Introduction

The recent 2019 UK General Election was a divisive and bad-tempered affair held against the background of Brexit stalemate and rows over the impact of austerity on public services. In the end Boris Johnson’s Conservative Party gained a majority of 80 seats as voters backed his promise to ‘get Brexit done’ and take the country out of the European Union by 31 January 2020 and voters also seemed to have reservations about Jeremy Corbyn, the opposition Labour Party leader.

The election was called at the end of October 2019 after parliament agreed that a fresh mandate was required to break the Brexit impasse. Campaigning officially began on 6 November and spanned six weeks. Polling day was Thursday 12 December.

1. News Consumption across Online and Offline Media

Some commentators have questioned the importance of traditional news formats such as TV and print in comparison with online and especially social media. In this first chapter, we explore the relative importance of these different media but also the way they were used in combination by different groups over the course of the campaign. We base this section largely on the survey data which allows us to compare usage of different sources of news and combine this with demographic and political information. We should bear in mind that as this is an online poll (covering ages 18–65 only) it is likely to underestimate the role of traditional media such as TV, print, and radio for those who are not online.

Later we’ll turn to the tracking data to understand online behaviour in detail, but first we need to understand more about how online fits into wider news diets.

Media Consumption Compared

We look first at the proportion of people who used TV, print, radio, online news, or social media for news overall in the last week of the campaign and compare this with consumption of election news (Fig. 1). TV and online news (including social media) were the most important sources of news overall, with 57% of our respondents using TV and 70% accessing online news. Within online, news websites reached more than half of our respondents (56%) but social media was also significant with four in ten (41%) having used it to get the news in the previous week. Only around a third used radio to get news and roughly a quarter of our online sample read printed newspapers. This consumption pattern matches findings that we have published in our previous Digital News Report based on data from the beginning of 2019.

As one might expect, consumption is slightly lower across the board when we ask just about election news. While reliance on online news drops by 14pp, television largely holds its share indicating that the medium may have played a bigger role around the election, at least for some. TV news spends a great deal of time covering election issues and that can reduce the space available for other news. By contrast, there tends to be a greater variety of news stories available online and in social media, which means more distractions – and more opportunities to avoid election news in particular.

Figure 1. Proportion that used different sources of news/election news in the last week of the campaignNEWSSOURCE_POST. Which, if any, of the following have you used in the last week as a source of news? NEWSSOURCEELECTION. Which of these have you used to get news about the 2019 UK General Election in the last week? Base: Total survey sample = 752.

When it comes to election news specifically, we also see differences around age and political allegiance – which we know are linked in the UK (Fig. 2). Younger users were much more likely to use online news and social media while older ones were more likely to rely on offline media – particularly TV, but also radio and print. There is a similar split when it comes to political allegiance, with Labour voters more likely to use online news and social media when compared with their respective counterparts. This is likely to be strongly linked to age because Labour voters tend to be younger on average.

Figure 2. Proportion that used different sources for election news in the last week of the campaign

NEWSSOURCEELECTION. Which of these have you used to get news about the 2019 UK General Election in the last week? P2019VOTEACTUAL. Which party did you vote for? Base: All survey respondents: 18-34 = 150, 35-65 = 602. All survey respondents that voted: Conservative = 213, Labour = 217.

These age-related differences between online and traditional media are striking but so too is the role played by social media within the online category. Four in ten (36%) of 18–34s say they used social media to access election news, only slightly fewer than the percentage that accessed online news sites (49%). One in seven (13%) of the younger group say they only consumed news from social media in the last week of the campaign. 21% of 35–65s say they only consumed TV news.

We’ll come back to perceptions of the importance of different sources of election news later in this report.

2. Online News Consumption in Detail

In this chapter we dig much more deeply into the detail of online news consumption using our tracking data with 1,711 participants aged 18–65, including all of those who took the survey. With the agreement of panellists, we monitored all the news websites and apps they visited on computers and mobile phones as well as their wider web usage. This methodology should provide more accurate data about what these respondents did rather than just what they remember when asked in a survey.

Taking the whole six weeks from 4 November to 15 December, we find that around three-quarters (72%) of the entire online panel visited one or more news websites during the election campaign, with the top 30 sites accounting for the vast majority of that. The average news user spent 16 minutes a week reading online news stories during the six-week campaign and accessed around 22 news stories each week across web and mobile. This is a significant amount of news consumption, but to put it in perspective, it is worth pointing out news makes up just 3% of all internet time – even during an election period.

