Attracting the news: Algorithms,platforms,and reframing incidental exposure

Abstract

This article argues for new approaches to the study of incidental exposure that better account for the role of algorithms, platforms, and processes of datafication in shaping the likelihood of news exposure online. It offers a critique of three themes prominent in the incidental exposure literature: (1) incidental exposure connotes accidental exposure to news on social media, (2) news content is ubiquitous on social media, and (3) incidental exposure can be conceptually distinguished from intentional exposure to news on social media. This article proposes a new metaphor to reframe research on incidental exposure: ‘attraction’ to news.

Keywords

Incidental exposure social media information inequality

The scholarly literature on incidental exposure to news has grown rapidly alongside the growth of social media platforms. It is a rich, optimistic literature that pays careful attention to the potential benefits of incidental exposure: learning about public affairs, increased exposure to diverse news sources, and enhanced engagement in civic life (e.g. Bode, 2016; Feezell, 2018; Fletcher and Nielsen, 2018; Oeldorf-Hirsch, 2018; Valeriani and Vaccari, 2016). At the same time, less attention has been paid to whether and how incidental encounters with news happen in the first place. Few studies ask who is incidentally exposed to news on social media and who is not, or theorize how such exposures occur (though see Boczkowski et al., 2018). This article undertakes a critical review of existing literature concerned with incidental exposure to news on social media, in order to advocate for increased attention to the role of platforms and algorithms in shaping the possibilities for incidental exposure. I argue against relying on the metaphor of incidental exposure as accidental exposure to news in this era of algorithmic curation, and call for an empirical debate over whether news exposure on social media is as ubiquitous as is often assumed.

In the second part of this article, I propose a new metaphor to help us think about the entanglement of actors and practices that produce news exposure on social media. My suggestion is to think in terms of attracting news, where attraction is characterized by fields of forces working in different ways. This new metaphor pushes us to investigate who attracts news on social media and who does not, and to consider the broader set of actors whose choices shape individual level news exposure. Users of algorithmically curated platforms are differentially attractive to news because of the entanglements of their own choices with datafication processes taking place within and across digital platforms, alongside the content production and dissemination practices of news publishers. When we move from thinking about incidental exposure as accidental exposure, this reframing focuses our attention on the complex ways in which power over incidental exposure is increasingly in the hands of platform companies. Finally, I illustrate how the attraction metaphor points in new directions for incidental exposure research.

Challenging key concepts of incidental exposure

The Oxford Dictionary offers two definitions for the word ‘incidental’ (Oxford Dictionaries, n.d.). The first emphasizes an incidental occurrence as something that happens as an ‘accompaniment to something else’, a byproduct activity with a strong element of serendipity: ‘occurring by chance in connection with something else’ (emphasis mine). The second definition emphasizes that an incidental occurrence happens ‘as a result of (an activity)’, in the sense of accompanying or attendant on. ‘The ordinary risks incidental to a fireman’s job’.

Both of these meanings are prominent in the literature on incidental exposure to news. Incidental exposure is often described as happening at random – the active metaphor is accidental exposure to news. Encounters with news are also often described as attendant on use of social media, a natural consequence of the ubiquity of news online. And incidental exposure is typically conceived of as distinct from intentional news exposure, rather than investigating the ways in which news seeking and stumbling upon news are intertwined on digital platforms. As I show below, these three premises in our understanding of incidental exposure need to be revisited in light of growth in algorithmically curated, data-driven platforms (Van Dijck et al., 2018).

Incidental exposure connotes accidental exposure to news

Many researchers credit Downs (1957) with directing scholarly attention toward the possibilities of incidental exposure to news. Downs distinguished between ‘sought-for’ information and ‘accidental’ information. Accidental information can be acquired at a much lower cost because accidents of exposure occur without any investment of effort. They are serendipitous: seeing 10 minutes of CNN in the doctor’s waiting room, or watching a breaking news segment that interrupts a sporting event.

Over time, the ‘accident’ metaphor became entwined with the second meaning of incidental exposure, that incidental exposure to news is a byproduct of doing other activities. Audiences stumble upon news while they are using media for another purpose, such as entertainment. Baum (2002) argued for the importance of ‘soft news’ as source of political information for Americans, proposing that soft news coverage of public affairs could reach citizens via a process of ‘incidental attention’. Tewksbury et al. (2001) explored whether publics could become ‘accidentally informed’ via incidental exposure to news online. They tied their arguments to specific technological developments, proposing that the then-new Internet portal sites (Yahoo!, Lycos, Excite) created opportunities for happy accidents of incidental exposure.

