Ordinary People and Social Media as Sources in Norwegian Newspapers

Abstract

Ordinary people have always been of interest to journalists, and social media has become a common place to find material for new stories. This paper presents a quantitative content analysis of Norwegian news articles that are based on social media posts published by ordinary people. The analysis focuses on the topics of the news stories, sources, headlines, lead paragraphs, use of amateur photos, and what news criteria they fulfill. The news articles generated covered a wide variety of topics but were mostly soft news. Social issues, culture, and politics were the largest categories. They were episodic in form and replicated much of the content from the original social media post together with an interview with the person who posted it. There was little use of other sources or follow-up stories. Most of the photos used were amateur photos. The journalistic processing of the material is at a minimum. Compared with Norwegian news in general, the sources in the material were slightly more female and of a wider age range, and they were picked up by journalists because of popularity cues on social media. For journalists, this poses an opportunity and a challenge to develop the stories into something more than mere snapshots of society. For the individuals involved it poses an opportunity to reach a larger audience, but also the challenge of context collapse when the audience shifts from their contacts on social media to the general news audience.

1. Introduction

“Facebook post: Truck driver Thor has had it!”

“Shared by millions around the world!”

“365,000 have seen this video!”

“Posted pictures of his 90-year-old father on Facebook–demands nursing home!”

These are some headlines from online Norwegian newspaper articles that are based on social media posts published by ordinary people. They often refer to how many times a post has been shared or a video has been viewed, or more vague concepts, such as “going viral” or “taking the internet by storm”. The newspaper articles do not usually go into detail regarding the topic of the social media post, nor do they present additional research or sources.

Social media has become a significant platform both for newsreaders and journalists. Newspapers often quote updates from ordinary people, as well as from famous politicians, sports stars, movie actors, and other celebrities. The present study presents a content analysis of news articles from Norwegian online newspapers that are based on social media posts made by ordinary people. It investigates which social media are most commonly used in news stories, what topics are covered, how the newspapers present them with pictures and videos, and use Goffman’s concept of performance to discuss the significance it might have for the involved people when their identity performance is moved to a different context where they cannot control it.

Ordinary people have always been of interest to news media but have generally become more visible in the past few decades, both in entertainment media and in the news (Eronen 2015; Palmer 2018; Schmieder 2015; Turner 2010). The use of amateur material and ordinary people as sources in news media has increased, and this includes amateur photos and videos (Karlsson et al. 2015; Mortensen 2016; Mäenpää 2014; Näsi 2015; Pantti and Sirén 2015; Schmieder 2015). The term ordinary people can be an ambiguous concept. In the context of this study, ordinary people are understood as people who are not part of an elite or celebrity group, nor hold any public position where media exposure is expected (Eronen 2015; Turner 2010). Ordinary people become news sources as witnesses, experts, heroes, community representatives, and so on (Palmer 2018). According to the study of news production by sociologist Gaye Tuchman, ordinary people in the news have significance not only because of the information they transmit but as representatives of a group (Tuchman 1978).

In terms of news production, a source is “any person or group to which information was attributed” (Kurpius 2002, p. 858). It includes any actors who provide background information or story suggestions, or who are observed or interviewed by journalists (Gans 2004). Research on the agenda-setting function of news media suggests that the selection of news sources affects how news is constructed, so studying the news sources is important (McCombs 2005; Messner and Distaso 2008). However, academics have rarely studied ordinary people as news sources (Palmer 2018).

Social media is a cheap and accessible source of information and have impacted the production and distribution of journalistic content (Ferreira 2021). Journalists use social media to monitor other news services, look for news events and find ideas for stories, and for audience contact, among other things (Weaver and Willnat 2016; Zhang and Li 2020). Journalists have their professional source cycles on social media and follow various elite profiles. It can be difficult for people with no direct relationship with the media and no public position to break into this source cycle (Brandtzaeg et al. 2016; Zhang and Li 2020). The present study focuses on the occasions when this does happen and investigates the type of news article that draws on ordinary people’s social media activity. Most of the cases in the material are picked up by journalists based on popularity cues on social media, providing a clear indicator that people find these stories interesting. The stories covered a wide variety of topics, often topics that rarely make news headlines. Despite this, the stories were rarely expanded with additional research or follow-up stories. There were merely presented as a snapshot, a small moment in time where an ordinary citizen for a short time meets a wider news audience than the contact network on social media.

2. Ordinary People in the News

While ordinary people always had some representation in news media, journalists are more likely to choose sources they regard as credible, knowledgeable, and powerful (Messner and Distaso 2008). Earlier studies have shown that sources have been dominated by people with a high level of education and income, and who are in leadership positions (Allern 1996; Olsson 1984; Sjøvaag and Kvalheim 2019; Vaage 1985). This is partly because they are involved in decision-making processes, so they are active producers of news material such as action plans, policies, and reports; hence, they are in positions that warrant a watchful eye from journalists. There are also pragmatic reasons; these sources are easy to reach, making it a less costly process for journalists (Allern 2015).

The use of ordinary people as sources has increased in recent years (Näsi 2015; Pantti and Sirén 2015; Schmieder 2015). Stories about ordinary people have always been of interest to news reporters, for instance, as a way of illustrating or giving a face to a current trend or major event, or as a method of explaining something complex through the use of personal examples (Eronen 2015; Schmieder 2015; Turner 2010). Social media have also become an important source in stories of a more mundane character (Farhi 2009; Hermida et al. 2012; Paulussen and Harder 2014; Singer et al. 2011). Studies on sources in Norwegian news have revealed that 10% of the sources are ordinary people with no connection to an elite group, and they are used mostly for stories on dramatic events, accidents, crime, or the closing down of a business, in such cases, they are often cast in the role of victim or an illustrative example (Allern 2015).