Tracking data. Base: All that accessed a website between Nov 4 and Dec 15 2019 = 1,666.

Figure 3. Reach of different types of news sites during the campaignTracking data. Base: All that accessed a website between 4 Nov and 15 Dec 2019 = 1,666.Note. Numbers in parentheses show the number of sites monitored in each category.

National and international news sites reached almost three-quarters (71%) of all website users, with hundreds of local sites together reaching four in ten (40%) – but with very low figures for each individual website (Fig. 3). Partisan digital sites such as the Canary and Novara Media and foreign sites like Russia Today (RT) collectively reached 6% of users. RT’s reach was roughly equivalent to that of the Brighton Argus, a popular local title, at around 1% of our sample. We also monitored traffic to over one hundred pro- and anti-Brexit pages on Facebook, but these were visited by less than 2% in total. Satire sites were used by around 3% of our sample.

Daily online usage fluctuated during the course of the campaign with significant peaks at the start of the campaign (6 November) and again at the end for the results (12 and 13 December) (Fig. 4). It is striking that overall online news consumption during the election was no higher than the month before – a period which saw a deal with the EU and a series of political manoeuvrings and high-profile parliamentary votes. On 29 October Boris Johnson finally got his wish for a pre-Christmas vote and the campaign started the following week.

Peaks of interest during the campaign itself were mainly focused around non-election stories such as flooding in the North of England, a terror attack in London, and a BBC interview with Prince Andrew where he tried to ‘clear the air’ over his friendship with Jeffery Epstein. Some of these became political issues which demanded a response and diverted the parties from their core messages for several days.

Figure 4. Average time spent with news sites per dayTracking data. Base: All that accessed a website between 4 Oct and 15 Dec 2019 = 1,806.

Figure 5. Other stories knocked the election off top spotFloods in Sheffield in the first week, Prince Andrew’s interview with the BBC in the second, and the London Bridge terror attack in the fourth all diverted attention from the election and demanded political responses. Far left: REUTERS/Jon Super; Centre: Mark Harrison/BBC; Right: REUTERS/Henry Nicholls

Which News Websites were Most Heavily Used during the Election Period?

The BBC was the most heavily used news website in terms of people reached (44%), the number of stories read, and the time people spent with those stories (Fig. 6). Right-leaning Conservative supporting titles, the MailOnline, the Sun, Telegraph, and Express, also attracted considerable online reach, as did the Guardian and the Mirror, which tended to be more sympathetic towards Labour.

Figure 6. Reach of specific news websites during campaignTracking data. Base: All that accessed a website between 4 Nov and 15 Dec 2019 = 1,666.

It should be noted that, for the BBC and Sky News, we only measured traffic to their news domain. For other websites, such as the MailOnline and the Guardian, we tracked all content including sport and a wider range of lifestyle content. It is likely that, when it comes to political and election coverage, there is an even bigger gap than is indicated here between the BBC and the rest.

Our approach, which is based on measuring access to hundreds of thousands of story URLs under a publisher domain, does not allow us to identify political news or election coverage specifically.

We should also be wary of just looking at website reach. We get a very different picture if we look at the time people spend with different websites. This is arguably a better measure of engagement and shows the importance of a small handful of sites in the UK. Taken together, the BBC (28%) and the MailOnline (21%) account for almost half of all time with news sites while the Guardian (7%), the Mirror (6%), and the Sun (5%) make a significant contribution (Fig. 7). Other sites that do well in terms of reach, however, such as the Express and the Independent, show very low levels of time spent, with around 2% of the overall news pie. In many cases, this is because traffic to these sites tends to be irregular use from social media or search rather than regular and loyal audiences.

Figure 7. Time spent with news sites during the campaign

Using the time-spent metric, we can also identify the amount of time people spent online with brands that supported the Conservatives (right-leaning) or Labour (left-leaning).