This basic framing of the relationship between characteristics of the media environment and opportunities for incidental exposure remains intact today. Media systems that offer a high degree of choice allow audiences to fulfill their content preferences; low choice media environments constrain options and provide more opportunities for incidental-as-accidental news exposure (Prior, 2007; Van Aelst et al., 2017). The growth of social media has been widely seen as heralding a renaissance of opportunities for incidental exposure to news.

Throughout, the metaphor of incidental-as-accidental exposure has remained intact. For example, Valeriani and Vaccari (2016) argued that ‘accidental exposure’ to news could promote political participation. Kim et al. (2013) analyzed the effects of ‘accidental news exposure’. Yadamsuren and Erdelez (2010) defined incidental exposure as ‘accidental discovery of useful and interesting news’. Yet, social media have created a challenge for the incidental-as-accidental exposure metaphor. Algorithmically curated platforms are fueled by the automated collection, analysis, and application of user data (Srnicek, 2017; Van Dijck et al., 2018). On platforms like Facebook, incidental encounters with news are entangled with prior choices made by individuals themselves and are therefore perhaps not so accidental after all. Feedback loops between user (and friend) preferences and future content exposure emerge, which are made possible by the application of user data to decision-making about what should appear in that user’s feed.

For example, Thorson et al. (2019a) showed that Facebook users who are algorithmically classified as interested in news and politics are more likely to be exposed to news – above and beyond their own self-reported content preferences. Users who have previously engaged with news content or who have liked news pages are more likely to see related content in the future. As Kaiser et al. (2018) phrase it, ‘though news contact on Facebook may be “incidental” in the sense of “unintentional,” it does not occur “by chance.”’

It is easy to stumble upon news on social media

Research on the effects of incidental exposure to news on social media tends to assume that news on social media is ubiquitous and thus easy to stumble upon: ‘the ubiquity of news on the Internet and within social media offers people the possibility to be exposed to news whether or not they actively seek it out’ (Gil de Zúñiga et al., 2017: 105). Müller et al. (2016) write that ‘scrolling through the latest posts of one’s Facebook news feed is often (more or less incidentally) related to stumbling upon news content’ (p. 431). Kim et al. (2013) argued that ‘with increased opportunities for unintentional exposure to news on the Internet, more people may be exposed to a greater number of stories about politics and public affairs, including mobilizing information’.

The idea that news is ubiquitous on social media is worth subjecting to empirical debate. Given platform company restrictions on access to individual level data about content exposure, we have little observational evidence concerning rates of news exposure on sites like Facebook. However, what data we do have suggest that rates of news exposure on social media may be lower than seen in self-report data. Furthermore, even some survey studies of social media exposure find relatively low levels of exposure.

A 2018 Pew study found that one third of Americans never get news on social media (32%) and another quarter report ‘hardly ever’ getting news on social media (21%). Another 27% say they ‘sometimes’ get news in this way, while only 20 percent report ‘often’ seeing news on social media. This suggests that while news exposure on social media is not an unusual occurrence, it is a much smaller subset of users who experience news as ubiquitous on the site – perhaps only the 20 percent who report ‘often’ seeing news on social media. The same survey asked Facebook users, ‘Do you ever get news or news headlines on any of the following sites?’ One third of Facebook users said ‘no’. A majority of users of other social platforms also responded no to this question: Instagram (68% never see news), YouTube (61% never see news), Snapchat (70% never see news), and WhatsApp (78% never see news).

My colleagues and I collected survey data from a national sample of 18–34 year olds 1 month before the 2016 US presidential election (see Appendix 1 for details). We asked Internet users whether they had seen any content on Facebook about the election in the previous week: 54 percent said they had seen nothing about the election on the platform in the past week. We also asked how many days in the previous week participants had read about state/national/global events on social media. Forty percent said ‘0 days’, and another 12 percent said ‘1 day’. Only 9 percent reported reading news on social media 6 or 7 days per week.

Facebook itself has reported that news stories make up less than 5 percent of users’ newsfeeds, on average (Kantrowitz, 2018). (Facebook has not made these data available for replication.) The rare scholarly studies of news exposure on social media that use digital trace data typically find low levels of news exposure as well. Wells and Thorson (2017) built a Facebook application to capture newsfeed data from a sample of college students, with participants’ permission. Over the course of a 1-week data collection, the average respondent saw 7.3 posts about public affairs news – a very small proportion of the number of items that appeared in their feed. Public affairs news comprised just 1.8 percent of the average feed. The median respondent liked zero pages from journalists or news organizations.

News sharing on social media may also be less common than generally assumed. Haenschen (2019) compared self-reports about frequency of sharing news links on Facebook with digital traces of participants’ actual sharing activity. She found that news sharing in general is rare, and that respondents typically over-report their news sharing activity. Möller et al. (2019) combined a user survey with web tracking data, including a measure of social media-driven news use (i.e. clicks to a news site that originated in a social media feed). They found that engagement with news content from social media was less common than visits to news sites via search results, and that the majority of social media-driven news use was undertaken by people who reported high levels of political interest. Those with low political interest were very unlikely to visit news sites from social media feeds. Taken together, empirical findings from surveys and observational data suggest that news exposure on social media may be less ubiquitous than often suggested in the literature.