Palmer (2018) conducted in-depth interviews in the USA with ordinary people who by chance ended up in the media spotlight. Some described it as fun, others said it was like death. For some, it was just a fun thing to do, but for others the experience was existential, and they started to question their own identity. Most subjects enjoyed the experience with a mix of altruistic and personal motives. The altruistic motives could be linked to bad experiences the informants have had about something and wanted to help others to avoid the same fate. Others hoped to educate the public or mobilize public action (ibid).

While most of the informants in Palmer’s study enjoyed the attention of a larger audience, they also resented how the journalists controlled the story and the access to this audience (ibid). Losing control over their own story was also an issue in a Norwegian interview study of people who had been through intense periods of public criticism in the news (Duckert and Karlsen 2017). The subjects were media commentators, editors, and athletes, and were all used to well-controlled media exposure, but when they faced heavy criticism, they lost control over their own stories. All the subjects in the book reported that this had given them various health problems, such as sleep deprivation, suicidal thoughts, anxiety symptoms, psychosomatic illnesses, and fatigue (ibid). One may wonder what it is like for a person who does not have media experience or support from media advisors.

On social media, people can write their own stories. Journalists can grant them access to a larger audience and will then control the stories by filtering the material from the sources, and selecting what to publish. This process changes the audience. Goffman (1956) uses the term “performance” to refer to how a person in a given context will choose which part of his or her identity to display to others. People are performing a facet of their identity, according to the impression they wish to convey to others. Which facets of the identity are being performed, and how it is performed, will change depending on who is present. If someone leaves or enters the room, the situation changes so the performance will also change. Knowing and understanding the audience is crucial for a successful performance.

Goffman’s work concerned physical spaces where the performance is crafted for a specific audience, but scholars from various disciplines argue that the concept of identity performance is also relevant to an online context (Buckingham 2008; Turkle 1996). Being active on social media is, among other things, to choose what part of yourself to present to an intended audience of friends, relatives, and acquaintances, but the borders between the intended audience and the actual audience are blurry, so there is a risk of a damaging mismatch between the intended and actual audience (Hodkinson 2016; Marwick and Boyd 2014). When social media goes viral, with the news coverage to follow, it disturbs this process even more, and the individual loses control over his or her self-presentation, as this is filtered by a journalist.

2.1. Social Media as a Source in News Production

Several studies from different countries show that social media are well established as a resource for journalists, who use them to gather and filter news, interact with the public, investigate stories, find sources and quotes for news stories, disseminate news, and promote specific events (Broersma and Graham 2012; Hermida et al. 2012; Ju et al. 2013; Kalsnes 2016; Lee and Ma 2012; Paulussen and Harder 2014; Rogstad 2014; Singer et al. 2011; Skogerbø and Moe 2015). Social media are a source both for hard news, such as national politics and international conflicts, and for soft news, such as show business, culture and arts, sports, and items of human interest (Paulussen and Harder 2014). Studies from Norway (Grøtte 2015) and Belgium (Paulussen and Harder 2014) have demonstrated that social media are frequently used as a source for celebrity news. Journalists also use social media to find user-generated content. User-generated content is most prevalent in popular culture-oriented stories and personal everyday life-oriented stories, and less so in news stories and informational stories (Jönsson and Örnebring 2011). For online Norwegian newspapers overall, 20% of the news articles are in the politics category, while 30% deal with sports, crime, and accidents (Elgesem 2015). Is this different when news media use ordinary people on social media as sources? RQ1 focuses on this:

RQ1: What topics are covered in the news stories that have ordinary people on social media as their source?

The typical source in Norwegian news media is a man between 20 and 66 years old who usually appears in a professional capacity (Dahlstrøm et al. 2017). News sources over the age of 67 are virtually non-existent (ibid.). The percentage of female sources has increased substantially, from 9% in 1979 to 19% in 1999 and 36% in 2015, though it is considerably higher (53%) in news relating to health, social issues, entertainment, and culture (Ibid.). Newspapers published user-generated content in the pre-digital era as well, for instance in the form of letters to the editor. Studies showed that it was the white middle class who wrote most of the letters to the editor (Forsythe 1950) and that the published letters mainly came from people with above-average income and education (Jönsson and Örnebring 2011). Social media have a broad demographic, and so it is tempting to assume that this leads to a broader representation of sources in news media and that it gives them a voice in the public discourse. This leads us to research question 2 regarding the gender and age of the sources.

RQ2: What are the demographic features of those who are used as sources?

Ordinary citizens are more often paraphrased than directly quoted, and referred to in collective and anonymous terms, as representing the public (Paulussen and Harder 2014). Is this also the case in these news articles? Research question 3 asks how ordinary people are referred to when news organizations use them as sources.

RQ3: How are ordinary people referred to when they are used as sources?

2.2. Journalistic Presentation

The emergence of affordable digital cameras and mobile phones led to the use of amateur images in news reporting, for instance during disasters and other dramatic events (Karlsson et al. 2015; Mortensen 2016; Näsi 2015; Pantti and Bakker 2009). But news services also use amateur images of a more mundane character to engage their readers and viewers (Mäenpää 2014; Näsi 2015). Segments such as “pictures of the day”, which feature natural scenery, cute animals, and local community events, serve mainly to create positive feelings (Näsi 2015). A Dutch study identified three categories of amateur pictures in news media; misfortune (hard news and information), memories (soft news, and personal and everyday oriented content), and sunsets (or more generally, weather and natural landscapes; (Pantti and Bakker 2009).

