ArticlePDF Available

The Small, Disloyal Fake News Audience: The Role of Audience Availability in Fake News Consumption

Article

The Small, Disloyal Fake News Audience: The Role of Audience Availability in Fake News Consumption

Abstract and Figures

In light of the recent U.S. election, many fear that "fake news" has become a force of enormous reach and influence within the news media environment. We draw on well-established theories of audience behavior to argue that the online fake news audience, like most niche content, would be a small subset of the total news audience, especially those with high availability. By examining online visitation data across mobile and desktop platforms in the months leading up to and following the 2016 presidential election, we indeed find the fake news audience comprises a small, disloyal group of heavy internet users. We also find that social network sites play an outsized role in generating traffic to fake news. With this revised understanding, we revisit the democratic implications of the fake news crisis.
Content may be subject to copyright.
https://doi.org/10.1177/1461444818758715
new media & society
1 –18
© The Author(s) 2018
Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/1461444818758715
journals.sagepub.com/home/nms
The small, disloyal fake
news audience: The role of
audience availability in fake
news consumption
Jacob L Nelson
Northwestern University, USA
Harsh Taneja
University of Illinois Urbana-Champaign, USA
Abstract
In light of the recent US election, many fear that “fake news” has become a force of enormous
reach and influence within the news media environment. We draw on well-established
theories of audience behavior to argue that the online fake news audience, like most niche
content, would be a small subset of the total news audience, especially those with high
availability. By examining online visitation data across mobile and desktop platforms in the
months leading up to and following the 2016 presidential election, we indeed find the fake
news audience comprises a small, disloyal group of heavy Internet users. We also find that
social network sites play an outsized role in generating traffic to fake news. With this revised
understanding, we revisit the democratic implications of the fake news crisis.
Keywords
2016 Elections, Fake News, News Audience, Audience Availability, Double Jeopardy,
Social Media, Audience Fragmentation, Elections, Mobile Internet
Concerns surrounding “fake news” have spread since the most recent US presidential
election. Many believe that “fake news” has become a powerful and sinister force in the
online news media environment, with dire consequences for democracy (Glaser, 2017;
Corresponding author:
Jacob L Nelson, Northwestern University, 2240 Campus Drive, Evanston, IL 60208, USA.
Email: jacobnelson4@gmail.com
758715NMS0010.1177/1461444818758715new media & societyNelson and Taneja
research-article2018
Article
2 new media & society 00(0)
Zengerle, 2016). As a result, news organizations and tech companies have taken steps to
stifle fake news production and dissemination (Owen, 2016). These efforts, and the dis-
course surrounding them, assume that fake news reaches a broad, susceptible audience,
who do not compare these stories against those from other sources. In this article, we
argue against this conclusion and substantiate this argument with an empirical analysis
of online audience data. Instead, we find that the fake news audience comprises a small
number of heavy Internet users, while the majority of news consumers continue to stick
to the most well-known of news brands.
These findings are consistent with prior media audience research. First, audiences
tend to disproportionately gravitate toward well-known, popular media outlets rather
than less established, niche offerings (Gentzkow and Shapiro, 2011; Prior, 2013; Taneja
et al., 2017; Webster et al., 2014). Second, audiences that do venture into less familiar
brands are typically heavy media users in general (Elberse, 2008). The term “fake news”
has become ubiquitous with remarkable speed, yet specific fake news outlets remain
relatively obscure compared with media institutions like CNN, The New York Times, and
Fox News. In short, there is a theoretical basis for why fake news sites would attract a
small portion of the overall news media audience, and why the people exposed to fake
news might actually spend more time-consuming content of all kinds than anyone else.
What follows is an analysis of online audience data that examines fake news con-
sumption patterns leading up to and following the November 2016 election. Contrary
to the discourse surrounding fake news, we find that only a small portion of the over-
all Internet audience visited fake news sites in 2016. We also find that visits to fake
news sites originated from social network sites (SNSs) at a much higher rate than
visits to real news sites, confirming the primary role social media played in spreading
fake news. We conclude with calls for future research into this unfolding fake news
crisis. Although its audience is small, its consequences for journalism and democracy
remain unclear.
The fake news crisis
As recently as 2012, scholars used “fake news” as a term that referred to “The Daily
Show” and “The Colbert Report”—late night television shows that blurred the line
between news and comedy (Borden and Tew, 2007; Day and Thompson, 2012). Now, the
term more commonly refers to false or misleading information made to look like a fact-
based news story in order to “influence public opinion or cull digital advertising dollars”
(Uberti, 2017). The sudden shift in the term’s meaning stems from a confluence of events
leading up to the election of Donald Trump as US president: The increasingly central role
of SNSs for news consumption (Gottfried and Shearer, 2016), the Russian propaganda
effort to produce and spread fake news stories during the 2016 presidential campaign
(Timberg, 2016), and the habit of some political elites to legitimize fake news stories by
passing them along to their followers using social media platforms like Facebook and
Twitter (Flood, 2016). Taken together, these circumstances have contributed to a media
environment where sensational headlines are easier to find than they are to verify. As a
result, two-in-three US adults now believe fake news causes confusion about current
events and issues, as well as basic facts (Barthel et al., 2016).
Nelson and Taneja 3
Yet, the few empirical investigations into fake news consumption that exist suggest
the fake news audience is actually small compared with the real news audience (Allcott
and Gentzkow, 2017; Guess, Nyhan, and Reifler, 2018; Nelson, 2017). In their analysis
of on online audience data from late October through late November of 2016, Allcott and
Gentzkow (2017) concluded “even the most widely circulated fake news stories were
seen by only a small fraction of Americans” (p. 21). Our analysis both supports and
expands on Allcott and Gentzkow’s, by looking at a year’s worth of real and fake news
consumption and by including data drawn from mobile browsers in addition to desktop.
While these findings may seem surprising to those caught up in the current discourse
surrounding fake news, they confirm some well-known theories of audience behavior,
which we review next.
Audience availability and the Law of Double Jeopardy
When it comes to understanding people’s motivations for media consumption, one often-
overlooked aspect is audience availability: the amount of time the audience has for using
the medium in the first place. For instance, audiences tend to choose television programs
based on their availability rather than their preferences. Even when program schedules
change, the specific time audiences set aside to watch television does not (Taneja and
Viswanathan, 2014; Webster and Wakshlag, 1983). As Webster (2014) explains,
“Audience behavior is often a two-stage process in which a decision to use a medium
precedes the selection of specific content” (p. 94). Even in the current media environ-
ment, where audiences have more autonomy and seemingly limitless options, they are
still bound by the fixed amount of time they have for media use. So, their availability to
take advantage of media tends to have a moderating effect on their individual character-
istics, such as media preferences (Taneja and Viswanathan, 2014).
Reckoning with audience availability behavior often complicates our understanding
of media choice. For example, many believe that as audiences age, they watch more
television news and less entertainment; however, older audiences watch more television
than younger audiences overall, so their time spent with both news and entertainment is
higher. At similar levels of availability, younger viewers watch as much news as older
viewers (Taneja and Viswanathan, 2014). These findings suggest that audiences who
have higher availability for media are more likely to have a more varied media diet, a
finding confirmed in audience studies across different media platforms (Taneja and
Viswanathan, 2014; Taneja et al., 2012). Audiences with more availability—heavy media
users—have more time to spend exploring different media offerings. On the other hand,
those with less audience availability (i.e. light media users) use their more limited time
to congregate within the most popular offerings (Elberse, 2008; Nelson and Webster,
2016; Webster et al., 2014). This often leads to a “winner-take-all” effect where the most
popular media products attract the most loyal audiences—a phenomenon referred to as
the Law of Double Jeopardy (Webster et al., 2014).