Incidental exposure is conceptually distinct from intentional exposure to news on social media

Given the entanglement of news exposure with prior user behaviors on algorithmically curated platforms, it becomes much more difficult to conceptually and empirically separate the concepts of incidental and intentional news exposure. Intentional news exposure is defined as using social media in order to seek out news. Incidental exposure occurs when a user is on social media for other purposes. The relationship between the two modes of news use is complex, because evidence suggests that intentional news seeking behavior and future (incidental) encounters with news are entangled in a feedback loop, such that more intentional engagement with news on a platform leads to more incidental exposure.

This feedback loop is made possible by datafication: the process by which social actions are quantified for use in tracking and algorithmic prediction (Van Dijck, 2014; Webster, 2011). On algorithmically curated platforms, incidental and intentional modes of news are intertwined, because users’ past behavior on the platform (and the behavior of their friends) is used to shape the array of content that is made visible in the future (DeVito, 2017). Conceptually, this suggests the need to distinguish between, say, a Facebook user who encounters a New York Times story because she previously liked the newspaper’s page (and who perhaps has previously clicked on news content) and a Facebook user who encounters the same story as a true accident, perhaps because it was shared by her Uncle Frankie – who even knew he read the newspaper? As Kümpel has suggested, we should begin to think of incidental exposure on a continuum. The extent to which news exposure on social media is serendipitous and empirically distinct from intentionality is a matter of degree.

Empirically, this means that survey measures of incidental exposure are often positively correlated with intentional exposure, a relationship that deserves to be probed. Ahmadi and Wohn (2018) found that intentional news use on social media (searching and seeking behaviors) had a strong, positive correlation with incidental exposure to news on social media. Information seeking motivation was also a strong, positive predictor of incidental exposure. The authors suggest these findings imply a paradox, such that ‘social media users may get exposed to information they need in an unintentional manner through purposeful information seeking’ (p. 5). Similarly, Oeldorf-Hirsch (2018) reported a .71 correlation between news seeking on Facebook and incidental encounters with news on the site. News exposures on social media depend on prior acts of intentionality, because the display of content on sites like Facebook is shaped by an algorithmic system optimized to keep users engaged via delivery of relevant content. Thus, the datafication of user behavior online is transforming the processes of news exposure at the individual level, blurring the lines between seeking and stumbling upon news.

News exposure as a system

As discussed above, the concepts that we typically draw on to analyze news exposure have not yet been sufficiently updated to take into account processes of datafication across digital platforms. Exposure to news on social media is the result of a series of intertwined choices made by individual users, news organizations, peers, algorithms, and other elements of the design of platforms (Thorson and Wells, 2016). News exposure on any algorithmically curated platform occurs within a sociotechnical system of information flows in which datafication processes enable feedback loops over time (Meadows, 2008). Because the Facebook newsfeed ranking algorithm is engineered to deliver content that users will find relevant and engaging (DeVito, 2017), individual choices to read, watch, or like news on social media will lead to increased future delivery of news to the feed (Thorson et al., 2019a). Intentional exposure to news on social media is thus causally related to future incidental exposure to news.

Up to this point, we have focused only on the role of individual choices and their entanglement with future news exposures through news feed ranking algorithms. However, there are more actors involved in shaping news exposure on Facebook than just users and algorithms. The choices made by friends and even news organizations themselves are also important shaping factors. People who are interested in news and politics are more likely to have friends who share news content on social media (Thorson et al., 2019a) – a typical finding of homophily as in many studies of social networks. This is another factor that explains why not all social media users are equally likely to ‘stumble upon’ news.

Much less widely studied are the ways that user choices about what to read and watch on social media are related to future news exposure through the choices made by news organizations. News organizations are increasingly reliant on digital data for story selection and advertising targeting, suggesting the emergence of data-driven processes for many aspects of news production and circulation (Lee and Tandoc, 2017; Van Dijck et al., 2018). In one recent example, Facebook hosted a 12-week program for local news publishers in the United States called the ‘Local News Subscription Accelerator’. A Facebook blog post reporting on the event captures the many ways in which local newsrooms are becoming more ‘data-driven’ (Mendoza, 2018). The Omaha World-Herald described ‘getting the newsroom on board’ with digital data, including choosing ‘story ideas with input from data and consumer behavior’. Newsday reported analyzing 15 months of content ‘to determine which content is most valuable to conversion/retention’. Other news organizations described increased use of paid targeting to reach subsets of Facebook audiences who might be interested in engaging with news content and in paying for subscriptions. For example, the The Atlanta Journal-Constitution noted their ‘improved targeting of high-propensity behaviors’. That is, they are increasingly effective at targeting paid news exposure to users who are likely to engage with that content or pay for a subscription.