Amateur footage is often of poorer technical quality than professional footage but offers a different kind of authenticity and proximity. Consider, for instance, the immediacy that amateur footage can provide in the case of dramatic events where professional photographers are not present (Karlsson et al. 2015; Mortensen 2016), and how the immediacy of amateur footage can provide additional information or perspectives regarding events that have already been covered (Näsi 2015). Poor technical quality can symbolize the reality of what is being presented, and it can give the images authenticity. This makes them interesting as news material (Andén-Papadopoulos and Pantti 2013; Mortensen 2016; Mäenpää 2014; Pantti and Bakker 2009; Pantti and Sirén 2015).

In the context of more mundane news stories, it may seem that amateur photographs are used primarily because they are, in most cases, cheaper than professional ones. But a picture of an ordinary person in a mundane news story links it to ordinary people. What is more, amateur photos portray intimacy through personal experience and feelings (Mortensen 2016; Pantti and Bakker 2009). Schmieder (2015) conceptualized amateur photos as “visual quotes” because journalists use them in much the same way as they use spoken quotes from interviewees. Images function as elements in the narrative that news workers aim to create. The data collected for the present study naturally do not tell us how amateur images are treated by news desks, but we can analyze the published results themselves with research question 4:

RQ4: What types of photos and videos are used in these news articles?

Studies from the Nordic countries indicate that social media are not replacing traditional sources, but rather are complementing them (Skogerbø and Karlsen 2014; Skogerbø and Moe 2015). Most news stories where social media are used as source material refer to other sources as well (Broersma and Graham 2012; Hladík and Štětka 2015; Paulussen and Harder 2014). Most of the content from social media that appears in traditional news outlets is textual information, except in the tabloid press, where pictures dominate (Hladík and Štětka 2015). The use of social media as a source certainly has a financial side to it. News organizations have had to downsize and reduce staff, and this has hurt journalists’ capacity to work with multiple sources (Dahlstrøm and Hognestad 2016). However, there is also the matter of verifying pictures, videos, and information found on social media, which can be difficult. There is no guarantee that using social media as a news source always will be cheaper, easier, and quicker than traditional news sources. Research question 5 focuses on the use of sources in the news stories, and whether the social media content remains the main source in the news article, or if the story is expanded to include other sources. The purpose is to discuss how social media can contribute to richer news reporting, but also how it can be used to quick copy-paste journalism.

RQ5: Is content from social media the main source of these news articles?

2.3. News Criteria

There are no definitive rules for how journalists decide which events should qualify as a news story, but there are some well-known news criteria. An incident is more likely to become a news story if it is close in time, space, and culture (Bro and Wallberg 2014; Wadbring and Ödmark 2016). Furthermore, negative, unexpected, or odd events are more likely to become news stories than positive and normal incidents, which means that accidents, crime, celebrity behavior, and sports are prioritized. Events linked to a political, economic, cultural, or athletic elite are more likely to become news stories, as well as events that can be simplified since complicated events rarely breakthrough as major news stories (Bro and Wallberg 2014; Wadbring and Ödmark 2016; Aalberg and Elvestad 2012). The chances of becoming a news story also increase if the subject matter is sex or if it can be linked to sex (Aalberg and Elvestad 2012). Traditional news criteria value negative news over positive news, but Wadbring and Ödmark (2016) remark that this seems to be different for online news, which tends to be more positive. They argue that new viral sites, with their focus on uplifting and spreadable stories, have had an impact on traditional news organizations in this respect (Wadbring and Ödmark 2016).

Online formats offer constantly updated information on likes, shares, and comments on social media (Al-Rawi 2017; Larsson and Ihlebæk 2016). Bruns (2016) argued that gate-watching within social media spaces has become prevalent. The traditional concept of gatekeepers refers to the power editors have to decide what material should be published, while gate-watchers refer to the power of the audience to decide which news stories deserve extra attention through re-distribution on social media (Bruns 2005, 2016; Singer 2014). This is also relevant to social media content published by ordinary people. Updates on Facebook and Twitter that are liked or shared by many users have greater visibility, which increases the chances of them being noticed by a news organization. The crowd engages in a kind of collaborative gate-watching action (Bruns 2016). It may be argued that the level of attention given to social media posts is becoming a news criterion, and that (as has been noted) many of the news stories derived from user-generated content will refer to the number of shares and likes in the headline or the lead paragraph. Therefore, the following hypothesis is proposed:

H1: 

The majority of articles refer to popularity cues.

The question then arises of whether all news organizations are equally keen on using ordinary people and social media as a source. Norwegian newspapers have traditionally been divided across three geographical levels. The local newspapers cover news from smaller towns or rural areas and cover a limited amount of national and international news. The regional newspapers are situated in larger cities, and cover news from there and the surrounding county, and sometimes also from neighboring counties. They carry some national and international news. The national newspapers have their main offices in Norway’s capital, Oslo, and feature news from the entire country, with an emphasis on the most populated areas. They carry more international news than regional newspapers. This structure dates from the days of paper distribution, but it is still basically intact. In addition, there are a few online newspapers that have never had paper editions, and TV channels that have branched out and created online newspapers as part of their services. The structure encompasses most of the country’s news production, so RQ6 investigates whether it influences the use of social media as a source.

RQ6: Which online newspapers most frequently use ordinary people on social media as sources?