The Law of Double Jeopardy posits that popular offerings will enjoy greater loyalty
than unpopular offerings (Ehrenberg et al., 1990; Martin, 1973). This occurs because
light users tend to only be aware of the most popular products, so they do not know what
they are missing (McPhee, 1963). And while heavy users are aware of both popular and
4 new media & society 00(0)
unpopular offerings, they comprise a smaller portion of the overall audience. These
heavy users will also take advantage of the popular as well as the unpopular rather than
forgoing the former for the latter. This explains why “80/20” distributions—when 20%
of the outlets account for 80% of the audiences—are so common when it comes to media
consumption. The biggest name products tend to “naturally monopolize” light users,
while the more obscure offerings attract heavy users that have the time to invest in a
variety of offerings, and the awareness of their existence to begin with (Elberse, 2008:
93). Accordingly, studies have suggested that most Internet users stick to popular news
sources, regardless of ideology, while a minority of omnivorous heavy users consume
ideologically extreme sources in addition to popular content (Gentzkow and Shapiro,
2011; Guess, 2016).
Online news usage and the fake news audience
Beyond audience availability and double jeopardy effects, there are other characteristics
of the online media system that increase the likelihood that audiences will be more
exposed to more popular content. Although no one outside of Facebook and Google
know exactly what goes into the algorithms that determine their link placements, it is
understood that content popularity plays an important role. Studies have repeatedly
shown that algorithmic recommendations and social information tend to enhance ine-
quality by increasing exposure to what is already popular (Bucher, 2017; Carlson, 2017;
Napoli, 2014; Salganik et al., 2006). As a result, a headline from a brand name in news
media that already has millions of followers (e.g. CNN) is likely to get pushed out to
audiences more often and more forcefully than a headline from a more obscure news site.
While the term “fake news” now seems inescapable, even the most prominent fake news
outlets (e.g. InfoWars, 70 News) are much less recognizable than media institutions like
The New York Times and The Wall Street Journal. Hence, although fake news stories
could certainly “go viral” and gain a lot of traction, they would need to overcome the
institutional forces working to marginalize them and favor stories from mainstream,
established outlets.
Even in a media environment where news consumption has grown increasingly
“incidental” and mediated by social media platforms (Fletcher and Nielsen, 2017,),
those with more availability will still likely be exposed to news from a wider variety
of sources, while those with less availability will likely continue to consume news
from the most popular sources. A recent field experiment of partisan news consump-
tion (arguably the closest parallel to fake news) revealed that when both liberal and
conservative news users were presented with a novel political issue within the “choice
architecture of passive Internet use” (i.e. the collection of recommendations through
search engines, trends, and social media cues), they did not change their existing news
usage patterns (Guess, 2016: 7). Light users may not “see” news from newer sources
because they have customized their news feeds to show less news and/or because the
algorithms underlying the news feeds are responsive to past behavior (Fletcher and
Nielsen, 2017). Heavy social media users, however, would be more likely to see
obscure news sources regularly pop up in their feeds. In other words, incidental news
usage may drive social media users toward fake news sources proportional to their
availability.
Nelson and Taneja 5
What follows is a test to see if this relationship exists. Our analysis examines whether
or not a majority of online news audiences congregate among the most popular news
sources, while those that get exposed to news from niche outlets—including fake news—
comprise a much smaller group of the Internet’s heaviest users. Our research questions
are as follows:
RQ1. What is the size and level of engagement of the fake news audience compared
with the real news audience?
RQ2. How does the overall online usage of the fake news audience differ from the real
news audience?
Recognizing that news usage peaks in months leading up to a general election, we pay
attention to whether or not the patterns we observe in RQs 1 and 2 are different before
and after the November 2016 elections. Furthermore, given the large amount of time
people spend browsing social media sites on mobile devices (Sonderman, 2014), the
formers’ purported linkages with fake news’ spread (Buchanan, 2017) and mobile’s
growing role as the platform of choice for news consumption (Nelson and Lei, 2017), we
also separately analyze both RQ1 and RQ2 for web browsing through mobile and desk-
top platforms.
Method
Our data come from the web analytic firm comScore. Each month, comScore (2013) uses
weights to estimate the online behavior of the total US online audience, based on data
from its panel of 1 million users. These data have been used in past analyses of digital
audience behavior (e.g. Gentzkow and Shapiro, 2011; Nelson and Webster, 2017). com-
Score data can be examined by platform (e.g. mobile or desktop specific) or in “multi-
platform” form, which combines and deduplicates its mobile and desktop users so that a
person who, for instance, logs onto a website via both an iPhone and a home computer is
only counted once. We used a combination of the three options throughout our analysis.
In general, our findings were consistent across platforms.
Defining fake news
We compared 30 fake news sites with 24 real news sites. The real news sites included a
mix of newspapers, broadcast, and digital-first publishers, all of which are established
brands within the news media sphere (e.g. Yahoo-ABC News, CNN, The New York
Times, The Washington Post, Fox News, and BuzzFeed). These sites were chosen
because, similar to fake news sites, they all maintain a national focus when it comes to
political news coverage as opposed to any sort of local targeting. Furthermore, they run
the gamut in terms of audience size when it comes to the overall Internet audience. For
example, The New York Times reached about 20% of the overall Internet audience in
November 2016, while C-SPAN reached only about 0.5%. Finally, these sites are fre-
quently included in studies of the news media environment. As a result, while this sample
6 new media & society 00(0)
is certainly not representative of all real news sites in the United States, we believe it
includes many of the nation’s most recognizable names in political news.
The fake news sites were more difficult to determine because “fake news” within the
context of the 2016 election lacks a clear and concise definition (Tandoc et al., 2017a).
For the purpose of this study, we wanted to exclude sites that publish satirical news (e.g.
The Onion) and focus solely on sites that literally fabricate the news. We referred to a list
of fake news outlets compiled by OpenSources, a research team led by media professor
Melissa Zimdars. OpenSources has a six-step process for analyzing news websites that
includes examining the site’s domain name, its “About Us” section, the sources its sto-
ries draw on, the writing style, the site’s aesthetic, and its social media presence.
OpenSources classifies the sites on this list using a series of tags like “fake,” “satire,”
“hate,” and “clickbait.” If a site provides accurate news with a distinct political slant, for
example, it is tagged as “political.” If a site promotes conspiracy theories (e.g. the Sandy
Hook massacre was staged, intelligence community leaks are the result of a US “Deep
State”) it is tagged with “conspiracy.” And if a source “entirely fabricates information,”
it is tagged as “fake.” The OpenSources team often used multiple tags for individual
sites; for example, Drudge Report was tagged as “political” and “bias,” while virallib-
erty.com was tagged as “fake,” “bias,” and “clickbait.”
The OpenSources list is constantly being updated but currently includes 834 sites. For
the purpose of this study, we focused solely on sites that met the most literal definition of
“fake news,” meaning they completely fabricate news stories. To do so, we only included
the 126 sites OpenSources had tagged as “fake” when we performed our analysis.
However, comScore requires at least 30 visitors from its panel to visit a site before it is
included in its panel. Of those 126, only 30 met that threshold in October of 2016. As a
result, the total number of sites in the sample was 54. One consequence of our sample
selection process is that it relied on a broad conceptualization of “real news” (e.g. both
C-SPAN and BuzzFeed) and a narrow conceptualization of “fake news” (i.e. sites that
are very literally fabricating the news).
Analysis
To address RQ1, we used comScore’s multi-platform data to examine visitation (meas-
ured by unique visitors) and engagement (measured by average minutes per visitor) for
real and fake news sites over time, in order to observe any changes in audience behavior
in the months leading up to and immediately following the presidential election. We did
this by comparing the monthly average visitors to real news sites with the monthly aver-
age visitors to fake news site between January of 2016 and January of 2017. We then
compared the monthly average time spent with real news sites with the monthly average
engagement time spent with fake news sites between January 2016 and January 2017.
Then, focusing exclusively on October 2016 desktop data,1 we ran a series of Point-
Biserial Correlation analyses to observe if a news site being fake was significantly related
to its audience size (measured by unique visitors), its audience engagement (measured by
average minutes per visitor), and the frequency by which its visits originate from
Facebook (measured by the percentage of visits coming from Facebook) as compared
with visits that occur immediately when the web browser is opened (measured by the
Nelson and Taneja 7
percentage of visits coming from what comScore labels “Logon”). Point-Biserial
Correlations follow the same assumptions as Pearson correlations and are used to test
associations between continuous and dichotomous variables. In this instance, the dichot-
omous variable was whether or not a news site is fake, while the continuous variables
were the audience size, engagement, and rate of visitors originating from both Facebook
and Logon.