These examples show that newsroom behaviors are entangled with data about user preferences, again mediated through algorithmic systems (Webster, 2011). Story topics that engage audiences on social media receive more news coverage, giving future audiences more opportunities to engage with those stories – and fewer opportunities to engage with other, less-covered stories (Bell et al., 2017; Welbers et al., 2016). Users who engage in behaviors that indicate likelihood to engage with news become more likely to be targeted with news content by news organizations. Thus, the reliance of news organizations on insights gleaned from user data produces additional routes through which intentional user choices are related, indirectly, to future incidental exposure to news.

Algorithmic systems and processes of datafication mean that control over news exposure is no longer solely in the hands of any single actor. Instead, news exposure is shaped by choices made by an assemblage of human and non-human actors, including media users, platform companies, algorithms, news organizations and journalists, and advertisers (Bucher, 2012; Van Dijck et al., 2018). These entanglements are reshaping news exposure in ways that have not yet been fully reflected in the main body of literature on incidental exposure (Bell et al., 2017; Kaiser et al., 2018; Napoli, 2014; Nielsen and Ganter, 2018; Thorson and Wells, 2016).

Attracting the news

In light of the above discussion, it is perhaps a good time to reframe incidental exposure around a new metaphor, one that helps us articulate the shift in power over exposure toward a broader assemblage of actors. A new metaphor should highlight the power of platforms and algorithms, as well as their entanglement with user preferences expressed through data. I propose we retire the metaphor of incidental exposure as accidental exposure (at least for use in analysis of news exposure via digital platforms) and explore instead the metaphor of ‘attraction to news’.

All metaphors have entailments that can help us expand our thinking in new directions (or, conversely, constrain our imaginations; Krippendorff, 1993). The entailments of attraction evoke both the idea of magnetic attraction – ‘A force under the influence of which objects tend to move towards each other’ – as well as the qualities of attraction to a person or an object, ‘the action or power of evoking interest in or liking for someone or something’, and ‘a quality or feature that evokes interest, liking, or desire’ (Lexico, n.d.). All of these entailments are useful in thinking through news exposure within algorithmically curated and datafied media distribution platforms.

First, some social media users attract more news than others. Audiences can attract news (rather than seeking it out or stumbling upon it) because user preferences are entangled with the choices of other actors on social media platforms. As outlined above, people’s actions online are read into data which are then made available to be acted on by various actors within an algorithmic system. On Facebook, the newsfeed ranking algorithm might prioritize exposure to news content for a user who has engaged with such content previously or who ‘liked’ a news page. These behaviors are read as signals that the user is interested in receiving similar content in the future (Thorson et al., 2019a). In this case, the user has attracted news through an algorithmic inference of her interests – she is classified as someone who likes news.

Users can also attract news because digital data allow news organizations to ‘see’ users who are likely to engage with news and take action based on that knowledge. Facebook and Twitter alike allow organizations to microtarget by interest and by types of behavior. That means a news organization can pay for a story to be shown to a user who has liked their page or who has been classified as having an interest in a particular topic, or in local news in general, or who has liked or followed the page of another news organization, or who simply is in a demographic category associated with high levels of news use. Datafication allows users to be classified as having a ‘high-propensity’ for news engagement, making such users more attractive targets for paid news content.

In both of the cases described above, the end user has not stumbled across the news story by accident, but rather has attracted that story into their orbit. It is important to note that attractiveness to news online is related to “real” user preferences, in the sense of the preferences that users would express on a survey. That is to say, a user’s initial attraction to news (i.e. interest in news) shapes the likelihood of becoming attractive to news online. Users who are more interested in news are more likely to engage with news online. In turn, those who have engaged with news content online are more likely to encounter news incidentally as algorithmic systems work to give that user content she will like (Thorson et al., 2019a).

We can also highlight a second-order sense in which users attract news online: aggregate user attention to news content is captured as data that is expressed to news organization through metrics such as engagement, clicks, views, and traffic. As noted above, evidence suggests these information regimes in turn shape what stories get covered and how they are presented, as journalists and news organizations are pressured by economic constraints to produce engaging content (Lee and Tandoc, 2017; Webster, 2011). Klinger and Svensson (2015) describe this process as a network logic – the ‘unwritten rules’ that govern production of content for social media that we all – including news organizations – are disciplined to follow because of our desire for visibility on the platforms (Bucher, 2012). Using the attraction metaphor, we can simply say that news users not only attract specific stories into their newsfeeds but also, on a more macro level, attract coverage of stories and news formats that data suggest they will find engaging.

