns about quality. However, affiliation did seem to affect the extent to which people were worried about COVID-19. As Figure 4 shows, on average, Democrats were the most worried and Republicans the least.
After asking people to provide their own list of trusted sources, we invited them to rank a preset list of information domains (news, search, etc.). Rankings were consistent with their free-response sources. On average, participants, and Democrats in particular, rated the news media as the most frequently used source. US government sites were the least frequently used by all segments except the Republicans, who referenced these sites second only to the news media. People in other political segments reported going to search or news aggregators or health organizations more than government sites. As one of our task-based research participants pointed out, government websites are notoriously hard to use, so this could be a factor in the lower frequency of use. It might also be a side-effect of trust. As we noted earlier, the US government was not a highly trusted source of news, which is unfortunate during a pandemic resulting in over 100,000 deaths in the US to date. See Figure 5 below.
When we asked people to recall which of a preset list of national news sources they had referenced in the previous 48 hours, CNN was top of the list for every political segment but Republicans, for whom Fox News was popular. The New York Times and NPR were popular for Democrats. See Table 2.
When we asked people to recall which of a preset list of data sources they had referenced in the previous 48 hours, the CDC was very popular across all political segments. Google’s COVID-19 tracker, which had recently been promoted by the White House, was also a top reference, particularly among the Unregistered voters. See Table 3.
While the relationship between political affiliation and behavioral measures was not statistically significant, we did see emergent trends in which Republicans appeared to take more risks than their counterparts. These trends may warrant a closer look in future research.
More Republicans said they ventured out of the home (e.g., for work or exercise) than the other political segments.
Republicans reported taking slightly fewer precautions (e.g., avoiding close contact with people outside of the home, disinfecting surfaces, and avoiding touching eyes, nose, and mouth) than their peers.
Worry was a more significant predictor of these behaviors. And, as discussed earlier Republicans were slightly less worried than other segments, and Democrats in particular. As such, this may explain why they reported taking fewer precautions than others.
Note that we also looked at these behaviors by gender, age, and stay-at-home orders, and did not see notable trends that would further explain these behaviors.
Politics, flaws in the ecosystem, and desired improvements
When we invited participants to suggest improvements to news and information ecosystem sources, we saw cross-partisan consensus for truthful reporting. Overall, about 32% specifically mentioned they wanted more facts and unbiased truth. Democrats had the most to say about improving the ecosystem, with 78% of this segment offering suggestions. They were unique in their desire for more statistics, data and graphs, easier to use and local information, and less from President Trump. Republicans wanted to see less political bias and more positive news. Independents expressed a desire for more medical sources in the news and more timely data availability.
This video highlights some of the difficulties participants have in establishing relevant/timely information.
Unregistered/other participants wanted to see the ecosystem provide more accurate and truthful reporting, and, overall, were less likely to offer specific recommendations. Democrats and Unregistered participants said they wanted more local information. Generally, suggestions aligned with the trends we described earlier in what each segment viewed as important.
About this study
This study combined survey and task-based research to understand how and where people get news and information about COVID-19, as well as the sources they trust. We conducted this research in early April 2020 (April 3 through April 16, 2020) at the start of the COVID-19 pandemic. At that time, many states had issued stringent restrictions and COVID-19 dominated the news.
The survey was US-based with 429 participants and was built and administered via SurveyMonkey. The majority of participants were recruited via a SurveyMonkey panel; a smaller sample of 30 participants came from a US-based TryMyUI panel. The TryMyUI participants also took part in an observational study from which we received videos detailing how they perform specific COVID-19 tasks. Tasks included a tour of each participant’s top COVID-19 information sites and answering questions about COVID-19 worldwide, US, and state cases, social distancing guidelines, testing procedures, and approved medicines (of which there were none at the time). For findings from the task-based research, see COVID-19: Looking for answers on the web. To the extent possible, the sample was balanced by the survey tool for gender and age.
As shown in Figure 6, our sample was geographically diverse, with representation from each of the four census regions in the US, and coverage from 46 out of 50 states. Unrepresented were Delaware, Maine, South Dakota, and Vermont.
Figure 7 shows the participant breakdown for key segments. Most participants were between the ages of 18-60 (88%) and the rest over 60. The gender breakdown was 57% female and 43% male. Political affiliation was 38% Democrat, 23% Republican, 25% Independent, and 14% Unregistered.
Our analysis and research continue. You can find these posts on our COVID-19 Research blog or connect with our authors on LinkedIn or Facebook: Marilyn Salzman, Bill Skeet, and Lori Eronimous. We'll be sure to announce new publications there.