We then looked at cross visitation patterns to see what percentage of visitors to fake
news sites also visited real news sites in October 2016. Doing so allowed us to under-
stand if the fake news audience exists in a “filter bubble” or if they have a more varied
news diet.
Audience duplication
We faced a limitation in our attempt to identify whether the fake news audience com-
prised heavy or light Internet users: comScore, like most audience measurement provid-
ers, does not report data at the individual user level (see Taneja, 2016). Therefore, we
were unable to directly segment users based on usage into heavy, medium, and light
users. Instead, we developed a method where we could use the audience duplication
among the sites in our sample to infer how the fake news audience as a whole behaved
relative to the real news audience. Audience duplication is a comScore tool that shows
what percentage of the total Internet audience overlapped in their visits to the sites in our
sample within our period of analysis. Audience duplication indicates the number of
unique visitors who overlapped on two or more different sites, the total amount of time
they spent on those sites (measured in average minutes per visitor), and the percentage of
the total Internet audience they comprised. This metric has been used in previous studies
to explore audience behavior both for television and web audiences (e.g. Taneja and
Webster, 2016).
To use this tool, we first separated the fake and real news sites into distinct groups. We
took Facebook and Google.com separately as proxies for overall Internet use, since these
sites cumulatively include almost all Internet users in the United States. For instance, in
October 2016, 90% of the total US online population visited Google.com, while 57%
visited Facebook, a site that is increasingly seen as the gateway to news consumption
(Barthel et al., 2015). Through a related analysis, we also confirmed that a disproportion-
ately high number of visits to fake news sites were from Facebook. Finally, since the
group of Google.com visitors is much larger than the group of Facebook visitors, it
includes most Internet users who do not use social media.
Next, we isolated the total unique visitors who visited Facebook and at least one other
fake news site and noted their average time spent online. We replicated this calculation
for Facebook and the group of real news sites. We repeated this analysis for each month
between January 2016 and January 2017. We performed the same procedure for each of
the 13 months under analysis with Google instead of Facebook, meaning we estimated
the average time spent by people who visited at least one real news site and Google or at
least one fake news site and Google. We then performed a similar analysis without
Facebook or Google to capture the average time spent on all fake and real news sites,
respectively, for the total audience who accessed at least either one fake news site
8 new media & society 00(0)
(henceforth “Fake News Users”) or one real news site (henceforth “Real News Users”).
The difference between the two time-spent measures indicates the extent to which each
group of users spends time on Facebook or Google and also corroborates that their usage
indeed captures variance when it comes to Internet use.
Since people access social networking sites and the wider web heavily from mobile
platforms, we followed these steps using both comScore’s desktop and mobile measure-
ment tools separately to see if there were any differences depending on the digital plat-
form people use for Internet access. In short, the audience duplication tool allowed us to
observe whether or not those who spend more time with both Facebook and Google (and,
by consequence, the Internet), thus demonstrating more availability for a medium, are
more likely to access fake news.
Results
Figure 1 reveals that the number of multi-platform monthly visitors (meaning online
visitors using either mobile or desktop devices or both) to an average real news site was
more than 40 times larger than the number of monthly visitors to an average fake news
sites throughout 2016. The audience size for an average real news sites was about 28 mil-
lion unique visitors, while the audience size for an average fake news site was only about
675,000. Figure 2, which shows the audience sizes for individual news sites, reveals that
the real news site audiences dwarfed fake news site audiences in October 2016, just
before the election. Even when the real news audience dropped suddenly, immediately
following the election, it still boasted millions more unique visitors than the fake news
audience. Furthermore, the drop seems to have only lasted a month, as it grew nearly
back to its November peak in January 2017.
Online audiences also tended to spend more time with real news than with fake news.
As Figure 1 shows, online audiences spent about 9 minutes per month with an average
real news site, while they only spent about half of that with an average fake news site.
Figure 2 shows that this trend stayed constant the month before the election. It is interest-
ing to note that engagement with real news increased following the election, while
engagement with fake news decreased.
Figure 3 reveals the total unique (unduplicated) audiences of the real and fake news
sites in our sample in October 2016. Not only was the individual reach of each fake news
site small, but this group of fake news sites collectively also reached a much smaller
audience than more than half of the real news sites we examined reached individually.
Furthermore, Figure 3 reveals that time spent with the fake news sites in our sample was
lower than time spent with the real news sites in our sample.
We then performed Point-Biserial Correlation analyses to see if a news site being
real or fake was significantly related to its audience size, engagement, and the source of
its traffic. The results, shown in Table 1, reveal a strong relationship between a news site
being fake and it having both a smaller audience and lower levels of engagement. The
results also reveal that visitors are significantly more likely to navigate to fake news
sites from the SNS Facebook. Although fake news sites have a positive relationship
with visits originating from both Facebook and Logon, the coefficient for the former is
not only larger but more significant as well. In other words, visits to fake news sites are
Nelson and Taneja 9
slightly more likely to stem from audiences first logging on to web browsers than visits
to real news sites; however, this effect is much more pronounced when it comes to visits
that originate from Facebook. For example, more than 50% of the visits to
Conservativedailypost.com originated from Facebook, while the same was only true of
about 10% of the visits to The New York Times.
We also examined the incidence of the fake news audience visiting real news sites. We
did this by observing what percentage of desktop and mobile visitors to each fake news
site in our sample also visited the 10 most popular real news sites in our sample in
October 2016. We then took the average cross visitation numbers between each fake
news site and the given real news site. Figure 4 shows quite a bit of overlap between the
real and fake news audiences. On average, for example, more than half of the visitors to
fake news sites in the sample also visited Yahoo-ABC. Furthermore, the percentage of
the fake news audience that visited real news sites was very strongly correlated (.94)
Figure 1. (a) Multi-platform audience size of real and fake news sites from January 2016;
through January 2017 (as measured by unique visitors, in thousands) and (b) Multi-platform
engagement with real and fake news from January 2016 to January 2017 (as measured by
average minutes per visitor).
10 new media & society 00(0)
with the overall popularity of the real news site. So, while only between about 20% and
30% of the fake news audience also visited sites like Fox and Politico, half or more vis-
ited The New York Times and Yahoo-ABC News. This suggests that rather than confined
to an echo chamber, the fake news audience also exposes itself to news content that a vast
majority of the online audience also consumes.
Next, we wanted to learn about the general online habits of the fake news audience.
Are they heavy Internet users in general, as the Law of Double Jeopardy and the audience
availability effect would posit? We investigated this using comScore’s audience duplica-
tion tool, which allowed us to compare the audience size and average time spent online for
the Fake News Users with the same metrics for the Real News Users, using time spent on
Facebook and Google, respectively, as proxies for audience availability. As already noted,
comparing the time spent on Facebook by real and fake news users, respectively, helps us
evaluate whether availability matters in incidental exposure to fake news versus real
news. And doing so for Google.com helps us evaluate if these availability effects, as seen
in the aggregate, matter when we also include users who may not be using social media.
Figure 5 demonstrate these effects using time spent on Google.com and Facebook.
com, respectively, for Real News Users and Fake News Users as proxies for audience
availability on both mobile and desktop platforms. In Figure 5, the series “mobile (Fake
News Users)” indicates the average monthly time spent on Google.com when surfing
Figure 2. (a) Multi-platform fake and real news audience size in October 2016 (in unique
visitors [in thousands]) and (b)Multi-platform fake and real news audience time spent in
October 2016 (in average minutes per visitor).
Real news sites are in red and fake news sites are in blue.
Nelson and Taneja 11
through mobile devices by users who accessed at least one Fake News Site in our sample
during the month. Likewise, the series “mobile (Real News Users)” reports the average
monthly time spent by users on Google.com who accessed at least one real news site in
our sample. The remaining two series show the time spent on Google.com among Fake
Figure 3. (a) Total fake and real news audience size, in unique visitors (in thousands), by
platform, in October 2016 and (b) Total fake and real news audience engagement, in average
minutes per visitor, by platform, in October 2016.
Table 1. Point-biserial correlation results.