Finally, the attraction metaphor highlights that news visibility is not dependent on the choices of any single actor, but rather is the result of a field of forces in interaction – a rough parallel to how we think about forces of magnetic attraction. Ananny (2016) suggests that to develop an ethics of algorithms requires considering not just lines of code or their designers but examination of a broader algorithmic assemblage, including the codes, human practices, and logics that shape the conditions and outcomes of algorithmic systems. If we want to understand incidental exposure to news content on social media, a similar task is required. The preferences of individuals are not the direct cause of news exposure on social media, although they remain important. Instead, exposure is mediated through algorithmic classification and ranking systems, and usage practices related to those systems on the part of media consumers and producers. Like magnetic fields, the strength of attraction to news varies across users. Magnetic forces not only attract but repel, reminding us that news absences on social media are also a valuable object of study: Who is not attractive to news? Who does not attract news? Finally, recognizing that not all incidental encounters with news are accidental encounters with news should help us to identify and analyze more deeply the true moments of serendipity – the moments of exposure when those who should by rights be unattractive to news get to see it anyway.

New metaphor, new research agendas

A new metaphor for incidental exposure points the way toward an expanded research agenda, one that should complement existing research focused on the effects of social media exposure to news. To understand the forces of news attraction requires us to attend to the practices of human and non-human decision makers that shape who is incidentally exposed to news and why, including making visible the diverse ways in which platforms exert control over exposure (Bell et al., 2017). Below, I briefly outline possible directions for future research.

The new (old) inequalities of news exposure

Existing scholarly discourse on incidental exposure predicts that such exposures can mobilize the non-politically interested, inform those who are less knowledgeable, and even increase the diversity of news sources used (Fletcher and Nielsen, 2018; Valeriani and Vaccari, 2016). The attraction metaphor inspires a new frame on this set of questions. First, we should ask, are all social media users equally attractive to news? Existing literature suggests that income and education level are powerfully related to news use in offline contexts, and that political interest as a motivator of news use has been growing over the past two decades (Min, 2010; Strömbäck et al., 2013). If incidental exposure is a leveler, we should expect that these factors are not significant predictors of exposure on social media. However, they are.

Using the 2016 survey of young adults referenced above, we find evidence of inequalities in incidental exposure. Young adults with higher levels of education were more likely to report seeing content about the election on Facebook (odds ratio = 1.38, p = .000) and reported engaging with public affairs news more often on social media (β = .18, p = .000). But why? Through what mechanisms do low socioeconomic status (SES) social media users become less likely to be incidentally exposed to news? One explanation could be that some people’s friends are less likely to post and share news than other people’s friends, an inequality rooted offline that translates into online networks. Indeed, we find that those with higher levels of education reported that their friends posted more about politics more than did the friends of lower education respondents (β = .10, p = .001). Education (but not income) is also related to levels of political interest, such that more educated respondents are higher in political interest (β = .15, p = .000). Education is also related to interest in news and current events (β = .17, p = .000), such that higher SES young adults are more interested in news. In turn, interest in politics predicts news engagement on social media as well as encountering election content on Facebook.

The characteristics social media users bring with them to the platform can shape their future exposure to news (Thorson et al., 2018). It would be difficult to hold a platform company responsible for user preferences socialized outside the platform. However, the attraction metaphor also points us toward analyses of the role that algorithms are playing to either ease or reinforce these inequities. For example, my colleagues and I have been studying interest classification algorithms on Facebook and their relationship to news exposure. We ask participants to download their Facebook interest classification data and provide it to us for analysis, as well as complete a survey (Thorson et al., 2019a). In one study of adults in a mid-sized American city, we analyzed the relationships among participant demographics, self-reported interest, and the way that Facebook algorithmically inferred what each person was interested in (see Appendix 1 and Thorson et al., 2019b). We found inequalities in the rate at which different kinds of people were classified as interested in news and politics. White respondents, more educated respondents, and those with higher levels of self-reported political interest were more likely to be algorithmically classified as interested in politics. Those with higher incomes, more education, and self-reported interest in news were more likely to be algorithmically classified as interested in news.

In another study using the same method (with a different sample), we found that algorithmic classification of interest in news and politics explains exposure to news on Facebook above and beyond self-reported interest (Thorson et al., 2019a). Furthermore, we established that becoming classified as interested in news on Facebook depends not only on your self-reported interest, but more importantly on your behavior on the platform (what pages you like), and on your friends’ behaviors (if your friends post about news and politics, you are more likely to be classified as interested in news).