In the UK, most national newspapers openly endorse either Labour or the Conservatives during election campaigns, and actively encourage their readers to vote for them. In 2019, the Mirror and the Guardian endorsed Labour, whereas the Daily Mail, the Daily Express, The Times, the Daily Telegraph, and the Sun backed the Conservatives (the i, the Independent, and the FT did not make an endorsement in 2019). However, strict laws mean that broadcasters like Sky, Channel 4, ITV, and the BBC must remain politically neutral, and devote equal airtime to parties and candidates.

Figure 8. Proportion of news site time spent with different types of outlet

Here we see that much of the time spent with online news media during the election was with brands that had obligations to be impartial (broadcast brands like BBC and ITV) or those that took a non-partisan view (like many local newspapers) (Fig. 8). Brands that openly supported the Conservatives (e.g. Mail, Telegraph, Times, Sun) accounted for nearly three times as much time as those that supported Labour (e.g. Guardian, Mirror).1

In combination with our survey evidence about the importance of television as a source of election news, it is clear that sources with obligations around balance and due impartiality remain critical for framing political news offline and online in the UK. This is a very different picture from the one we see in the United States where some commercial broadcast media (and their websites) take a partisan approach to much of their coverage and public media are less influential.

Having said that, obligations around impartiality are not the same as being seen to be fair and balanced. There was significant criticism during and after the campaign on social media and from participants and candidates about hidden and overt biases.2 Senior BBC and ITV correspondents were regularly accused of bias from both Conservative and Labour supporters. We explore these issues further in Chapter 4 where we look in more detail at the issue of trust.

Where did Younger and Older Groups Get their Online News during the Election?

We’ve already seen in Chapter 1 that young people were more likely to get their news online than older groups so it is something of a surprise to find in our tracking data that the young were accessing fewer online mainstream news sites overall. Two-thirds (66%) of 18–35s accessed any news site during the six-week campaign compared with three-quarters (76%) for 35–65s (Fig. 9). This suggests that young people are turning to other sources of online news.

Young people used mainstream brands (BBC, MailOnline, the Guardian, Sun, and Mirror) during the campaign, but they used them proportionally less. The main exception was the Guardian which had a slightly higher proportion of under-35s than over-35s and the same was true of digital-born brand BuzzFeed, though with much lower reach overall.

Figure 9. Reach of selected news sites by ageTracking data. Base: 18-34s/35s-65s that accessed a website between 4 Nov and 15 Dec 2019 = 673/1051.

Even more striking was the difference in engagement amongst under-35s, who spent less than half as much time reading news stories (8 minutes/week) when compared with over-35s (22 minutes/week) (Fig. 9). We know that young people are less interested in politics and less likely to vote, but given they also consume less TV and radio news this is worrying from a democratic perspective.

It is possible that social media is making up some of this gap, but other research suggests that this tends to involve catching up on headlines, watching short videos, and sharing memes rather than engaging with deep or nuanced coverage.3 In our survey, under-35s say they used news websites for election news more often than social media and far more than TV or print. They also said that news websites were the most important in helping them understand the issues in the election. And yet 8 minutes a week (less than 2 minutes a day), on average, spent reading all news websites suggests that this group has been paying little attention to their own medium of choice.

The Role of News Apps

Our tracking methodology also allows us to look at the time spent with different news apps during the election and to compare this with other apps that people use. Here we find that, among app users, BBC News (10%) and Sky News (3%) had by far the highest reach of any news app during the campaign but there was very little traction for others (Fig. 10). Overall, 18% accessed one or more news apps during the campaign compared to 66% that used Facebook, 64% WhatsApp, and 27% Twitter. People who used apps spent more time on average than those who used the mobile or desktop websites but in total time with news apps was less than 1% of all app usage.

Figure 10. Reach and time spent with apps during the campaignTracking data. Base: All that accessed the internet using a mobile app between 4 Nov and 15 Dec 2019 = 758.

In total the BBC app accounted for more than half of all news app users. News app users tend to be brand loyalists, older, and more interested in news. Our tracking data show reach of the BBC News app is 6% with 18–24s but as high as 16% with 55-65s.

Despite these significant numbers, far more people used news websites via mobile or desktop browsers. The vast majority are not sufficiently interested in news to download a specific news app. When it comes to apps, it also worth noting how news compares in terms of both reach and time spent with social apps like YouTube, Facebook, WhatsApp, and Twitter. According to our data, under-35s who used Facebook spent on average nearly 3 hours a week using the app during the election period, Instagram users spent just over an hour each week, and Twitter users spent just under an hour. This compares with an average of 30 minutes for this group reading all news apps put together.