Unique visitors Average minutes
per visitor
% of visits from
Facebook
% of visits from
Logon
Fake news –0.5445*** (–4.689) –0.367** (–2.837) 0.6184*** (5.675) 0.3816* (2.4237)
df 52 52 52 52
N54 54 54 54
df: degrees of freedom.
t-values in parentheses; *p < 0.05, **p < 0.01, ***p < 0.001.
12 new media & society 00(0)
News Users and Real News Users when using desktop platforms. Figure 5 replicates simi-
lar data but indicates the average time spent on Facebook.com (instead of Google.com) by
both Fake News Users and Real News Users for both mobile and desktop Internet access.
Based on the data in Figure 5, we find that the average time spent per month on
Facebook by those who accessed fake news was 1317 minutes (standard deviation
[SD] = 107) on desktop and 1359 minutes (SD = 289) on mobile compared with 358 min-
utes (SD = 34) on desktop and 575 minutes (SD = 75) on mobile for the real news audi-
ence. The time spent on Google was an average of 273 minutes (SD = 30) on desktop and
174 minutes (SD = 36) on mobile for fake news users as compared with 147 minutes
(SD = 17) on desktop and 35 minutes (SD = 10) on mobile for real news users.
These data clearly show that within both desktop and mobile platforms, Fake News
Users comprised heavy Facebook and Google users who, on average, spent at least dou-
ble the amount of time on either of these websites via both desktop and mobile devices
as did Real News Users. Thus, we can see that fake news usage is well correlated with
higher audience availability on both Facebook and Google compared with real news
usage.2 We also find that the time spent by real news users on Google and Facebook is
either at or below the average time spent on these platforms across all users (not just
news users), while time spent by fake news users is much higher than the average.3 Since
each of heavy Facebook users and heavy Google users would, on average, spend more
time on the Internet than lighter users of these services, one could speculate that Fake
News Users would be heavier Internet users than Real News Users.
Discussion
Journalists, scholars, and policymakers alike are currently very worried about the adverse
effects of fake news. These concerns stem from the belief that audiences who stumble
across fake news are not also exposed to real news that would push them to question
Figure 4. Percentage of desktop and mobile visitors to fake news sites who also visited real
news sitesa in October 2016.
aReal news sites ranked in ascending order by unique visitors per month.
Nelson and Taneja 13
what is true and what is false. However, prior audience behavior scholarship offers rea-
son to believe a counterfactual to this narrative: in this scenario, light Internet users stick
mostly to establishment sources for their online news consumption, while heavy users
venture out into the long tail of available news media, thus finding themselves exposed
to niche offerings like fake news. This article provides evidence that this may indeed be
the case.
Our findings indicate that the fake news audience is small and comprises a subset of
the Internet’s heaviest users, while the real news audience commands a majority of the
total Internet audience. Not only are established news brands’ audiences significantly
larger, but they are more loyal as well. Fake news sites, on the other hand, suffer the
downside of the Law of Double Jeopardy—small, disloyal audiences. These findings are
consistent with prior research that concludes that light users tend to stick to the most
popular content, while heavy users consume both the popular content while also ventur-
ing out to more obscure fare (Elberse, 2008).
Our analysis of Facebook use also shows that audience availability plays an important
role in determining fake news exposure even when news consumption itself is incidental.
The use of these platforms alone does not lead to the consumption of fake news. Rather,
it is the amount of time a user spends with Facebook that is likely correlated with levels
Figure 5. (a) Average time spent (minutes) on Google.com by users who visited at least
one real news or at least one fake news website on mobile and desktop, respectively, and (b)
Average time spent (minutes) on Facebook.com by users who visited at least one real news or
at least one fake news website on mobile and desktop, respectively.
14 new media & society 00(0)
of fake news use. Conversely, we find that the time Real News Users spent on Facebook
is only a little higher than the average time spent on the platform by all Facebook users.
At least based on our findings, even if light users are getting exposed to fake news on
their Facebook feeds, they are less likely to click on them. Thus, our findings conform to
the idea that algorithmic intervention generally does not disrupt established usage pat-
terns. In light of this theoretical basis, we believe that fake news is likely to remain a
small audience phenomenon, especially considering that even during an election cycle,
when news appetite is highest (Mitchelstein and Boczkowski, 2010), we saw that the
audience for fake news was still both small and disloyal.
Our findings both corroborate and complicate our current understanding of fake news
consumption. For instance, they suggest that consistent with the current discourse sur-
rounding the role of social media in the spread of fake news, a heavy Facebook user is
far more likely to encounter fake news than a lighter Facebook user. Yet, the fake news
audience itself remains quite small and non-exclusive, especially compared with the real
news audience. Considering this small subset of heavy Facebook users is exposed to real
news as well as fake news, it is worth wondering why they continue to consume both.
Shouldn’t the former persuade these users to discount the latter? This question gets at a
larger limitation of our study, which is that we cannot say with any certainty whether or
not heavy Internet users are savvier online news consumers. One might think that a per-
son who spends more time with news overall would be have a stronger news literacy, but
in a media environment so saturated with different, often opposing declarations of fact,
that may no longer be the case.
We do not mean for our findings to suggest that fake news is not disconcerting. It is
disturbing that fake news exists regardless of the size of its audience. Compared with real
news production, which requires resources like journalists, editors, willing sources, and
time, fake news production requires little more than a creative person with an Internet
connection. As a result, it is much easier for fake news producers to flood the news media
environment with fake news than for real news producers to counteract them. This sug-
gests that even if individual fake news sites continue to draw small audiences, fake news
in general could have a corrosive effect on the news media environment, by making it
more difficult for audiences to determine what is real and what is not.
In a way, it is even more disturbing that it is the heavy Internet users who appear to be
the ones consuming the bulk of fake news, just to satiate their insatiable appetite for jour-
nalism. It is one thing to venture out into the long tail when it comes to creative forms of
media—catching an episode of a niche show is unlikely to affect your opinion of “This Is
Us,” and vice versa. But, what happens when a heavy media user comes across a fake
news story that explicitly calls out an established news source as fraudulent? This sug-
gests that real news is failing to function as a corrective for the misperceptions perpetu-
ated by fake news, which would be consistent with recent studies in political communication
(Nyhan and Reifler, 2010).
We faced a number of limitations with this study that stem primarily from our depend-
ence on comScore data. For example, while comScore captures news exposure within
mobile news apps, it does not capture news exposure within social media platform apps
(e.g. Facebook and Twitter). In other words, we were able to include the people who logged
onto The New York Times via the publisher’s app but not the people who loaded a Times
Nelson and Taneja 15
story within their Facebook app’s internal browser. Considering the increasingly large role
social media platforms play in news consumption, this is an unfortunate limitation.
However, because in-app visits were omitted from comScore estimates for both real and
fake news sites, this limitation may not have affected the difference between these audi-
ences, which has been the focus of this study. Furthermore, comScore data does not pro-
vide data about the audience’s socio-economic status, their political attitudes, media
literacy, or trust in news. There is a rich scholarship that explores the relationship between
audience demographics and news consumption habits (e.g. politically active citizens con-
sume more partisan news (Levendusky, 2013), Republicans are more likely to avoid coun-
ter-attitudinal information (Garrett and Stroud, 2014), and people’s news media preferences
vary depending on the region in which they live (Althaus et al., 2009)). Although our find-
ings indicate that audience availability plays an important role in accounting for fake news
consumption, experimental and survey research that captures these more granular audience
traits might further explain engagement with fake news.
In addition, our analysis was unable to capture the way that the fake news crisis has
spread throughout the “dark social web,” which comprises web-based messaging plat-
forms, such as WhatsApp and WeChat. These platforms have large audiences outside the
United States, in places where mobile devices are how most people access the Internet.
Communication on these either occurs via private, one-on-one or small group corre-
spondences, and often all the information is in the message itself and does not link to the
Web (Napoli and Obar, 2014). Hence, assessing the spread of misinformation and fake
news on these platforms is not possible with third party databases, such as comScore. In
other words, a far more consequential fake news crisis may be unfolding within rela-
tively walled-off spheres of communication. Finally, we could not assess the way that
audiences reacted to the fake news they were exposed to. Some scholars have begun
exploring this question, by analyzing how audiences authenticate the accuracy of the
news stories they consume (Tandoc et al., 2017b). We look forward to seeing more work
in this vein.