In these two studies, we see evidence that attractiveness to news can be written into data, and through subsequent analysis and transformations within a platform can affect the visibility of news in ways that are outside of individual control. This is not to say that classification algorithms are intentionally biased against any particular group, but rather that the particular configuration of user practices, data collection, and algorithmic techniques in use at Facebook today serve to reify differences between the news rich and news poor that existed long before digital platforms came to be.

Targeting by news organizations and political actors

A second area for expanded research is to consider the ways in which strategic microtargeting by politicians and news organizations may shape the likelihood of incidental exposure. Incidental exposure to news and political information on social media is not only determined by the organic spread of messages and peer sharing but also by paid sponsorship of content. There is a growing literature on the ways politicians and related groups use targeting on social media (e.g. Kim et al., 2018; Kreiss and McGregor, 2018). Less studied have been the paid targeting choices made by news organizations. News publishers are using paid media to increase visibility of news stories, as well as to promote subscriptions (Mendoza, 2018). Therefore, we need to analyze who they are targeting, and whether these targeting practices affect rates of incidental exposure across different segments of users.

One way to investigate this question is to use data from Facebook’s advertising library, a database that provides public access to advertisements created by pages on the platform (Facebook, n.d.). This database provides access to ads that are being run by publisher pages on the site. However, what we cannot see using these data are the decisions made by these organizations about how to target paid content. Audience targeting choices could work to broaden the audience for news content on social media, for example, if a newspaper chose to pay for a visibility boost to populations underserved by news. On the other hand, news organizations could reinforce existing inequalities by targeting their existing audiences or people who are similar to those audiences – those with ‘high-propensity’ to engage with news. Future studies of news dissemination practices should consider how the choices made by news publishers on social media (what to post, what to pay for, what target audiences to focus on) shape who attracts incidental exposure to news.

Conditional media effects

As noted above, the attraction metaphor creates room to articulate incidental-ness as a matter of degrees: social media users who are strong attractors of news become so at least in part because their prior behavior has indicated an interest in such content. Kümpel has argued we should, therefore, consider the incidental-ness of a news exposure on a continuum – some news exposures are more serendipitous than others – and has proposed that the degree of incidental-ness might shape willingness to engage with news content as well as downstream effects. A useful turn in the literature concerned with effects of incidental exposure has been the articulation of a distinction between incidental contact with news versus actual engagement with a news story, defined as clicking on or reading the story (Kaiser et al., 2018; Karnowski et al., 2017; Kümpel, 2018; Lee and Kim, 2017). We might expect that users for whom exposure to news on social media is highly predictable (that is, those for whom incidental exposure is not accidental at all) will be more motivated to process – and engage with – news content they encounter than would users who have the much rarer experience of a truly serendipitous encounter with news. This in turn should have implications for learning and mobilization effects.

For example, Karnowski et al. (2017) found that interest and prior knowledge of news were more important predictors of news engagement than any social source characteristics of a message. Drawing on interviews, Kümpel (2018) found that users’ decisions to read articles they find on social media depend on their pre-existing interest in the issue. Lee and Kim (2017) demonstrated that learning from social media news exposures was fully mediated by actual news engagement (time spent reading the article encountered). Together, these relationships lead to what Kümpel describes as a Matthew Effect: those who are already interested in news are more likely to be exposed to news and also to engage with news, thus gaining more from each encounter than the less interested.

Future studies should consider building in digital trace measures to survey data collections as a way to account for the individual-level degree of attractiveness for news in their models. For example, as noted above, my colleagues and I (Thorson et al., 2019a) combined Facebook user surveys with Facebook profile data. Such a data collection enables researchers to include behavioral traces of ‘attractiveness to news’ (e.g. number of news pages ‘liked’, number of news posts ‘liked’, and number of algorithmically inferred news interests) in models predicting learning or participation outcomes.

Conclusion

The purpose of this article is to highlight the need to reframe incidental exposure research for the rapidly evolving era of datafication and platformization. On algorithmically curated platforms like Facebook, content selection is shaped not only by user preferences and news gatekeeping processes but also by the dynamics of platforms themselves (Van Dijck et al., 2018). Current configurations of platforms, users, and news organizations appear to be designed to reinforce existing patterns of attractiveness to news, but these configurations are subject to change. At the time of writing, Mark Zuckerberg indicated Facebook will pivot away from its investment in the newsfeed as a central home for visibility on the platform (Statt, 2019, 30 April). The Facebook mobile app is undergoing a redesign that will highlight groups and events rather than the newsfeed exclusively. Each of these changes will affect patterns of news exposure on the platform, highlighting the power of platform dynamics in the broader system of news exposure.

