Discussions over misinformation, disinformation, and ‘fake news’ have reignited interest in news literacy. A wide range of different actors — from educators to technology companies — believe that raising news literacy would make people better able to separate fact from fiction, potentially limiting the spread of false information and leaving them better equipped to navigate partisan media environments. Others, however, have struck a note of caution by arguing that we need to think carefully about what news literacy should look like.¹ In the past, news literacy largely meant teaching people to be sceptical, or giving them ways of questioning the stories told by the mass media. How useful are such skills in a world where many believe that trust in institutions, including the news media, is already dangerously low?

In this section we will describe how we measured individual news literacy, before taking a more detailed look at its relationship with different news diets. For us, news literacy refers to knowledge about how the news is made: who makes it, how it is selected, and how it is financed. To do this, we asked respondents in certain countries a series of multiple-choice questions with correct answers. By combining these data with that from other parts of the survey, we can see differences in attitudes and behaviours between those with high and low levels of news literacy. This in turn provides an indication of the possible effect that increasing news literacy might have.

Measuring News Literacy

To establish a proxy measure of news literacy, we asked respondents three factual questions. Each probed a different dimension of how the news is made. The questions were multiple choice with a single correct answer. Each respondent’s level of news literacy was determined by the number of correct answers they were able to provide. Of course, three questions cannot accurately measure exactly how knowledgeable a person is about an issue as complex and multifaceted as news production. However, they can be used to establish a reliable proxy, and there is a long history of the use of factual questions in survey research to establish knowledge levels among respondents.

The Questions

Our first question asked respondents to identify which news outlet from a list of four is not primarily funded by money from advertising.² This question essentially asks respondents to identify the public broadcaster from a list of television and print outlets that are funded by advertisers. The available options were adjusted to make them country-specific, but always comprised of a commercial broadcaster and two popular newspapers. Here, as with all of the other questions, we included a ‘Don’t know’ option which was treated as incorrect.

Across the 18 countries we have included in the analysis here, just over half (52%) were able to correctly identify the public broadcaster.³ This varies nationally, with higher figures in countries like the UK (73%), where the public broadcaster is by far the most widely used news source. Only 46% of respondents in the US correctly identified PBS. Across all countries, 15% of respondents incorrectly thought that various commercial TV news outlets or newspapers were not primarily funded by advertising. Around one-third (34%) said they did not know.

Our second question asked respondents who they thought was typically responsible for writing a press release. Just 31% of respondents across all countries were able to do this. This figure rises to nearly half in Sweden (45%) and Denmark (47%). Around a quarter across all countries incorrectly thought that they are written by journalists working for news organisations.

Our third question asked about how news is selected on social media, which as we have documented for several years has emerged as an important source of news for many people. Just under a third (29%) correctly stated that most of the individual decisions about news people see on Facebook are made by computer analysis of what stories might interest them. More than one in ten (12%) said that these decisions were made by journalists working for news organisations, with a similar number (11%) believing that Facebook employs journalists for this task. Just under one in ten (9%) thought the selection process was random.

We can convert these responses into a news literacy scale. When we do this, we can see that news literacy is much lower than many within the news industry might like or expect. We can see that one-third (32%) did not get any of these questions correct. A similar number got just one correct — normally the first question on public broadcasters. Just 10% answered all three correctly. We’ve attached labels to each of these groups ranging from ‘very low’ to ‘very high’ to indicate how we think this translates into news literacy, and will use them throughout the rest of this section.

People with Higher News Literacy Prefer Newspaper Brands

In the remainder of this section, we will focus on how different levels of news literacy are related to different news diets. We start by looking at people’s main source of news. In general, the preference for newspapers and newspaper websites (which we have grouped together here) is more widespread among those with higher levels of news literacy; rising from 20% to 34%. Conversely, the preference for television and television/radio websites is more widespread among those with low levels. The preference for social media as a news source is largely consistent across all groups, but is slightly higher among those with the lowest levels of news literacy (15% compared to 10%).

People with High Literacy Use Social Media Differently

Those with higher levels of news literacy may rely less on social media for news, but they appear to be more discerning when they do use it. When deciding whether to click through to a story, they are more likely to pay attention to a range of different credibility cues. Compared to those with lower levels of news literacy, they are more likely to say that the news brand, the headline, and the person who shared the story are important in deciding whether it is worth their time. The exception to this rule is the number of comments, likes, or shares, which is the least important cue across all groups, but is more important among those with the lowest level of news literacy. However, they are also less likely to share or comment on news themselves, so the simple idea that low-quality news is primarily spread by people with low news literacy may only be partly true.

News Literacy and News Brands

News literacy is also strongly associated with different news diets. Unsurprisingly, people with higher levels of news literacy tend to consume news from a wider range of sources. Thinking about online use only, people with the highest levels of literacy use on average roughly twice as many news brands each week as those with the lowest levels (4.22 compared to 2.39 across all markets).

The specific brands used by different groups also varies a lot in some countries. In US (and also Germany), lists of the most popular brands become inverted as news literacy rises. Yahoo! and Fox are the most popular news brands among those with the lowest levels of news literacy, but among those with the highest levels they are overtaken by brands like the New York Times and the Washington Post. This is highly likely to be due to differences in content. Importantly, Yahoo! and Fox are used by roughly similar numbers of people within each news literacy group, but the reason they fall behind is because certain other brands become much more popular as literacy rises.

We also see elements of this pattern in the UK. As in the US, people with higher levels of news literacy are more discerning, and have a (collective) sense of what brands are worth paying more attention to. Use of the Guardian, for example, rises sharply as literacy increases. But perhaps the key difference is that the same brand (BBC News online) is the most widely used across all levels of literacy.

Differences between brands appear to be less important in the Nordic countries. Here, people’s news diets tend to be similar across different levels of news literacy, with most brands becoming consistently more popular as literacy rises, and the rank order remaining largely the same. This is likely to be due to similarities in terms of tone and coverage across most brands. The exception to this rule is commercial television news, such as MTV in Finland and TV2 in Norway. They do not become more popular as literacy rises.

Complex Relationship between Trust and News Literacy

Finally, we consider the relationship between trust and news literacy. Many people hope that increasing overall levels of news literacy will reverse the decline in news trust we see in many countries. This sounds like a reasonable assumption, but as we suggested at the beginning of this section, news literacy may also go hand in hand with a high degree of scepticism. Even if we focus on news production, the more people know about how the news is made, the more knowledgeable they will be about its limitations and imperfections. This may be why we see only a very small increase in trust levels as news literacy increases.

We also see that trust in news from search engines and news from social media becomes less widespread as news literacy increases. One possible reason for this may be that those with high news literacy are better able to use credibility cues to identify untrustworthy news on search and social. But it may also be because much of the discussion about the impact of algorithmically driven platforms has so far focused on the risks, with terms like ‘echo chamber’ and ‘filter bubble’ starting to enter the vernacular.

These discussions are important, but we should not lose sight of some of the potential benefits highlighted by academic research. The use of social media for news has been associated with more diverse news diets, increases in political participation, and modest depolarisation of political attitudes.⁴ As search engines and social media become more important to the news ecosystem, any attempt to raise news literacy should also aim to improve the knowledge of both the positive and negative outcomes.