Research reveals the practices and perceptions of the public regarding COVID-19 news and information.
By: Marilyn Salzman, Bill Skeet, and Lori Eronimous
We surveyed more than 400 people across the US in early April 2020 at the height of the COVID-19 pandemic to understand how and where people get news and information about COVID-19, including what sources they trust and what channels they use. We also watched how a subset of the participants use the internet to find answers to key questions about the coronavirus. This article is the first of a series in which we will discuss the lessons learned from this research.
Summary of Findings
Most important information. What people say is most important about COVID-19 is often multi-dimensional. The most common needs include statistics about cases, infections and deaths, and prevention, safety, or spread.
Most trusted sources. The large majority of participants (86.0%) cite a source from the following information domains as their most trusted source: the news media (41.3%), medical community (33.6%), or government (11.2%). The number one most trusted source is the CDC, with 18.9% of the participants putting it at the top of their list.
Few consume a balanced diet of information. Nearly one in three (30.1%) survey participants list sources from only one of three information domains: news media, the medical community, and government. Far fewer, one in ten participants (10.0%), report a diverse set of trusted sources drawing from all three domains.
People want high-quality information. Survey participants suggest that the COVID-19 information ecosystem could be improved by ensuring information that is accurate, trustworthy, impartial, unbiased, and sourced from medical and health experts.
Overview & methodology
The online survey was conducted between April 3 and April 16. It was a US-based survey leveraging an initial geographically diverse email invitation garnering 12 participants, followed by a SurveyMonkey panel of 379 participants, and a TryMyUI panel of 38 participants. 30 of the TryMyUI participants also took part in task research from which we received videos detailing how they perform specific COVID-19 tasks. To the extent possible, the panels were balanced by the survey tool for gender and age.
As shown in Figure 1, we have fairly good representation from each of the four census regions in the US, and coverage from 46 out of 50 states. (Unrepresented are Delaware, Maine, South Dakota, and Vermont.)
Figure 2 shows the participant breakdown for key segments. Most participants are between the ages of 18-60 (88%) and the rest are over 60. The gender breakdown is 57% female and 43% male. Political affiliation is 38% Democrat, 23% Republican, 25% Independent, and 14% Unregistered.
What information is most important?
To understand what topics mattered the most, we asked: “Briefly describe what information about COVID-19 is most important to you.” We allowed an open-ended response so participant responses were not constrained. We analyzed these responses and several themes emerged.
Multi-dimensionality. For many people, the answer to this question is complex. They want answers to a multi-dimensional set of questions, as illustrated by the participant responses below.
"It’s important for me to know where the virus is spreading and the number of cases around me. It’s also important for me to know the current social distancing guidelines as well as if antibody tests come out."
Statistics. Two out of five participants (41.6%) say they are most interested in statistics about cases, deaths, infection, and recovery. Common terms describing information needs include: “statistics” or “stats”; “number” of cases, infections, or deaths; “rate” of infection, spread, or death, “recent” or “new” cases.
"Up-to-date stats on infection numbers and deaths, but also recoveries."
"Demographics of the COVID-related deaths, vaccine and treatment progress, cases in my immediate area."
Prevention, safety & spread. More than a third of the participants (34.0%) say that they want information about the prevention, safety, or spread of COVID-19. Common descriptors are: “stay healthy” or “stay safe”; “prevention”; how to “protect” myself or loved ones; how the virus is “spread”; how “contagious” it is or how it is “transmitted”.
"How to protect you and the ones you love from it and the effects it's having on the world."
Symptoms, testing, & treatment. Another important topic is how to recognize symptoms and get testing or treatments, with 17.9% of participants citing this as most important. Common terms include: “symptoms”, “testing” or “treatment”, “medication” or “vaccine”.
"Where are we on a cure or vaccine, how are treatment plans progressing, what are the plans to help people medically?"
Local information. The desire for information at the state or city level is also evident from the comments, with 15.4% of the participants citing this as important. We also see this theme in their news consumption, as discussed later. Participants often use these phrases: cases or information “near me” or “my location”; effect on “my” community, area, family; “local” statistics; “city”, “county”, or “state” guidelines, testing, etc.
"The number of cases in my state, the number of cases in my county, school information in regard to my university, stay at home orders and anything my state decides."
Other. Other themes include:
Government response
Unbiased facts or scientific truth
Economic or business impact
When will it end?
Social distancing
General fears
"When will the world open up again?"
“The truth. It is important to know exactly what is going on without political bias.”
What sources do people trust most?
Next, we asked: “Which sources do you trust to provide COVID-19 information? (up to 5 in order of trust).” We allowed open-ended responses so each participant could describe up to five sources, starting with their most trusted source, in their own terms. Figure 3 is a word cloud illustration of the most trusted sources. The larger the text, the more frequent the source.
Many participants (41.0%) are quite specific about their most trusted source, identifying institutions and people by name. These include the CDC (listed by 18.9% of the participants), CNN (8.4%) Fox (4.0%), the governor of their state (4.2%), President Trump (3.0%), and Dr. Fauci (by 2.6%). Figure 4 shows a breakdown of these sources.
Many other most trusted entries are more general, including local or national news, local or state government, health organizations or departments, and even the internet or search. Figure 5 breaks down the domains of information that people tap into.
What other sources do people trust?
Survey participants could list up to five trusted sources when asked: Which sources do you trust to provide COVID-19 information? (up to 5 in order of trust). As shown in Figure 6, people tend to reference more than one source.