3. Method

The data analyzed in the present study consist of news articles from Norwegian online newspapers that are based on social media posts published by ordinary people. Criterion sampling was used, in which predefined criteria were set up to collect cases. The following criteria were used:

• Norwegian online newspapers only were selected. Paper editions, radio, and TV were not included.
• Only editor-controlled online newspapers were included. Alternative news sites, weekly magazines, and viral sites were excluded.
• The main content of the news story had to be based on a social media post. Stories in which tweets or other quotes from social media were used as additional elements in a larger story were not included.
• The social media posts had to be by ordinary people, who in the present context are understood as people who are not media personalities, and who do not hold positions that are scrutinized by the media.

3.1. Data Collection

The data were collected using Retriever (https://www.retrievergroup.com, accessed on 30 August 2017), a service that allows searches of any Norwegian newspaper. The search period spanned the first six months of 2017. Two sets of search words were used. The first were the phrases “goes viral”, “explodes on the internet”, “takes off”, and “is shared”. These are typical phrases used in Norwegian newspapers when addressing viral trends. There has not been a thorough mapping of what phrases are most commonly used to describe virality in Norwegian newspapers, so the selection of phrases is done on the author’s assumption after reading many stories on virality. This of course mainly resulted in articles that included popularity clues somewhere in the text.

A separate search was completed on the words Facebook, Twitter, Instagram, Snapchat, Tumblr, LinkedIn, and social media. They were chosen because they were the most popular social media in Norway during the time of the data collection (Medienorge 2021). Twitter has fewer users than these but was chosen because it is very popular among journalists (Kalsnes 2016). Including other social media such as Yodel, Saharah, and Discord as search words might give different results today, as they are popular among young users, and of course, TikTok, which has become very popular in Norway in the last few years (Medienorge 2021).

To reduce the number of irrelevant hits, the words “police”, “Trump”, “TV”, “betting (sport)”, “reality show”, and the names of specific reality shows were excluded. Fifty-nine online newspapers were searched. Seven of them were nationwide newspapers (the largest ones were selected), nine were regional newspapers, and 43 were local. Regional and local newspapers from across the country were selected to ensure the selection was geographically dispersed. This resulted in a return of 291 articles.

3.2. Analysis

The unit of analysis was the news article, not the original social media post. The coding focused on the type of publication, the topic of the article, whether the headline and lead paragraph featured the topic or the post’s virality, and which sources and pictures were used. The topic describes the main theme of the article. The categorization scheme was taken from Kalsnes and Larsson (2017) who, inspired by Sjøvaag et al. (2012), analyzed article topics according to nine main categories:

• Politics (including war and terrorism, demonstrations, public administration, harassment, political humor, and regional politics)
• Economy (including personal economy, business, finance, and markets)
• Crime (including murders, violence, trafficking, policy issues, and trials)
• Social issues (including work, health and education, environmental issues, consumer, construction, animal welfare, and traffic issues)
• Culture (including the arts, media, royalty, curiosities, leisure, nature photos, and popular culture)
• Sports
• Accidents
• Weather
• Science and technology issues

The author of this paper coded the material, using a coding key specifically developed for this study. A coding reliability test was performed by having a student assistant code 20% of the material, or 60 cases, using the same coding key. Using Messner and Distaso’s (2008) description of Holsti’s method, we tested the coder reliability. The similarity between the main coder and the test coder was 95% for the topic, 100% on the gender and age of the main source, and 92% on the type of newspaper. This implies that the coding reliability is satisfactory. This method does not take into account the instances of coding similarity that could have arisen by chance. This means that the reliability might be exaggerated using this method. Still, Holsti’s method is a good measure of coder reliability for this study. This is because it was the manifest meaning of a limited material that was coded, not the latent meaning, and, to avoid shared preconceptions, the assistant was not involved in the development of the coding key.

3.3. Limitations

The material was too limited to allow for generalizability, but the findings will hopefully initiate a discussion about news organizations and user-generated material from social media, and the use of ordinary people as sources. The data did not hold information on how news organizations found the material on social media. It can only be said that they found certain posts newsworthy. There was no bias in the selection, but there was the possibility that some news articles were left out. Searches on the term “social media” and related words gave a huge amount of hits in Retriever, and most of them were not relevant to the present study. However, excluding the irrelevant cases was a tedious process, so some relevant examples may have been overlooked. Some articles may also have been represented more than once. Several newspapers cooperate through networks; they can re-publish stories, sometimes without any changes, sometimes with. However, this is in principle not very different from when several newspapers publish the same news story released by a news agency. It is not, therefore, considered to be a flaw.

4. Results

4.1. A Variety of Topics

RQ1 concerns the topics covered in the news stories.

Table 1. Topic of article.

Topic	Percent
Social issues	48
Culture	32
Politics	11
Sports	4
Crime	3.5
Economy	0.5
Accidents	0.5
Weather	0.5
Science/technology	0
Total	100
N: 291

displays that social issues were by far the largest category, followed by culture and politics. There were few articles on sport and crime, and almost none on the economy, accidents, weather, or science and technology. There were great variations within the three largest categories, which are discussed below.

4.1.1. Social Issues

The most prominent topics were health care, elderly care, consumer action, and traffic issues.

This news story (Figure 1) from a local newspaper is a typical example. A lady visited her mother at a nursing home and was not happy with the unappetizing dinner her mother was being served. She took a picture of the food and posted it on Facebook, stating in the accompanying text that she was angry and that our elders deserve better. The post received a large number of likes, comments, and shares. The newspaper referred to the virality of the post both in the headline and in the lead paragraph. The main picture in the news article was taken from the Facebook post.