When it comes to news audiences in the United States, our findings suggest that the
current news crisis may be less about an abundance of fake news than a lack of trust
when it comes to real news. Only about a third of Americans currently have confidence
in the mass media “to report the news fully, accurately and fairly”—the lowest level of
trust since Gallup began asking the question in 1972 (Swift, 2016). News organizations
were concerned with waning public trust before the election but since then, these wor-
ries have only amplified (Kantar, 2016). It is worth considering how the diminishing
trust in established news media and its weakened relationship with usage might shift the
balance of power between real and fake news going forward. If future generations feel
no trust toward what are currently the most well-known news brands, while fake news
producers continue to find novel ways to game social media algorithms, we could
expect to see a future where news audiences are increasingly exposed to fake news. In
light of these circumstances, we believe the current crisis facing political journalism is
not how to deal with fake news but instead how to persuade audiences that real news is
more credible—even when it tells them things that run counter to how they would like
the world to be.
16 new media & society 00(0)
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this
article.
Notes
1. The tool used to explore the sources of online traffic is only available with comScore’s desk-
top data.
2. It must be clarified that these user groups are non-exclusive. As the preceding cross visiting
analysis corroborates, most fake news users are a small subset of real news users.
3. The average time spent per user per month on Facebook alone during our data analysis period
was ranged about 300 minutes on desktop and 1000 minutes on mobile. The same figure for
Google.com is 160 minutes on desktop and 94 minutes on mobile.
References
Allcott H and Gentzkow M (2017) Social Media and Fake News in the 2016 Election. Stanford,
CA: Stanford University. Available at: http://web.stanford.edu/~gentzkow/research/fak-
enews.pdf
Althaus SL, Cizmar AM and Gimpel JG (2009) Media supply, audience demand, and the geog-
raphy of news consumption in the United States. Political Communication 26(3): 249–277.
Barthel M, Mitchell A and Holcomb J (2016) Many Americans believe fake news is sow-
ing confusion. Pew Research Center, 15 December. Available at: http://www.journalism.
org/2016/12/15/many-americans-believe-fake-news-is-sowing-confusion/
Barthel M, Shearer E, Gottfried J, et al. (2015) The evolving role of news on Twitter and Facebook.
Pew Research Center, 14 July. Available at: http://www.journalism.org/2015/07/14/the-
evolving-role-of-news-on-twitter-and-facebook/
Borden SL and Tew C (2007) The role of journalist and the performance of journalism: ethical
lessons from “fake” news (seriously) (Author abstract). Journal of Mass Media Ethics 22(4):
300–314.
Buchanan M (2017) Commentary: why fake news spreads like wildfire on Facebook. Chicago
Tribune, 3 September. Available at: http://www.chicagotribune.com/news/opinion/commen-
tary/ct-perspec-fake-news-google-facebook-0904-story.html
Bucher T (2017) The algorithmic imaginary: exploring the ordinary affects of Facebook algo-
rithms. Information, Communication & Society 20(1): 30–44.
Carlson M (2017) Facebook in the news. Digital Journalism 6: 4–20.
comScore (2013) comScore media metrix description of methodology. Available at: http://www.
journalism.org/files/2014/03/comScore-Media-Metrix-Description-of-Methodology.pdf
(accessed 30 October 2015).
Day A and Thompson E (2012) Live from New York, it’s the fake news! Saturday night live and
the (non)politics of parody. Popular Communication 10(1–2): 170–182.
Ehrenberg ASC, Goodhardt GJ and Barwise P (1990) Double jeopardy revisited. Journal of
Marketing 54(3): 82–91.
Elberse A (2008) Should you invest in the long tail? Harvard Business Review 86(7–8): 88–96.
Fletcher R and Nielsen RK (2017) Are people incidentally exposed to news on social media?
A comparative analysis. New Media & Society. Epub ahead of print 17 August. DOI:
10.1177/1461444817724170.
Flood B (2016) Mike Huckabee apologizes for sharing fake news about “liberal, Jewish” students.
The Wrap, 18 November. Available at: http://www.thewrap.com/mike-huckabee-facebook-
fake-news-liberal-jewish-students-apology/
Nelson and Taneja 17
Garrett RK and Stroud NJ (2014) Partisan paths to exposure diversity: differences in pro- and
counterattitudinal news consumption. Journal of Communication 64(4) 680–701.
Gentzkow M and Shapiro JM (2011) Ideological segregation online and offline. Quarterly Journal
of Economics 126(4): 1799–1839.
Glaser A (2017) Apple CEO Tim Cook says fake news is “killing people’s minds” and tech
needs to launch a counterattack. Recode, 12 February. Available at: http://www.recode.
net/2017/2/12/14591522/apple-ceo-tim-cook-tech-launch-campaign-fake-news-fact-check
Gottfried J and Shearer E (2016) News use across social media platforms 2016. Pew Research
Center, 26 May. Available at: http://www.journalism.org/2016/05/26/news-use-across-
social-media-platforms-2016/
Guess AM (2016) Media choice and moderation: evidence from online tracking data. Available at:
https://www.dropbox.com/s/uk005hhio3dysm8/GuessJMP.pdf?dl=0
Guess, Andy, Brendan Nyhan, and Jason Reifler. 2018. ‘Selective Exposure to Disinformation:
Evidence from the Consumption of Fake News During the 2016 US Presidential Campaign’.
https://www.dartmouth. edu/~nyhan/fake-news-2016.pdf
Kantar (2016) Brand and trust in a fragmented news environment. Reuters Institute for the
Study of Journalism. Available at: http://reutersinstitute.politics.ox.ac.uk/sites/default/files/
Brandandtrustinafragmentednewsenvironment.pdf
Levendusky, M. (2013). How Partisan Media Polarize America. Chicago: University of Chicago
Press.
McPhee WN (1963) Formal Theories of Mass Behavior. New York: Free Press.
Martin CR (1973) The theory of double jeopardy. Journal of the Academy of Marketing Science
1: 148–155.
Mitchelstein E and Boczkowski PJ (2010) Online news consumption research: an assessment of
past work and an agenda for the future. New Media & Society 12(7): 1085–1102.
Napoli PM (2014) Automated media: an institutional theory perspective on algorithmic media
production and consumption. Communication Theory 24(3): 340–360.
Napoli PM and Obar JA (2014) The emerging mobile Internet underclass: a critique of mobile
Internet access. The Information Society 30(5): 323–334.
Nelson JL (2017) Is fake news a fake problem? Columbia Journalism Review, 31 January.
Available at: http://www.cjr.org/analysis/fake-news-facebook-audience-drudge-breitbart-
study.php
Nelson JL and Lei RF (2017) The effect of digital platforms on news audience behavior. Digital
Journalism. Epub ahead of print 2 November. DOI: 10.1080/21670811.2017.1394202.
Nelson JL and Webster JG (2016) Audience currencies in the age of big data. International Journal
on Media Management 18(1): 9–24.
Nelson JL and Webster JG (2017) The myth of partisan selective exposure: a portrait of the online
political news audience. Social Media + Society. Epub a head of print 7 September. DOI:
10.1177/2056305117729314.
Nyhan B and Reifler J (2010) When corrections fail: the persistence of political misperceptions.
Political Behavior 32(2): 303–330.
Owen LH (2016) Dec. 15 Clamping down on viral fake news, Facebook partners with sites like
Snopes and adds new user reporting. Niemanlab, 15 December. Available at: http://www.
niemanlab.org/2016/12/clamping-down-on-viral-fake-news-facebook-partners-with-sites-
like-snopes-and-adds-new-user-reporting/
Prior M (2013) Media and political polarization. Annual Review of Political Science 16: 101–127.
Salganik MJ, Dodds PS and Watts DJ (2006) Experimental study of inequality and unpredictabil-
ity in an artificial cultural market. Science (New York, N.Y.), 311(5762), 854–6.