On average, participants cite 3.4 sources (out of a minimum of 0 and a maximum of 5) with over 40% listing 5 sources. Note that some participants have more than five trusted and stacked up a few similar sources for a single response. We counted these only once in our analysis, although they contribute to the word clouds and counts for specific sources (e.g., CDC, CNN, etc.). Also of note are the non-trivial number of single-sourcers who list only one source (12.2%) and no-sourcers who say they don't have a trusted source (2.8%).
A look beyond the number of sources to the types of sources, or information domains, shows that definitive information domains -- the news media, medical community, and government -- make up 81.7% of all trusted sources listed. See Figure 7.
News is dominant. When people describe trusted sources for COVID-19, news sources are dominant (46.2% of all sources listed). CNN and Fox account for 22.7% of these. Local news sources account for 17.2% and other national and international news, syndicated media, or general news account for the rest.
Medical and health organizations are critical. One in four trusted sources listed (24.9%) are from the health and medical community.
Government is also an important trusted source. 10.6% of the trusted sources listed are government with several mentions specifically of the White House COVID-19 Task Force.
Other sources. A notable percentage of participants are trusting other ways of gathering information: 9.8% of the trusted sources are from search, data/news aggregators, or social media, and most of the remaining 8.5% of the listed sources are friends, family, or self.
Trust profiles emerge. An analysis of the aggregated set (profile) of trusted sources listed by each participant suggests that some trust profiles deserve a closer look in future articles.
For example, we can characterize people by the sources they follow to create trust profiles. Table 1 showcases a few emerging trust profiles. It shows that many people (31%) follow the CDC as a trusted source. A notable percentage of the participants consider key news venues as trusted, including CNN (23.5%), Fox (12.1%) and, the New York Times (10.3%). Others follow their governor, President Trump, or a data aggregator site such as the Worldometer. Note that people may follow more than one source as trusted, so they could be a CDC and a local news follower.
Taking a profile view of the news sources and domains demonstrates that many people supplement their most trusted source with other types of information. See Table 2. Where 86% of participants listed news, medicine, or government as their most trusted domains for information, only about 30% listed only those sources. The remaining two-thirds trust other domains (search, family and friends, etc.) for information. In particular, it is striking to see that a sizable percentage of people (18.9%) list only news sources as trusted. This highlights the significant impact news can have on shaping peoples’ knowledge, behaviors, and worries, particularly if it is a single source such as CNN or Fox.
Diversity of trusted information
As shown earlier in Figure 6, most people (85%) list two or more trusted sources of information. This is considered a good practice because no source is perfect all the time, especially when in the midst of a health crisis where information is changing daily and sometimes even hourly. Nevertheless, more than one in ten people (12.2%) list only one trusted source. Of those who listed two or more trusted sources, some people only listed sources from the same domain, such as multiple news sources. This could give consumers a false sense of diversity, while in fact, their news diet is not balanced.
Fox and CNN are frequently listed among the trusted news sources (12.1% and 23.5% respectively). Yet, relatively few participants (6.3%) list both (Fox and CNN) as trusted sources. This small group also uniquely lists the largest average number of trusted sources (4.4 sources out of 5). For comparison, those who claimed CNN (but not Fox) as a trusted source, list an average of 4.0 sources, and those who cite Fox (but not CNN) list an average of 3.2 sources. The latter is on par with those who identify neither CNN or Fox and list an average of 3.2 sources.
Finally, about one in five users (19.1%) report a person--such as their governor (9.3%), President Trump (7.7%), or Dr. Fauci (6.5%)-- as one of their trusted sources.
People want high-quality information
A concluding question in the survey asked people to offer their thoughts on how they would improve the sources of COVID-19 information online: How would you improve the COVID-19 websites/apps and news available to you online? Participants could provide up to three recommendations.
About a quarter of all participants offer no suggestion. Of those who do, the most popular recommendation is for more facts and accurate information that is impartial, unbiased, and clearly sourced from science, medicine and health experts.
"Only proper researchers and doctors knowledgeable should be allowed to speak to the subject, not politicians."
All other suggestions for improving sources of COVID-19 information are comparatively less common, but interesting nonetheless. For instance, the next most popular suggestion is for information to be easier to access and more up-to-date. Following that are recommendations for more data, graphs, statistics and maps. The popularity of sites such as worldometer.com and Bing surface during the task research sessions, which augment the online survey.
"Make it more accessible, I still need to pay for some subscriptions in order to view briefings or other videos. It should be free."
Finally, some participants indicate a desire for more positive news, less politics, and more information on prevention, preparedness and safety.
"Show more prominently the cases recovered and just give more hopeful news especially about drugs that are working and antibiotics news."
Key takeaways and next steps
When it comes to news and information about COVID-19, people are interested in the latest statistics and trends about the disease, how they can protect themselves, and prevent infection. More people trust the CDC than any other source. But in general, most people trust national news media to provide the information they need.
The majority of people list either the news media, the medical community or the government as their trusted source and many cite a combination of these.
People tend to list multiple trusted sources, which is good, but often they are of the same type, which is not as good. The lack of diversity is particularly evident in certain groups of news consumers.
One thing all seem to agree on is that COVID-19 news needs to be focused on facts, accuracy and science and leave opinion, politics and bias out of the news.
We will dive deeper on these and other findings in future articles.
Really interesting! Would love to see demographics broken down further by socioeconomic status/level of education and mapping responses by region. Great work!
This is great research and well-done. A few things surprised me, one was that many people want more fact-based and scientific research-based information. It leads me to believe that more people are critical thinkers than I thought before. Thanks and can't wait for your next report!
I totally love this. Keep up the good work. Want to learn more.
Very interesting results. The single source behavior is surprising to me. I am curious to learn more about the profiles and their unique behavior as you progress.