Footnotes

Appendix 1 Acknowledgements

Author would like to thank the special issue editors and to the anonymous referees for their thoughtful feedback. In addition, she would also like to thank many people who took time to discuss these ideas with her, especially Anna Kümpel, Mel Medeiros, and Kelley Cotter.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

Author biography

Kjerstin Thorson is an Associate Professor in the College of Communication Arts & Sciences at Michigan State University. Her research explores how people use digital and social media to learn about and participate in politics, and how social media platforms are reshaping the visibility of news and politics.

References

Ahmadi

Wohn

(2018) The antecedents of incidental news exposure on social media. Social Media+ Society 4(2): 2056305118772827.

Ananny

(2016) Toward an ethics of algorithms: Convening, observation, probability, and timeliness. Science, Technology, & Human Values 41(1): 93–117.

Baum

(2002) Sex, lies, and war: How soft news brings foreign policy to the inattentive public. American Political Science Review 96(1): 91–109.

Bell

Owen

Brown

, et al. (2017) The platform press: How Silicon Valley reengineered journalism. Available at: https://academiccommons.columbia.edu/doi/10.7916/D8R216ZZ

Boczkowski

Mitchelstein

Matassi

(2018) News comes across when I’m in a moment of leisure: Understanding the practices of incidental news consumption on social media. New Media & Society 20(10): 3523–3539.

Bode

(2016) Political news in the news feed: Learning politics from social media. Mass Communication and Society 19(1): 24–48.

Bucher

(2012) Want to be on the top? Algorithmic power and the threat of invisibility on Facebook. New Media & Society 14(7): 1164–1180.

DeVito

(2017) From editors to algorithms: A values-based approach to understanding story selection in the Facebook news feed. Digital Journalism 5(6): 753–773.

Downs

(1957) An Economic Theory of Democracy. New York: Harper.

10.

Facebook (n.d.) Available at: https://www.facebook.com/ads/library/

11.

Feezell

(2018) Agenda setting through social media: The importance of incidental news exposure and social filtering in the digital era. Political Research Quarterly 71(2): 482–494.

12.

Fletcher

Nielsen

(2018) Are people incidentally exposed to news on social media? A comparative analysis. New Media & Society 20(7): 2450–2468.

13.

Gil de Zúñiga

Weeks

Ardèvol-Abreu

(2017) Effects of the news-finds-me perception in communication: Social media use implications for news seeking and learning about politics. Journal of Computer-mediated Communication 22(3): 105–123.

14.

Haenschen

(2019) Self-reported versus digitally recorded: Measuring political activity on Facebook. Social Science Computer Review. Epub ahead of print 16 January. DOI: 10.1177/0894439318813586.

15.

Kaiser

Keller

Kleinen-von Königslöw

(2018) Incidental news exposure on Facebook as a social experience: The influence of recommender and media cues on news selection. Communication Research. Epub ahead of print 23 October. DOI: 0.1177/0093650218803529.

16.

Kantrowitz

(2018) 18 February. Available at: https://www.buzzfeednews.com/article/alexkantrowitz/facebook-wont-release-its-5-of-news-feed-content-is-news

17.

Karnowski

Kümpel

Leonhard

, et al. (2017) From incidental news exposure to news engagement. How perceptions of the news post and news usage patterns influence engagement with news articles encountered on Facebook. Computers in Human Behavior 76: 42–50.

18.

Kim

Chen

De Zúñiga

(2013) Stumbling upon news on the Internet: Effects of incidental news exposure and relative entertainment use on political engagement. Computers in Human Behavior 29(6): 2607–2614.

19.

Kim

Hsu

Neiman

, et al. (2018) The stealth media? Groups and targets behind divisive issue campaigns on Facebook. Political Communication 35(4): 515–541.

20.

Klinger

Svensson

(2015) The emergence of network media logic in political communication: A theoretical approach. New Media & Society 17(8): 1241–1257.

21.

Kreiss

McGregor

(2018) Technology firms shape political communication: The work of Microsoft, Facebook, Twitter, and Google with campaigns during the 2016 US presidential cycle. Political Communication 35(2): 155–177.

22.

Krippendorff

(1993) Major metaphors of communication and some constructivist reflections on their use. Cybernetics & Human Knowing 2(1): 3.

23.

Kümpel

(2018) The issue takes it all? Incidental news exposure and news engagement on Facebook. Digital Journalism 7: 165–186.

24.

Kümpel

(this issue) The Matthew Effect in social media news use: Assessing inequalities in news exposure and news engagement on social network sites (SNS). Journalism.

25.

Lee

Tandoc

Jr (2017) When news meets the audience: How audience feedback online affects news production and consumption. Human Communication Research 43(4): 436–449.

26.

Lee

Kim

(2017) Incidental exposure to news: Predictors in the social media setting and effects on information gain online. Computers in Human Behavior 75: 1008–1015.

27.

Meadows

(2008) Thinking in Systems: A Primer. Hartford, VT: Chelsea Green Publishing.