18 new media & society 00(0)
Silverman C, Hall E, Strapagiel L, et al. (2016) Hyperpartisan Facebook pages are publishing false
and misleading information at an alarming rate. Buzzfeed, 20 October. Available at: https://
www.buzzfeed.com/craigsilverman/partisan-fb-pages-analysis?utm_term=.peg6RnKE3-.
pvVKrpRlG
Sonderman J (2014) Mobile and social media are intricately linked. American Press Institute,
10 June. Available at: https://www.americanpressinstitute.org/publications/reports/white-
papers/mobile-and-social-media/
Swift A (2016) Americans’ trust in mass media sinks to new low. Gallup, 14 September. Available
at: http://www.gallup.com/poll/195542/americans-trust-mass-media-sinks-new-low.aspx
Tandoc EC, Lim ZW and Ling R (2017a) Defining “fake news.” Digital Journalism. Epub ahead
of print 30 August. DOI: 10.1080/21670811.2017.1360143.
Tandoc EC, Ling R, Westlund O, et al. (2017b) Audiences’ acts of authentication in the age of fake
news: a conceptual framework. New Media & Society. Epub ahead of print 21 September.
DOI: 10.1177/1461444817731756
Taneja H (2016) Using Commercial Audience Measurement Data in Academic Research.
Communication Methods and Measures, 10(2–3), 176–17.
Taneja H and Viswanathan V (2014) Still glued to the box? Television viewing explained in a
multi-platform age integrating individual and situational predictors. International Journal of
Communication 8: 2134–2159. Available at: http://ijoc.org/index.php/ijoc/article/view/2841/1193
Taneja H and Webster JG (2016) How do global audiences take shape? The role of institutions and
culture in patterns of web use. Journal of Communication 66(1): 161–182.
Taneja H, Webster JG, Malthouse EC, et al. (2012) Media consumption across platforms: identify-
ing user-defined repertoires. New Media & Society 14(6): 951–968.
Taneja, H., Wu, A. X., & Edgerly, S. (2017). Rethinking the generational gap in online news use:
An infrastructural perspective. New Media & Society. doi:10.1177/1461444817707348
Timberg C (2016) Russian propaganda effort helped spread “fake news” during election, experts
say. The Washington Post, 24 November. Available at: https://www.washingtonpost.com/
business/economy/russian-propaganda-effort-helped-spread-fake-news-during-election-
experts-say/2016/11/24/793903b6-8a40-4ca9-b712-716af66098fe_story.html?utm_term=.
eebfde8a22f8
Uberti D (2017) “Fake news” is dead. Columbia Journalism Review, 14 February. Available at:
http://www.cjr.org/criticism/fake_news_trump_white_house_cnn.php
Webster JG (2014) The Marketplace of Attention: How Audiences Take Shape in a Digital Age.
Cambridge, MA: The MIT Press.
Webster JG and Wakshlag J (1983) A theory of television program choice. Communication
Research 10(4): 430–446.
Webster JG, Phalen PF and Licthy LW (2014) Ratings Analysis: Audience Measurement and
Analytics. 4th ed. New York: Routledge.
Zengerle P (2016) Clinton calls “fake news” a threat to U.S. democracy. Reuters, 9 December.
Available at: http://www.reuters.com/article/us-usa-clinton-fakenews-idUSKBN13X2R6
Author biographies
Jacob L Nelson is a doctoral candidate at Northwestern University’s Media, Technology, and
Society program. He will be joining the Walter Cronkite School of Journalism and Mass
Communication at Arizona State University as an assistant professor in August 2018. He researches
issues in news production and consumption.
Harsh Taneja is an assistant professor at the Institute for Communications Research at University
of Illinois Urbana-Champaign. His research explores the connections between media use and its
associated social and institutional contexts.
... An analysis of the state of the art of research carried out in the topics of trustworthiness, 'fake news', filter bubbles and echo chambers and the like present relevant learnings for EUNOMIA. Especially after the 2016 US Presidential elections, studies on the trustworthiness of information distributed online were carried out in increasing numbers (e.g., Allcott & Gentzkow 2017;Lewandowsky, Ecker, & Cook 2017;Nelson & Taneja 2018;Pennycook & Rand 2019a-c). While research after the 2016 US Presidential elections has a rather negative focus on disinformation, misinformation and 'fake news' 1 , there is a body of research prior to this event focusing on the positive aspects and potential of social media as a source of information on news and public affairs, as a forum of political expression and participation, or a means to empower users to actively participate in the news distribution process. ...
... Audience availability refers to the amount of time the audience has for using a specific medium in the first place (Nelson & Taneja, 2018). Originated from research on TV consumption, this holds not only true for television programmes, for which audiences stay more or less the same for a specific time of the day, but also for 'new media': here, too, audiences are bound by the fixed amount of time they have for media use, despite having more autonomy and options. ...
... Originated from research on TV consumption, this holds not only true for television programmes, for which audiences stay more or less the same for a specific time of the day, but also for 'new media': here, too, audiences are bound by the fixed amount of time they have for media use, despite having more autonomy and options. Research further shows that audiences with more availability -heavy media users -have more time exploring different media options, which light media users spend their time on the most popular offerings (Nelson & Taneja 2018). ...
Technical Report
Full-text available
This report documents the outcomes of an analysis of user behaviour on social media regarding the approval, assessment and evaluation of information and information sources, feeding into the further development of the EUNOMIA toolkit. Both individual and collective behaviour was analysed. On the one hand, there are factors that cause and explain individual behaviour, such as cognitive biases and psychological effects that influence a single person’s behaviour. An example is the so-called truth effect, i.e. the fact that repetition and familiarity with content make it more believable. On the other hand, group effects and social norms additionally influence the individual’s behaviour. Studies have shown that we are more likely to believe a piece of information if our social circles also accept it (Lewandowski et al. 2012; Eckles & Bakshy 2017; Lazer et al. 2017; Sloman & Fernbac, 2018; Karduni 2019). The task of user behaviour analysis included (i) a literature review; (ii) a workshop with end users and experts; and (iii) an online survey. We identified explanations for collective and individual user behaviour in assessing, sharing and distributing (mis)information, building on (i) the theory of cognitive dissonance and the theory of selective exposure; (ii) the third-person effect; (iii) the concept of opinion leadership; (iv) the concept of information gatekeeping; (v) the truth effect; and (vi) explanations for the persistence of misinformation and (vii) audience behaviour. Insights explain, for example, how users on social media tend to surround themselves with information that confirms their own interests, values and beliefs in so-called ‘filter bubbles’ or ‘echo chambers’. Furthermore, we were able to identify strategies to influence or reward preferable behaviour to avoid the spread of misinformation in the form of nudges, building on certain heuristics (i.e., mental shortcuts) and psychological or social effects. We also identified approaches for the correction of misinformation (e.g., providing explanations, targeting opinion leaders), as well as strategies to avoid their spread (e.g., triggering a thinking process before information is read).
... Explanations for such significant differences between countries, ranging from political polarization to diverse use of social media, are common in recent research (Nelson & Taneja, 2018;Spohr, 2017;Urman, 2019), but many others are possible and reasonable, and they have not been sufficiently tested. One explanatory factor differentiating the experience of fake news among people from diverse countries could be cultural, a fact well documented in cross-national and cross-cultural studies in other fields, and also in media and journalism studies (Esser & Hanitzsch, 2012). ...
... As it is understandable with a phenomenon as new and expansive as fake news, the interest in measuring its spread and penetration, both within a country and comparatively between countries, has been a topic of great interest from the outset. However, beyond partial and generally inconclusive studies-especially in the United States (Allcott & Gentzkow, 2017Nelson & Taneja, 2018;Vosoughi et al., 2018)-little is known about the objective, measurable, quantifiable severity of this infodemic. Lazer et al. (2018) echoed the unknown scope of this problem when highlighting that there are surprisingly few scientific answers to these basic questions. ...
... Much of the research on how fake news works has focused on analyzing the reception and processing of this information by people, individually or in specific social groups. Numerous studies from the different perspectives of psychology, social psychology, consumer behavior, and media use have explained how the success of fake news depends to a large extent on factors and processes such as political partisanship and bias (Pennycook & Rand, 2019;Van der Linden et al., 2020), echo chambers or filter bubbles (Spohr, 2017;Törnberg, 2018), confirmation bias (Kim et al., 2020), selective exposure (Barnidge & Peacock, 2019;Hameleers et al., 2020), availability bias (Nelson & Taneja, 2018), information avoidance (Kim et al., 2020), "news-finds-me" perception (Gil de Zúñiga & Diehl, 2018), and other cognitive and emotional mechanisms (Bakir & McStay, 2017). ...