Consumer Action

The articles on consumer action were mostly focused on social media posts in which consumers complained about products they had purchased, poor customer service, or public service. The complaints were often directed against a commercial company, a transport provider, or a public institution. In particular, there were many articles about beer. One of the major grocery store chains in Norway had announced that they were cutting down on the number of distributors, including breweries. This caused a social media uproar from people complaining that their local brand of beer would be removed from the shelves.

Traffic Issues

The traffic category was dominated by the so-called “war” between cyclists and car drivers. These posts often included videos from dashcams or mobile phones.

Figure 2 shows a typical example. A bus driver filmed a cyclist with his mobile phone (in some cases they use dashcams), posted the video on Facebook, and complained that the cyclist was in the middle of the road and that he slowed down the traffic. The headline of the news article focused on the cyclist’s refusal to give way, and the lead paragraph stated that the video was “spreading like wildfire on Facebook”. The article fueled an angry debate in the comments section of the newspaper. Such heated debates were a common feature in cases concerning cyclists and motorists.

4.1.2. Culture

Curiosities, media, and leisure were the main topics in the culture category.

Curiosities

There were various entertaining oddities in the material, such as advertisements for houses with very unusual interiors, spectacular marriage proposals, pets, and creative and humorous content.

Figure 3 is a screenshot from a news story concerning a humorous video that had received a great deal of attention on Facebook. In it, a young man talks in a colorful dialect about how life in the northern parts of Norway is much better than in the south. He is wearing a beach outfit while surrounded by snow, and while he talks, he adjusts a beach chair and places it in the snow. This ironic video was posted in March when the Oslo-based newspapers usually report on the warm spring weather in the south, whilst the northern parts of the country are still covered in snow. The video plays to the time-honored antipathy between north and south in Norway.

The video was picked up by one of the major tabloids in Norway, and republished under the heading “viral videos”, a heading under which they publish various videos from around the world. The video was republished in full length, without any further comments or interviews, or any other form of journalistic processing.

Media

Articles in this category invariably had social media as the topic. They included stories where somebody had shared a photo that had then received huge numbers of shares and likes. Many of the stories originating from Instagram were for the most part in the media category. Most of them were not based on a single posting but were rather small portraits of a person of local interest. They were not professional influencers but mimicked the style of the more famous influencers. The articles contained a general description of the person’s Instagram content, which featured either fashion and make-up, or fitness and health. Only local and regional newspapers ran these kinds of articles.

Figure 4 is a typical example of a media story from a regional newspaper (Drammens Tidende). The headline stated that “Sofie has a quarter of a million followers on Instagram”. The lead paragraph repeated this information but phrased it differently. The article contained an interview with her, and a few pictures from her Instagram profile, which consisted mainly of artistic fashion selfies.

Leisure

Articles in the leisure category dealt with people’s hobbies, such as restoring old cars, travel, and outdoor activities.

4.1.3. Politics

The political stories overlapped with other topics, particularly elderly care and traffic issues. The main subjects in this category were humor, stories about harassment, and regional policy.

Figure 5 features an article from the local newspaper Midtsiden. The story concerned an online petition against the building of a road bridge. The lead paragraph stated that the “goes viral on Facebook” with 829 signatures and served as an example of the relative measurement of what counts as “going viral.” The article replicated the petition from the Facebook page and used a picture from a previous rally about the same issue. This was an example also of how traditional forms of citizen activism merge with digital tools such as online petitions and social media.

Humor and Harassment

The humor category mainly contained jokes and satire about well-known politicians. The harassment articles dealt with people making harassing, hateful, or threatening remarks toward well-known politicians.

4.1.4. Advertisements

Advertisements overlapped with several main categories, but also stood as a category in their own right. People advertise items for sale on social media, and occasionally the ads become news stories. There were three types of advertisements: (a) humorous and ironic advertisements, for instance from someone selling a car who described in full detail how poor it is and how badly they wanted to get rid of it; (b) sale advertisements for houses or apartments with unusual interior design; and (c) job seekers who had published an open job application on social media (usually Facebook).

Figure 6 is a typical example of an open job application on Facebook. The person posted a picture of themself with sections of their resume. They described their difficulty in finding employment and how motivated they were. The news article republished most of the Facebook posts, including the picture, with a screenshot of the text from the post. The newspaper even replicated the original spelling errors. Was this done on purpose, or was it merely a consequence of fast cut-and-paste journalism? The headline and lead paragraph both emphasized the virality of the post.

4.2. Gender and Age of the Sources

RQ2 investigated the gender and age of the source creators. Of those who posted the content on social media, 57% were men and 43% were women. Their age was mentioned in only 26% of the news articles. The range was from 11 to 90. The largest group of posters was between 20 and 29 (55% of all cases). Twenty percent were below the age of 20, and fourteen percent were between 30 and 39. Since age was only stated in 26% of the features, and we did not want to embark on the imprecise task of guessing people’s age based on photographs, we could not make any assumptions about whether the sources recruited from social media were younger or older than the average age of traditional news sources. It may also have been the case that the newspapers simply found it more interesting to state the age of their sources when they were young.

The people interviewed were in most cases (66%) referred to by their names only (i.e., without a title). In 14% of the articles their professions were referred to, in 6% their family position (i.e., mother, father, and so on), and in 6% their capacity as a performer of a specific activity (e.g., scuba diver, mountain climber, amateur photographer, and so on). A total of 3% of the articles named them as the owner of something, usually a house or an apartment, and 3% as a volunteer in an organization. In 2% of the articles, they were anonymous.