When they are using these apps they will, of course, come across news, but according to our survey data, with the exception of Twitter, they are primarily there to connect with friends rather than catch up with news (Fig. 11). It is notable how under-35s consider all of these apps more important for news than older groups – especially Twitter (33%).

Figure 11. Relative importance of news for users of different social platforms

Top Stories during the Election

As previously discussed, our methodology measured news usage during the election period from hundreds of designated websites, but this is not the same as measuring election news. To understand the relative interest in the election compared with other stories, we looked at the most viewed 100 stories for each week with our panellists and measured the proportion that were election related.

On average, we find that around four in ten (38%) of the most popular online stories were about the election during the six-week campaign, though this proportion was as high as 61% in the final week (Fig. 13). As previously noted, the period also saw major floods in the North of England, Prince Andrew’s ill-fated interview about his friendship with Jeffrey Epstein, and a terror attack on London Bridge in which two people died. The bulk of the most-read stories – including almost all the top stories each week – came from BBC News, which as we’ve already seen is by far the most used news website in the UK. The most-read story in the period from the start of the campaign to a few days after polling was the BBC’s ‘As it happened’ election results page, which was read by around 4% of the sample. If we also include the month before the campaign, the most-read story was the BBC’s poll-tracker, which was read by around 5% (Fig. 12).

Another successful piece of evergreen content was the BBC’s ‘Who Should I Vote For?’, which enabled users to compare the manifesto commitments of the main parties. This interactive guide was used in similar proportions by younger and older groups and reached about 2% of our sample.

Figure 12. Word cloud drawn from the top headlines and two of the most popular news stories

On a weekly basis, the most read election stories tended to involve news rather than opinion, with ‘live pages’ run by websites like the BBC and the Guardian accounting for a significant proportion of traffic. These contain a feed of updating posts about developments in the election mixed with social media reactions in reverse chronological order. Our data show that these live pages had dwell-times up to ten times higher than a typical story and many visitors would normally return several times in a day.

Conservative-supporting media such as the MailOnline chose a political agenda that focused on issues such as anti-Semitism in the Labour Party and the level of spending commitments in the party’s manifesto. News headlines from right-wing media were often overtly partisan, framed as part of a larger narrative around the dangers of a hard-left Corbyn government:

In addition, both left- and right-leaning media from a print background carried opinion pieces with a clear political perspective:

• Dominic Lawson: If you think the NHS is well run, vote Labour | Comment | The Sunday Times
• Marina Hyde: In this climate, how does Boris Johnson not melt with shame | Opinion | The Guardian

Partial headlines, personal attacks on politicians, and overt press support for a particular party may have had some effect in this campaign but our data show that most of these stories gained little reading attention. Brands like the Express achieved some reach over the course of the campaign but its stories accounted for 2% of total reading time. Stories from the MailOnline did attract more attention, though story analysis shows that much of this was for non-election-related news. The Guardian was, if anything, much more influential, with its political coverage achieving high reach with younger voters and appearing regularly in the most-read lists.

Election Results Day

Figure 14 which tracks engagement levels for popular news sites, shows how audiences turned to the BBC for reliable information and analysis as the results came in on election night. Traffic to the BBC was around four times higher than its nearest competitors (the Guardian and MailOnline) during these peak times.

Figure 14. Hour by hour consumption of news websites for election resultsTracking data. Base: All that accessed a website between 12 Dec and 13 Dec 2019 = 1,243. Note: figures indicate total minutes spent with each site by panellists. A figure of 0 indicates that no panellists used a site in that hour.

By morning, we see another significant peak with online both a supplementary and primary source of news. The bulk of usage came around 7 am, with mobile phones a convenient way to get updates on the national picture and on local results. The peaks for MailOnline and the Guardian came a few hours later, which suggests a certain amount of complementary usage as users looked for additional perspectives and analysis on the results.