Article
[Abstract] This article analyzes the extent to which certain cultural dimensions explain the intensity with which citizens of different countries perceive the presence of fake news in their daily lives. The research is based on the Flash Eurobarometer survey conducted in 2018 about fake news and disinformation online in 25 European countries, and adopts the Hofstede cultural dimensions as a model of cultural analysis. The study uses multilevel regression analysis to test individual and macro-level indicators that explain variations in perceptions of fake news. The findings reveal a clear, direct relationship between uncertainty avoidance, masculinity, and fake news exposure, as well as an interaction of these cultural dimensions with age, but not with the other individual and media use related variables. These results have theoretical and practical implications, especially from the point of view of the design of public policies to fight disinformation in the European Union (EU).
... On the contrary, many readers are regularly exposed to cross-cutting news media (Guess & Nyhan, 2018). Even in the case of misinformation, readers who fully disregard established news outlets are a slim minority (Nelson & Taneja, 2018). Overwhelmingly, empirical research demonstrates evidence for a news audience that is more connected than it is disparate, more concentrated than it is fragmented. ...
... Because of a lack of familiarity and fewer resources, smaller offerings in many markets also experience the most disloyal audiences (Ehrenberg et al., 1990). The same holds true for digital news, where, for example, the audience for misinformation is both disloyal and dwarfed by that of mainstream media (Nelson & Taneja, 2018). ...
Article
Full-text available
Many attention markets exhibit stable patterns of concentration, where a few producers attract and sustain a far greater share of the audience than others. This inequality often follows patterns consistent with cumulative advantage, a process in which performance compounds over time. Attention to news sources online possesses these characteristics; however, online audiences also fragment across many disparate news producers. How do social media and recommender systems contribute to these attention dynamics? In this study, we examine two paradigmatic models: concentration driven by cumulative advantage and fragmentation driven by stochasticity. We evaluate these models against a large-scale empirical dataset of news source attention in the popular social media site Reddit. While we find high levels of attention concentration, we do not find the stable popularity over time that characterizes cumulative advantage. Rather, sources gain and lose popularity seemingly at random, aligning with a stochastic model. These results demonstrate the persistence of attention inequality, even in the absence of a strong driving mechanism. They also suggest that social media systems can undermine the accumulation of attention to the most prominent news sources. Digital attention markets striving for more equitable allocation require novel mechanisms of organizing and distributing information.
... False news reports can go viral and be widely disseminated within a short time frame. Because of the ease of accessing different types of content on social media platforms such as Facebook, many people rely on social media to consume news over traditional news sources such as print and television (Nelson and Taneja, 2018;Gerbaudo, 2018). Fake news, disguised as real news, makes it almost impossible to detect as most individuals may not be able to determine the authenticity that separates facts from made up 'news' (Pennycook and Rand, 2019). ...
... Another feature of Facebook is Facebook Groups which is a community of like-minded users. It is possible to promote fake groups and increase membership in these groups so fake news content can be seeded to amplify differences and generate emotional reactions (Lazer et al., 2018;Nelson and Taneja, 2018). ...
Article
Full-text available
Facebook is the largest social media platform that is used by all generations of users, as well as small and large businesses. Many users consider Facebook as a primary news source even though the news on Facebook is not authenticated. This 'fake news' can be used for financial or political gain and can also impact consumer behaviour towards products. The purpose of this study was to investigate advertising response behaviour and fake news perception among multi-generational Facebook users, in conjunction with other variables such as gender. Using a survey, data were collected from a multi-stage quota sample of 400 respondents in the USA. A scale was developed and psychometrically tested as part of the study to determine fake news perception. Findings of this study showed that the frequency of Facebook use was consistent among generations, with Baby Boomers being most active in reading posts, and Gen Y users being most active in posting to Facebook. Gen Y users found Facebook advertisements to be most relevant. Results can be used to drive engagement with Facebook users and develop campaigns that use actionable segmentation schemes. Implications of fake news perception are discussed, and future research directions are provided.
... Nevertheless, other studies contradict the findings, showing that online misinformation consumption is limited to only a small audience (Allen et al., 2020;Grinberg et al., 2019). By comparing audience traffic data of 30 fake news sites and 24 real news sites based in the United States in 2016, Nelson and Taneja (2018) showed that the number of monthly visitors to "an average real news site was more than 40 times larger than the number of monthly visitors to an average fake news site" (p. 3,727). ...
Article
Full-text available
In the wake of the U.S. 2016 Presidential Election, concerns about misinformation traversing on social media have heightened. Since then, much of the public discourse has been on developing effective strategies to combat the spread of misinformation online. While several studies have focused on the effects of a mixed/hybrid regime of information channels on political participation and campaigns, we know little about how the existence of a hybrid media system exposes people to misinformation during an election cycle. Using a nationally representative survey administered during the 2019 Indonesian election (N=1,820), we find evidence for the prominence of traditional media as well as face-to-face discussions: political use of traditional media such as newspapers and TV as well as sharing of political information through face-to-face discussions are found to be positively associated with at least one measure of misinformation exposure. As for the social media communicative pathways, only political use of WhatsApp and Instagram are found to be positively associated with misinformation exposure; no similar effects are observed for Facebook and Twitter attests, to some extent, to the efficacy of strategies aimed at combating misinformation implemented on such platforms. By considering social media, traditional media, and face-to-face communication in a context of a less digitalized hybrid media environment, this paper provides a more comprehensive framework and novel empirical data to study misinformation exposure beyond the context of Western democracies.
... The heavy media users are aware of most of the information alternatives that are offered by the media ecosystem since they have the time to investigate more. According to a study [67], the heavy media users cover a smaller part of the overall audience but they are also the ones that exploit the popular, as well as the unpopular, news outlets; thus, they are far more prone to propagate fake news content [68]. On the other hand, those with less free time, and thus less availability, are characterized as light media users [69]. ...
Article
Full-text available
The emergence of the Internet and web technologies has magnified the occurrence of disinformation events and the dissemination of online fake news items. Fake news is a phenomenon where fake news stories are created and propagated online. Such events occur with ever increasing frequency, they reach a wide audience, and they can have serious real-life consequences. As a result, disinformation events are raising critical public interest concerns as in many cases online news stories of fake and disturbing events have been perceived as being truthful. However, even at a conceptual level, there is not a comprehensive approach to what constitutes fake news with regard to the further classification of individual occurrences and the detection/mitigation of actions. This work identifies the emergent properties and entities involved in fake news incidents and constructs a disinformation blueprint (DCAM-DB) based on cybercrime incident architecture. To construct the DCAM-DB in an articulate manner, the authors present an overview of the properties and entities involved in fake news and disinformation events based on the relevant literature and identify the most prevalent challenges. This work aspires to enable system implementations towards the detection, classification, assessment, and mitigation of disinformation events and to provide a foundation for further quantitative and longitudinal research on detection strategies.
Chapter
In a so-called post-truth era, research on the subject of the spread of mis- and disinformation is being widely explored across academic disciplines in order to further understand the phenomenon of how information is disseminated by not only humans but also the technology humans have created (Tandoc, Sociol Compass 13(9), 2019). As technology advances rapidly, it is more important than ever to reflect on the effects of the spread of both mis- and disinformation on individuals and wider society, as well as how the impacts can be mitigated to create a more secure online environment. This chapter aims to analyse the current literature surrounding the topic of artificial intelligence (AI) and the spread of mis- and disinformation, beginning with a look through the lens of the meaning of these terms, as well as the meaning of truth in a post-truth world. In particular, the use of software robots (bots) online is discussed to demonstrate the manipulation of information and common malicious intent beneath the surface of everyday technologies. Moreover, this chapter discusses why social media platforms are an ideal breeding ground for malicious technologies, the strategies employed by both human users and bots to further the spread of falsehoods within their own networks, and how human users further the reach of mis- and disinformation. It is hoped that the overview of both the threats caused by and the solutions achievable by AI technology and human users alike will further highlight the requirement for more progress in the area at a time when the spread of falsehoods online continues to be a source of deep concern for many. This chapter also calls into question the use of AI to combat issues arising from the use of advanced Machine Learning (ML) methods. Furthermore, this chapter offers a set of recommendations to help mitigate the risks, seeking to explore the role technology plays in a wider scenario in which ethical foundations of communities and democracies are increasingly being threatened.