28.

Mendoza

(2018) Inside the local news subscriptions accelerator: Publishers share their results. Available at: https://facebookjournalismproject.com/article/inside-the-local-news-subscriptions-accelerator-publishers-share-their-results/

29.

Min

(2010) From the digital divide to the democratic divide: Internet skills, political interest, and the second-level digital divide in political internet use. Journal of Information Technology & Politics 7(1): 22–35.

30.

Möller

van de Velde

Merten

, et al. (2019) Explaining online news engagement based on browsing behavior: Creatures of habit? Social Science Computer Review. Epub ahead of print 10 February. DOI: 10.1177/0894439319828012.

31.

Müller

Schneiders

Schäfer

(2016) Appetizer or main dish? Explaining the use of Facebook news posts as a substitute for other news sources. Computers in Human Behavior 65: 431–441.

32.

Napoli

(2014) Automated media: An institutional theory perspective on algorithmic media production and consumption. Communication Theory 24(3): 340–360.

33.

Nielsen

Ganter

(2018) Dealing with digital intermediaries: A case study of the relations between publishers and platforms. New Media & Society 20(4): 1600–1617.

34.

Oeldorf-Hirsch

(2018) The role of engagement in learning from active and incidental news exposure on social media. Mass Communication and Society 21(2): 225–247.

35.

Oxford Dictionaries (n.d.) Incidental. Available at: https://en.oxforddictionaries.com/definition/incidental

36.

Pew Research Center (n.d.) Available at: https://www.journalism.org/2018/09/10/news-use-across-social-media-platforms-2018/

37.

Prior

(2007) Post-Broadcast Democracy: How Media Choice Increases Inequality in Political Involvement and Polarizes Elections. Cambridge: Cambridge University Press.

38.

Srnicek

(2017) Platform Capitalism. Polity. Cambridge, UK.

39.

Statt

(2019) Available at: https://www.theverge.com/2019/4/30/18523265/facebook-events-groups-redesign-news-feed-features-f8-2019

40.

Strömbäck

Djerf-Pierre

Shehata

(2013) The dynamics of political interest and news media consumption: A longitudinal perspective. International Journal of Public Opinion Research 25(4): 414–435.

41.

Tewksbury

Weaver

Maddex

(2001) Accidentally informed: Incidental news exposure on the World Wide Web. Journalism & Mass Communication Quarterly 78(3): 533–554.

42.

Thorson

Wells

(2016) Curated flows: A framework for mapping media exposure in the digital age. Communication Theory 26(3): 309–328.

43.

Thorson

Cotter

Medeiros

, et al. (2019a) Algorithmic inference, political interest, and exposure to news and politics on Facebook. Information, Communication & Society. Epub ahead of print 1 January. DOI: 10.1080/1369118X.2019.1642934.

44.

Thorson

Medeiros

Cotter

, et al. (2019b) Platforms, algorithms, and unequal exposure to news. Unpublished Manuscript.

45.

Thorson

Edgerly

(2018) Political inequalities start at home: Parents, children, and the socialization of civic infrastructure online. Political Communication 35(2): 178–195.

46.

Valeriani

Vaccari

(2016) Accidental exposure to politics on social media as online participation equalizer in Germany, Italy, and the United Kingdom. New Media & Society 18(9): 1857–1874.

47.

Van Aelst

Strömbäck

Aalberg

, et al. (2017) Political communication in a high-choice media environment: A challenge for democracy? Annals of the International Communication Association 41(1): 3–27.

48.

Van Dijck

(2014) Datafication, dataism and dataveillance: Big Data between scientific paradigm and ideology. Surveillance & Society 12(2): 197–208.

49.

Van Dijck

Poell

De Waal

(2018) The Platform Society: Public Values in a Connective World. Oxford: Oxford University Press.

50.

Webster

(2011) The duality of media: A structurational theory of public attention. Communication Theory 21(1): 43–66.

51.

Welbers

Van Atteveldt

Kleinnijenhuis

, et al. (2016) News selection criteria in the digital age: Professional norms versus online audience metrics. Journalism 17(8): 1037–1053.

52.

Wells

Thorson

(2017) Combining big data and survey techniques to model effects of political content flows in Facebook. Social Science Computer Review 35(1): 33–52.

53.

Yadamsuren

Erdelez

(2010) Incidental exposure to online news. In: Proceedings of the 73rd ASIS&T annual meeting on navigating streams in an information ecosystem, Pittsburgh, PA, 23–28 October, vol. 47, p. 22. Silver Spring, MD: American Society for Information Science.

54.

Yurieff

Kelley

(2019) 10 April. Available at: https://www.cnn.com/2019/04/10/tech/facebook-integrity-updates/index.html