Figure 15. Much of the battle for election traffic was via mobile phones

Election results pages from BBC News, The Times, Guardian

Conclusion

Despite the importance to the Brexit outcome, it is clear that the British people were not fully engaged by an election just a few weeks before Christmas. There was a peak of interest at the beginning of the campaign and another one when the result became clear, but overall the amount of news consumed online was pretty similar to a non-election period and interest fell over the course of the campaign. Individual news organisations with highly engaged and politically interested users may have seen more activity, but we find no evidence of an overall increase.

Frustration with the drawn-out Brexit process may have played a part in this – along with lack of trust in politicians. With many people saying they worry that they can’t tell the difference between what is real and fake online it is perhaps not surprising that many turn as much to their friends or colleagues to help make up their mind as they do to the media these days.

But even if the news media play a smaller role than in the past, our report does show that the vast majority of our sample consistently consumed the reporting of a relatively small number of mainstream news brands operating across a combination of TV, radio, print, and online. Television remains an important source of news overall as well as helping people understand the issues.

But TV is unlikely to maintain this level of importance in the years ahead as the population ages and fewer people watch traditional TV news. If these trends continue it is not clear which media will fill the gap or how fair and balanced that will be. News media did reach over 70% across online sources but much of this was shallow and fragmented – accounting in total for just a fraction of all internet time.

The evidence that younger voters spent just 8 minutes per week with online news media (and even less with election news) and that more than one in ten (13%) of this group only used social media for election news will be of considerable concern for those who worry about political engagement and the health of democracy. This is not so much a concern about political filter bubbles, since our data suggest social media users tend to be exposed to a wider range of political news, but more a worry about why younger people seem to find existing news media so unappealing , and about the unreliability and the lack of transparency over what people are exposed to.

In terms of news brands, the presence of the BBC is both unsettling and reassuring. Its political coverage dominates online news coverage as much as it does via TV and radio, suggesting a less plural media than some might like. On the other hand, it could be seen as a positive that most of the news people were exposed to came from an organisation that has particular obligations to be fair and balanced during an election, conducted regular fact-checking of politicians, and published a wide range of information about party positions. Given the importance of this output, it is scarcely surprising that the BBC’s coverage has come under such scrutiny from both sides and was the subject of more complaints than ever before.

Despite this, our data suggest these criticisms from political parties, activists, and media commentators were not fully shared by ordinary people. The majority of those who expressed a view felt the BBC had done a good job. BBC stories were widely viewed, and read equally by Conservative and Labour voters. More widely, however, it is true that Labour supporters feel more suspicious of the media than Conservatives. There is historic resentment on the left about the relentless personal attacks by right-wing titles which they believe make it harder for a Labour leader to ever win an election. But it is also clear from our data that the right-wing media – with the exception of the Daily Mail – have lost influence and reach with the move to online and continue to do so with each passing election. Indeed, there are more left-leaning options online and some of these, like the Guardian, have gained much greater reach and influence than they ever had in print.

It was probably never the case that headlines in a national newspaper endorsing one party or another could win or lose an election. That is even less the case today as politicians develop their own channels and as personal networks have become more important in influencing choices. This report shows that mainstream media still play an important role in scrutinising politicians and helping to inform voters, but they no longer have a monopoly on these activities.

Before the election, much of the public debate in media circles related to concerns about political polarisation fuelled by online and social media, but our analysis suggests an even bigger problem may be the lack of significant engagement with news by significant portions of the population.

News media looking to increase their relevance may wish to draw the lessons from this campaign to convince those disappointed with the coverage that they can do a better job in the months and years to come.

Methodology

This study was commissioned by the Reuters Institute for the Study of Journalism to understand how news was consumed during the 2019 UK General Election campaign. The data were collected by Netquest using a combination of passive tracking and two online questionnaires.

Netquest tracked the computer and mobile use of 1,711 people aged 18 to 65 from 4 November to 15 December 2019. A subset of these panellists was also invited to complete an online survey shortly after the official start of the election campaign (fieldwork 13–19 November), and again immediately after polling day (fieldwork 13–19 December): 840 panellists completed the first survey, and 752 of these also completed the second survey (retention rate: 90%).

For most of the statistics we present here, the base is everyone in the panel that accessed a website at least once during the campaign from either a computer or mobile device (1,666 people). Where we refer to the use of apps, the base is everyone that accessed the internet using an app during the campaign (758).