4.3. User-Generated Visuals: Photos and Videos

RQ4 dealt with the type of visual material that was used in the articles. The results showed that there was a significant proportion of amateur pictures and some videos in the material. All but one of the articles was illustrated with pictures, videos, or screenshots. Seventy-four percent contained user-generated material as the principal visual element. Thirty-three percent had a private photo of the user as its main picture, while thirteen percent had a private photo of an object. The same percentage contained a screenshot from the original post as its main picture, and 11% used video taken from the social media post. Sixteen percent used professionally taken photographs of the interview subject as their main visual. The remaining 10% included other types of professionally taken photographs.

Thirteen percent of the articles featured video, with eleven percent of them having video as the main visual. All the videos were user-generated and were taken from the original social media post. The majority of the videos were within the humor, arts, or nature sub-categories. Several of the videos dealt with issues such as elderly care, traffic safety (as has been noted), and animal welfare.

4.4. Additional Sources

A traditional journalistic principle is to refer to several sources. RQ5 asked whether content from social media was the main source in the articles and whether additional sources were used. The results are presented in

Table 2. Sources.

Sources	Percent
SoMe post only	7
SoMe post and individual	50
Additional sources	43
Total	100
N: 291

.

“SoMe post only” refers to articles where they only used material from the social media post. “SoMe post and individual” refers to articles where they used material from the social media post in combination with an interview with the person who posted it.

Fifty-seven percent of the articles used only one source, namely the originator of the social media post. Seven percent used social media posts exclusively. The most common format for the articles was to replicate some of the content from the social media post along with an interview with the person who posted it. Forty-three percent of the articles referred to additional sources. In 84% of these, the social media user was still the main source. Additional sources were most commonly other ordinary people, private companies, and public offices.

4.5. Virality as a News Criteria

H1 proposed that the majority of articles would refer to popularity cues. The material did not contain any information about why the news workers found certain social media posts to be newsworthy, but the headlines and lead paragraphs usually provided some justification as to why the article should be read. The hypothesis was strengthened: 57% of the headlines referred to the virality of the social media post, and 82% did so in either the headline or lead paragraph, or both. Forty-five percent of the articles dealt with the virality of the social media post in the main body as well, while 54% of the articles used most of the main body to address the topic of the post. Only 1% of the articles developed the subject in the main body, which suggested that the topic itself was rarely the news criterion; virality was the main selling point. Although 43% of the articles used additional sources, only 7% of the articles elaborated on the topic using additional research. In most cases, supplementary sources provided little more than a short comment. It appeared that in the majority of cases, the topic of the post was not newsworthy in itself, but that the popularity cues from social media made it newsworthy. There does not seem to be a definitive measure of what counts as going viral. Some of the articles covered social media posts that had been shared 25,000 times, while in other articles, a few hundred reactions on social media qualified as going viral.

4.6. Geography

Popularity cues as a news criterion were related to the geographical area covered by the newspapers. Unlike local and regional newspapers, the national newspapers chose stories from across the country. This was one reason why they generated articles from popular social media posts more often, as displayed in

Table 3. Type of newspaper.

Type of Newspaper	Papers	Articles	Average
Local newspapers	44	145	3
Regional newspapers	8	37	5
National newspapers	7	109	16
Total	59	291

.

There are only a few national newspapers in Norway. Seven of them were included in the present study. They each had 16 articles that are based on social media posts in the period covered. The regional and local newspapers averaged five and three articles, respectively. This was partly because social media posts had to be relevant to the specific areas covered by these outlets, or they had to be posted by someone in those areas; national newspapers were not so confined. National newspapers also employ more people, so it may have been that they had more resources to monitor social media.

5. Summary and Discussion

The standard procedure for the newspapers was to have an interview, usually by phone, with the person who published the post. This would serve as the main source for the article. The body of the text usually contained large parts of the original post. Additional sources were not commonly used. The articles were episodic in form, and they focused more on individual relevance than on societal issues. This finding was consistent with earlier research that showed that newspaper articles featuring user-generated content consisted mainly of personal and everyday life-oriented stories, rather than news or information (Jönsson and Örnebring 2011). Only on a few occasions did the newspapers elaborate on the topic with multiple sources, further research, or follow-up stories. The newspapers did not employ a great number of resources in the majority of the articles.

This low-budget journalism was also reflected in the use of visual material, which consisted principally of private photos. The main picture in the article was usually a portrait or headshot taken from the person’s social media page, occasionally a screenshot of the social media post, and in some cases a user-generated video. Only a few articles featured photos taken by professional photographers. Almost all the amateur photos and videos in the material were of a mundane type, connected to soft news and person-oriented content. The photos did not hold much factual information, but signals that the source is from the same group as the readers (Näsi 2015). The information in the articles lay mainly in the written text, with the photos serving as illustrations, although it seemed unlikely that the news articles would have been published without accompanying photos or other visual material.