Chapter
Full-text available
Various recent events, such as the COVID-19 pandemic or the European elections in 2019, were marked by the discussion about potential consequences of the massive spread of misinformation, disinformation, and so-called “fake news.” Scholars and experts argue that fears of manipulated elections can undermine trust in democracy, increase polarization, and influence citizens’ attitudes and behaviors (Benkler et al. 2018; Tucker et al. 2018). This has led to an increase in scholarly work on disinformation, from less than 400 scientific articles per year before 2016 to about 1’500 articles in 2019. Within social sciences, surveys and experiments dominated in the last few years. Content analysis is used less frequently and studies conducting content analyses mostly use automated approaches or mixed methods designs.
Article
Full-text available
Occupational health and safety are critical in promoting the wellness of organizations and employees. The COVID-19 pandemic is one of the most life-threatening viruses encountered in recent history, providing a unique opportunity for research to examine factors that drive employee safety behavior. Drawing from terror management theory, we propose and test a moderated mediation model using data collected from employees working during a peak of the pandemic. We identify two sources of influence - one external (i.e., media exposure), and one internal (i.e., HR practices) to the organization - that shape employees' mortality salience and safety behaviors. We find that COVID-19 HR practices significantly moderate the relationship between daily COVID-19 media exposure and mortality salience, with media exposure positively associated with mortality salience at lower levels of HR practices but its effects substituted by higher levels of HR practices. Moreover, our results also show that mortality salience spurs safety behaviors, with age moderating this relationship such that younger - but not older - employees are more likely to engage in safety behaviors due to mortality salience. Taken together, we offer theoretical implications for the safety behavior literature and practical implications for organizations faced with health crises or having employees who commonly work in hazardous conditions.
Article
Full-text available
Journalism researchers and professionals are increasingly focused on the mobile platform, which many believe has become the most popular method of digital news consumption. However, few studies that examine digital news use draw from observed audience behavior. Furthermore, little is known about whether or not different digital platforms affect how audiences actually attend to news. This article addresses these gaps. Using online audience data, we find that over the past year, the mobile news audience has indeed surpassed the desktop news audience in size, and that a subset of this audience devotes an enormous amount of time to news apps, especially compared with time spent on desktop devices. In light of our findings, we argue the mobile news audience should be seen not as one homogenous mass, but as two distinct groups—mobile news browsers and mobile app users. We conclude by discussing the implications of these findings for news production.
Article
Full-text available
Through an analysis of relevant literature and open-ended survey responses from 2501 Singaporeans, this article proposes a conceptual framework to understand how individuals authenticate the information they encounter on social media. In broad strokes, we find that individuals rely on both their own judgment of the source and the message, and when this does not adequately provide a definitive answer, they turn to external resources to authenticate news items.
Article
Full-text available
Many assume that in a digital environment with a wide range of ideologically tinged news outlets, partisan selective exposure to like-minded speech is pervasive and a primary cause of political polarization. Yet, partisan selective exposure research tends to stem from experimental or self-reported data, which limits the applicability of their findings in a high-choice media environment. We explore observed online audience behavior data to present a portrait of the actual online political news audience. We find that this audience frequently navigates to news sites from Facebook, and that it congregates among a few popular, well-known political news sites. We also find that political news sites comprise ideologically diverse audiences, and that they share audiences with nearly all smaller, more ideologically extreme outlets. Our results call into question the strength of the so-called red/blue divide in actual web use.
Article
Full-text available
This paper is based on a review of how previous studies have defined and operationalized the term “fake news.” An examination of 34 academic articles that used the term “fake news” between 2003 and 2017 resulted in a typology of types of fake news: news satire, news parody, fabrication, manipulation, advertising, and propaganda. These definitions are based on two dimensions: levels of facticity and deception. Such a typology is offered to clarify what we mean by fake news and to guide future studies.
Article
In any given time period, a small brand typically has far fewer buyers than a larger brand. In addition, its buyers tend to buy it less often. This pattern is an instance of a widespread phenomenon called “double jeopardy” (DJ). The authors describe the wide range of empirical evidence for DJ, the theories that account for its occurrence, known exceptions and deviations, and practical implications.
Article
Scholars have questioned the potential for incidental exposure in high-choice media environments. We use online survey data to examine incidental exposure to news on social media (Facebook, YouTube, Twitter) in four countries (Italy, Australia, United Kingdom, United States). Leaving aside those who say they intentionally use social media for news, we compare the number of online news sources used by social media users who do not see it as a news platform, but may come across news while using it (the incidentally exposed), with people who do not use social media at all (non-users). We find that (a) the incidentally exposed users use significantly more online news sources than non-users, (b) the effect of incidental exposure is stronger for younger people and those with low interest in news and (c) stronger for users of YouTube and Twitter than for users of Facebook.
Article
The move to high-choice media environments has sparked fears over audience fragmentation. We analyze news audiences across media platforms (print, television, and online) in 6 countries, going beyond platform-specific, single-country studies. We find surprisingly high levels of news audience duplication, but also that cross-platform audiences vary from country to country, with fragmentation higher in Denmark and the United Kingdom than in Spain and the United States. We find no support for the idea that online audiences are more fragmented than offline audiences, countering fears associated with audience segmentation and filter bubbles. Because all communication exists in the context of its audience, our analysis has implications across the field, underlining the importance of research into how trends play out in different contexts.
Article
Following the 2016 US presidential election, many have expressed concern about the effects of false stories ("fake news"), circulated largely through social media. We discuss the economics of fake news and present new data on its consumption prior to the election. Drawing on web browsing data, archives of fact-checking websites, and results from a new online survey, we find: 1) social media was an important but not dominant source of election news, with 14 percent of Americans calling social media their "most important" source; 2) of the known false news stories that appeared in the three months before the election, those favoring Trump were shared a total of 30 million times on Facebook, while those favoring Clinton were shared 8 million times; 3) the average American adult saw on the order of one or perhaps several fake news stories in the months around the election, with just over half of those who recalled seeing them believing them; and 4) people are much more likely to believe stories that favor their preferred candidate, especially if they have ideologically segregated social media networks.
Article
Our study investigates the role of infrastructures in shaping online news usage by contrasting use patterns of two social groups,millennials and boomers,that are specifically located in news infrastructures. Typically based on self reported data, popular press and academics tend to highlight the generational gap in news usage and link it to divergence in values and preferences of the two age cohorts. In contrast, we conduct relational analyses of shared usage obtained from passively metered usage data across a vast range of online news outlets for millennials and boomers. We compare each cohort's usage networks comprising various types of news websites. Our analyses reveal a smaller than commonly assumed generational gap in online news usage, with characteristics that manifest the multifarious effects of the infrastructural aspect of the media environment, alongside those of preferences.
Article
In May 2016, two stories on Facebook’s Trending Topics news feature appeared on the site Gizmodo, the first exposing the human curators working surreptitiously to select news and the second containing accusations that certain curators censored conservative voices. These revelations led to public outcry, mostly from conservatives upset at possible bias and journalists critical of Facebook for its largely secretive and haphazard approach to news. This study examines the response to the controversy as metajournalistic discourse—talk about news that seeks to define appropriate practices and legitimate news forms. It identifies a fundamental divergence in the public articulation of Facebook’s role in the larger news ecosystem. Facebook reacted to the controversy by formulating an approach to news as a form of content that, like other content on the site, should be personalized, organized according to popularity in the form of user engagement, and free of editorial control from the social media site. By contrast, journalists positioned news as purposively selected and shared while placing Facebook as an active participant within the news ecosystem and therefore beholden to an enhanced institutional commitment to public responsibility.