The tracking data were weighted to be representative of the British population aged 18 to 65 in terms of age, gender, and region, based on the latest population estimates from the Office for National Statistics.6 The survey data were weighted to be representative of the British population aged 18 to 65 in terms of age and gender. As the focus of the research was on the UK General Election, we excluded panellists from Northern Ireland due to differences in the political system.

The focus of this research is on the passive tracking of online media use. Limitations associated with panel recruitment and the practicalities of the tracking methodology meant that we were not able to recruit a sufficient number of people aged over 65 to make it representative of the entire British population. Therefore, we restricted the study to those aged 18 to 65. This is important to keep in mind here because age is an important predictor of both vote choice and news use. As such, the data will tend to under-represent Conservative voters. This means that our survey sample contained slightly more Labour voters than Conservative voters, but this is consistent with post-election polling from YouGov that also showed that Labour had a higher vote share than the Conservatives among the under-65s.7

The focus of this study is online news use. We tracked the use of 35 of the most popular mainstream national and international news websites for broadcast, newspaper, and digital-born outlets (e.g. bbc.co.uk/news, dailymail.co.uk, huffpost.co.uk), 879 local newspaper websites (e.g. birminghammail.co.uk, liverpoolecho.co.uk), 20 specialist, alternative, or partisan websites (e.g. thecanary.co, breitbart.com, newstatesman.co.uk), seven satirical news websites (e.g. thepoke.co.uk, thedailymash.co.uk), and 125 dedicated politics pages on Facebook (e.g. Leave.EU, EvolvePolitics). We also tracked the use of 16 dedicated mobile news apps (e.g. BBC News app, Sky News app). Unless otherwise specified we use ‘news’ to refer to the 35 most popular mainstream national and international news outlets.

Passive tracking data are generally considered to be a more reliable way of measuring online news use than surveys. However, the tracking software is only designed to collect data on URLs accessed using computers running Windows and Mac OSX operating systems. The mobile tracking software can be installed on both Apple and Android smartphones and tablets, but only the top-level domain for each URL accessed (e.g. bbc.co.uk, theguardian.com) can be recorded, as well as which apps are shown on the main screen. As the passive tracking software records the URLs loaded by browsers, it is not possible to track what people ‘see’ on their screens but do not click on. As such, passive tracking does not record, for example, news snippets displayed on people’s social media feeds or on apps that use their own in-built browsers. Panellists are able to turn off the tracking software for 15-minute periods at any time, and this is not recorded in the data.

As tracking software only measures online use, we supplemented it with online survey data on people’s attitudes and their offline news use. Online samples will tend to under-represent the consumption habits of people who are not online – typically those who are older, less affluent, and with lower levels of formal education. It is also important to note that online surveys rely on recall, which is often imperfect or subject to biases. We have tried to mitigate these risks through careful questionnaire design and testing. We have made particular use of survey questions from the British Election Study, and we are very grateful to them for making their survey questions publicly available.8

The size of the survey sample means that some subgroups in the data are too small to meaningfully compare to others. For example, it is possible to compare Conservative voters with Labour voters, but not with voters for other parties. Even comparisons between Conservative and Labour voters should be treated with caution because with roughly 200 voters for each in the sample of 752, only differences of around 10pp will be statistically significant using a chi-squared test (p < .05). Where appropriate, we flag differences between groups that we do not believe to be statistically significant using this test in both the text and the figures.

About the Authors

Richard Fletcher is a Senior Research Fellow at the Reuters Institute for the Study of Journalism and leads the Institute’s research team. His primary research interests are global trends in digital news consumption, comparative media research, the use of social media by journalists and news organisations, and more broadly, the relationship between technology and journalism. He is lead researcher and co-author of the main Digital News Report – the world’s largest annual survey of global news consumption.

Nic Newman is Senior Research Associate at the Reuters Institute for the Study of Journalism, where he is lead author of the annual Digital News Report and the annual Journalism Trends and Predictions report. He has recently published a detailed study on news podcasting. Nic is also a consultant on digital media, working actively with news companies on product, audience, and business strategies for digital transition.

Anne Schulz is a postdoctoral Research Fellow at the Reuters Institute for the Study of Journalism. Her doctoral work focused on populism, media perceptions, and news consumption. She is researching questions surrounding news audiences and digital news.

Published by the Reuters Institute for the Study of Journalism with the support of the Google News Initiative.