Previous research has indicated that social media has become a source of hard and soft news (Paulussen and Harder 2014). The majority of the articles in the present study must be regarded as soft news, but they covered a great variety of topics. The stories based on Instagram posts covered topics such as fashion and fitness, while those based on Facebook posts, which dominated the material, covered a greater variety of subjects. The largest topics were health care, elderly care, consumer action, traffic issues (social issues), curiosities, media and leisure (culture), humor and harassment, and regional politics (politics). There were also some articles covering personal sales ads and open job applications. There were not many examples of typical firestorms in the material. This can be because of the focus on single media posts going viral. The few examples of firestorms or negative electronic word-of-mouth were related to harassment in the politics category, and customer complaints in the consumer action category. Popular news topics such as sports, crime, and accidents were barely represented in the material. The same applied to the economy, the weather, and science and technology. A previous content analysis of Norwegian online newspapers showed that sport, crime, and accidents comprised 30% of the news articles. Twenty percent of the articles were in the politics category (Elgesem et al. 2010), compared with 11% in the present material. Even though some of the articles in the social issues category, particularly those concerning health care and elderly care, often contained political references, it seems that when source material shifts to social media, it leads to different topics. The story selection here is based on popularity cues, where the audience has given input that they think these stories are interesting or amusing. In one way, this is a free-market research and a democratic way of story selection. However, one must also be aware that popularity cues on social media can be manipulated. Still, with some know-how, it is possible to use the activity of ordinary people on social media as indicators of which types of stories and topics they are interested in, and these topics often differ from regular news topics. It is also an opportunity to come in contact with people outside the usual source pool. It is an opportunity missed when newspapers merely present a snapshot instead of developing the topic with further research and investigation.

The sources in the present material consisted of men (57%) and women (43%). In Norwegian news in general, the proportion of female sources is 36% but is considerably higher (53%) in social issues, health, and entertainment/culture news (Dahlstrøm et al. 2017). The latter figure means that it is not possible to conclude that social media has led to the use of more female sources. The proportion of female sources in Norwegian news has increased slowly but steadily over the last four decades (ibid.), so other factors than social media have probably been more significant. It is nevertheless plausible that social media has allowed journalists to come in contact with sources outside the elite groups they are routinely in contact with.

Sources over the age of 67 are virtually non-existent in Norwegian news (Dahlstrøm and Hognestad 2016). There were a few sources over the age of 60 and even 90 in the present study, and some sources were under 20, but 66% of the sources were in the age range of 20–59. In legacy media, 86% of the sources were between 20 and 66 (Dahlstrøm et al. 2017). This is not sufficient to claim that the age range is wider when social media is used as a source. Also, the age of the source was mentioned in only 26% of the articles. That the 20−29 age group was the largest in this material does not necessarily mean that the sources were younger on average than the sources for news in general. It might simply be an effect of when newspapers find it appropriate or interesting to state the age of their sources.

When ordinary people are used as sources in Norwegian news media, they are usually presented as victims or illustrative cases in stories of dramatic events, accidents, crime, or the closing down of a business (Allern 2015). It has also been found that ordinary citizens are more often paraphrased than directly quoted, and are referred to in collective and anonymous terms (Paulussen and Harder 2014). This will in most cases imply that they lose control over their own story, as Palmer (2018) explains. In the material in the present study, ordinary people, as the main source, were treated differently. They were the main characters in the stories; the main photo in the item was usually of them, they were directly quoted, and they occupied most of the space in the text. When elite sources were included, their function was mainly to comment on or react to statements made by ordinary people. It is plausible to assume that this allows them to be more in control of the story, although it is still the journalists who filter the story and grant access to a larger audience. The news audience is different from the intended audience on social media, opening up the risk of context collapse (Hodkinson 2016; Marwick and Boyd 2014). The sources in the present study were usually referred to by their names only (i.e., their profession or position was rarely mentioned). Using Tuchman’s (1978) notion that ordinary people appear in the news as representatives of a group, rather than themselves, we see that they are not given the individual traits that they display themselves on social media. While there may be good journalistic reasons for this, it also means that the subjects lose control over this part of what Goffman (1956) sees as identity performance. Parts of their identity performance are now in the hands of the journalist. The informants in Palmer’s (2018) study often gave altruistic motives for speaking to the press (educate others, raise awareness), and giving away control over the identity performance might be necessary to gain access to a larger audience, but they also resented how the journalist controlled the story and access to the audience (Palmer 2018). Losing control over their own story was the biggest problem for people who found themselves in a media controversy in Norway (Duckert and Karlsen 2017). However, this might be different when so many of the news articles build on what they have already published on social media. In most cases in the present study, very little of the social media was changed before it was published as a news story.

Events that are close in time, space, and culture are more likely to become news stories, and so do negative, unexpected, and odd occurrences and events that can be linked to political, economic, cultural, or athletic elites. Accidents, crime, celebrities, and sports are therefore more likely to be covered (Wadbring and Ödmark 2016). It is also helpful if the story has a sex interest (Aalberg and Elvestad 2012). Previous studies have indicated that in news stories involving Facebook, Twitter, and YouTube, the social media platform was the subject of the story itself or connected to the story’s subject (Paulussen and Harder 2014). The great majority of the articles in the present study referred to social media in their headlines and lead paragraphs, especially numbers of likes and shares, and although social media themselves were not the subject of the story, they were certainly given prominence as a justification of why the story existed at all. A high number of shares, comments, and reactions increases the possibility that the post will be noticed by news workers, and will be taken as an indication of relevance (Bruns 2016). Such popularity cues were mentioned in the headline or lead paragraph in a large proportion of the news articles. This strengthened the hypothesis that popularity cues on social media are becoming a news criterion in their own right. Hence, collective gate-watching in the form of shares, comments, and reactions on social media can influence the news agenda. However, news organizations must consider whether the figures might have been manipulated, either automatically or by organized campaigns.

According to Schmieder (2015), non-elite sources are ordinarily regarded as being more closely related to regional than metropolitan journalism. This is not the case with social media; national news services, with their main offices in the larger cities, were responsible for nearly half of the articles examined in the present study. National newspapers can pick social media stories originating from anywhere in the country, but local newspapers in most cases need geographical proximity for a social media post to be of interest. There was also a clear tendency for local newspapers to publish social media-related stories from their geographical area after they had been covered by a national outlet. This suggests that the national press spends more resources monitoring and using social media to generate news stories.

6. Conclusions

Previous researchers have pointed out that journalists are now more likely to use mobile phones or various forms of online information than to meet people face-to-face, and there have been concerns that this has led to more homogeneous content based on the same online sources, and that journalists are less likely to question the material they access (Davis and Love 2018; Elvestad and Philips 2018). It seems from the material in the present study that online sources can be just as varied, or limited, as face-to-face sources. But it also seems that the journalistic processing of the material found on social media is minimal, with few questions asked and little research completed. Is this the result of the journalistic evaluation, time pressure, or economic restraints?

The use of social media as a source no doubt has a part-financial motive. News organizations have had to downsize and reduce staff, and the journalists themselves have found that this has hurt their capacity to work with multiple sources (Dahlstrøm and Hognestad 2016; Holt and Karlsson 2015). But there are also journalistic reasons for using ordinary people and social media as sources. For instance, social media can create an impression of immediacy, which is a highly sought-after value in news reporting. Immediacy has a time and space dimension; the journalists are bound to report the latest news as soon as possible, and events that are geographically close to the audience appear to be more important. It is also a reflection of the perceived cultural and social space between the audience and the event. Immediacy is also perceived presence, in that communication appears as a personal conversation, a direct address. Much of this can be seen in social media postings; they are mostly personal in their address, they are culturally proximate, and they are often written by peers. The social media posts in the material in the present study also represent more spontaneous and unfiltered outbursts than are to be found in traditional media interviews, press conferences, press releases, and so on.

Paulussen and Harder (2014) argued that social media is not very likely to lead to an increased diversity of voices in the news, because professional journalists still tend to use established sources of information, and this network of sources is difficult to break into. Studies from the Nordic countries have indicated that social media do not replace the usual elite sources, but rather serve as an addition (Skogerbø and Karlsen 2014). This suggests that social media can contribute to a greater variety of news sources, but we cannot expect this effect to be huge. There were slightly more female sources for the material here than in Norwegian news in general, but at the same time, the topics were to a large extent in the social issues category, where female sources are more common. This suggests that other factors are likely to be more important than social media when it comes to source diversity in the news, but that social media can be a contributing factor.

Both print and online newspapers have attempted various strategies to engage and maintain their readers. Letters to the editor have long been an important way for the public to express their opinions (Bruns 2005; Ihlebæk and Krumsvik 2015). In online newspapers, comment sections are a common feature, but they have not gained a wide audience (Hermida and Thurman 2008; Bergström and Wadbring 2014). A Swedish study claimed that most readers have never posted in comments sections, and they were used less and less frequently after 2011 (Almgren and Olsson 2016). The various social media attract a broad range of users, and there are great variations in use; some are lurkers who rarely contribute content or interact, some use the platforms mainly for socializing, some are eager debaters, and others put them to a wide variety of purposes (Brandtzaeg 2012).

The use of social media by traditional media can be seen as part of the “hybrid news system,” where sources circulate between the two sectors (Chadwick 2011, 2013; Hladík and Štětka 2015), or as a continuation of the mutual influencing of each other’s agendas (Kalsnes 2016). Earlier studies have shown how newspapers influence the TV agenda, and how campaign websites influence the news agenda (Messner and Distaso 2008). That social media can influence the news agenda should therefore come as no surprise. Social media constitute a place where journalists can come in contact with ordinary people and vice versa. In Palmer’s study (Palmer 2018), many of the subjects contacted journalists themselves, because they had information or experiences that they thought needed to be shared with a larger audience. We do not know the motives of the persons in the present study, but it seems like they turn to social media, rather than a journalist when they have something to say. It is easier to publish it yourself, although for a smaller crowd than to get the attention of a journalist. This has contributed to social media becoming one of the arenas for public debate. It also means that content that was intended for friends and relatives sometimes gets a larger and more diverse audience.

Social media functions as a more popular arena for discussion and debate on current topics than the newspapers’ comments sections; and, instead of letters to the editor, many seem to prefer to post on social media if they have something to say, or something funny, or if they are simply annoyed about something. In effect, parts of the public debate have moved to social media. Newspapers have a democratic responsibility to be a forum for public debate (Dahlgren 2009; Ihlebæk and Krumsvik 2015), so it is natural for them to include social media as part of their commitment. It is interesting to investigate more about how this affects the persons involved, and the public debate itself.

7. Further Research

Journalists possess the power to frame stories, selecting the angle and preferred interpretation to present to the audience (Jensen 2016). In this respect, it would be interesting to investigate how journalists monitor and use social media to find material, and why they choose to present the stories in the way that they do. Interviews with journalists might shed more light on the criteria that are applied when user-generated social media content is selected for publication. It would also be interesting to study how becoming news subjects in this way affects the individual. Popularity cues can function as a news criterion, but one should also look at how such cues can be manipulated, either algorithmically or by organized campaigns, in an attempt to affect the news agenda.

Research with different search words might also give different results. There is no thorough mapping of what phrases are most commonly used to describe virality, and this also changes over time. For instance, when English texts are translated into Norwegian, it can bring along anglicisms, affecting the way we describe virality in Norwegian. After the time of the data collection, anglicisms such as “trending” and “trender” have been used this way.

Including other social media such as Yodel, Saharah, and Discord as search words might give different results today, as they are popular among young users, and of course, TikTok, which has become very popular in Norway in the last few years.