Public Opinion Quarterly, Vol. 84, Special Issue, 2020, pp.

236–256

“TAKING THE TEMPERATURE OF THE ROOM”

HOW POLITICAL CAMPAIGNS USE SOCIAL MEDIA TO
UNDERSTAND AND REPRESENT PUBLIC OPINION

SHANNON C. McGREGOR*

Downloaded from https://academic.oup.com/poq/article/84/S1/236/5871904 by guest on 10 June 2025

Abstract For most of the twentieth century, public opinion was nearly
analogous with polling. Enter social media, which has upended the social,
technical, and communication contingencies upon which public opinion
is constructed. This study documents how political professionals turn to
social media to understand the public, charting important implications for
the practice of campaigning as well as the study of public opinion itself.
An analysis of in-depth interviews with 13 professionals from 2016 US
presidential campaigns details how they use social media to understand
and represent public opinion. I map these uses of social media onto a the-
oretical model, accounting for quantitative and qualitative measurement,
for instrumental and symbolic purposes. Campaigns’ use of social media
data to infer and symbolize public opinion is a new development in the
relationship between campaigns and supporters. These new tools and sym-
bols of public opinion are shaped by campaigns and drive press coverage
(McGregor 2019), highlighting the hybrid logic of the political media
system (Chadwick 2017). The model I present brings much-needed atten-
tion to qualitative data, a novel aspect of social media in understanding
public opinion. The use of social media data to understand the public, for
all its problems of representativeness, may provide a retort to long-standing
criticisms of surveys—specifically that surveys do not reveal hierarchical,
social, or public aspects of opinion formation (Blumer 1948; Herbst 1998;
Cramer 2016). This model highlights a need to explicate what can—and
cannot—be understood about public opinion via social media.

Shannon C. McGregor is currently an assistant professor in the Hussman School of Journalism

and Media at the University of North Carolina at Chapel Hill, Chapel Hill, NC, USA, and a senior re-
searcher at the UNC Center for Information, Technology, and Public Life. This research was completed
while McGregor was an assistant professor in the Department of Communication at the University of
Utah. The author wishes to thank all the campaign staffers who provided their time and insights for
this project. The author also wishes to thank Gina Chen, Regina Lawrence, Thomas Johnson, Stephen
Reese, Talia Stroud, Kevin Coe, and Emily Van Duyn for their feedback on an earlier version of this
study. Many thanks go to the three anonymous reviewers and the editors for providing feedback and
suggestions that significantly improved the study. *Address correspondence to Shannon C. McGregor,
Hussman School of Journalism, University of North Carolina, Carroll Hall, Chapel Hill, NC 27599,
USA; email: [email protected].
doi:10.1093/poq/nfaa012 Advance Access publication July 15, 2020
© The Author(s) 2020. Published by Oxford University Press on behalf of the American Association for Public Opinion Research.
All rights reserved. For permissions, please e-mail: [email protected]
Campaign Use of Social Media as Public Opinion 237

Public opinion is endlessly in transition (Herbst 2011), accompanied by

changes in meaning, definition, and measurement. Enter social media, a not-
able aspect of the hybrid media system (Chadwick 2017), which facilitates
interactions between the media, political actors, and the public. These shifting
dynamics have restructured the social, technical, and communication contin-
gencies upon which public opinion is constructed. Alongside surveys, myriad
campaigns, journalists, and scholars fashion likes, comments, follows, and

Downloaded from https://academic.oup.com/poq/article/84/S1/236/5871904 by guest on 10 June 2025

retweets into an emergent form of public opinion. The process by which elites
construct public opinion is one way that we may observe the shifting nature
of public opinion.
This study documents how political professionals turn to social media to
understand the public, charting important implications for the practice of cam-
paigning as well as the study of public opinion itself. An analysis of in-depth
interviews with professionals from 2016 US presidential campaigns details
how they use social media to understand and represent public opinion, and
how it is measured and deployed. I map campaign uses of social media to
understand and convey public opinion onto a theoretical model, accounting
for quantitative and qualitative measurement, for instrumental and symbolic
purposes.
At first blush, campaigns’ use of social media data to infer and symbolize
public opinion may appear to put the public closer to politics—but results
show a mostly top-down affair, driven by campaigns. These new tools and
symbols of public opinion are shaped by campaigns and drive press coverage
(McGregor 2019), highlighting the hybrid logic of the political and media
system (Chadwick 2017). Political opinions on social media are highly ma-
nipulable, and “authentic” users who engage with politics on any given social
media platform are not representative of any larger public—in particular, they
tend to be more partisan, polarized, and uncivil (Kushin and Kitchener 2009;
Groshek and Koc-Michalska 2017; Sydnor 2018; Cohn and Quealy 2019;
Wojcik and Hughes 2019). This study accounts for how elites use social media
to assess and portray the public with consequences for how the public in turn
conceives of themselves.

Political Campaigns and Public Opinion

To understand the role that social media play in the definition and measure-
ment of public opinion, it is necessary to consider how political professionals
understand the public. Campaigns convey a reflection of the public back to
itself through their constructions of public opinion, which inform strategic
decisions and campaign communication. How do political professionals in-
corporate social media into their definitions of the “slippery entity” of public
opinion (Herbst 1998)?
238 McGregor

To make sense of how campaign professionals use social media to under-

stand public opinion, we can think along two lines: its uses and its data.
Opinion data can be used for both instrumental and symbolic purposes by
campaigns (Herbst 1993). As Herbst (1993, p. 158) concludes in Numbered
Voices, “The ability to conduct polls and use them privately (instrumental) and
the ability to manipulate these data once they are publicized (symbolic) should
both be evaluated if one is to understand the alleged power of polls.” As this

Downloaded from https://academic.oup.com/poq/article/84/S1/236/5871904 by guest on 10 June 2025

study will show, campaigns utilize social media data for both purposes. Like
polling, social media data are aggregated and understood using quantitative
methods. But a unique aspect of social media is that data are also represented
and understood in qualitative ways—drawn from individuals’ conversations
with their online networks.
Instrumental thinking ascribes the increased quantification and focus on
statistics to the desire to “routinize processes of observation” (Herbst 1993,
p. 12). In this sense, public opinion data are collected and utilized with a par-
ticular end in mind. Instrumental uses refer to “… the straightforward, mani-
fest use of numerical data” (Herbst 1993, p. 29) as functional data for political
professionals. Campaigns have long used polls for instrumental purposes—to
inform everything from efforts to mobilize and persuade the public to deci-
sions about where to focus on-the-ground efforts and the strategic deploy-
ment of resources. The instrumental use of social media to understand the
public has been examined but focuses narrowly on its potential as compared
to polls. Scholars have rightly noted that the user base of any social media
site is not representative of any country’s population, and thus is a poor fit for
electoral prediction or as a stand-in for polls (Jungherr 2015; Schober et al.
2016). Work comparing social media metrics to election outcomes and polling
is undertheorized (Jungherr 2015)—there is scant evidence linking the motiv-
ations to, for example, tweet about a candidate with motivations to vote for
a candidate. Yet, some people do express their public attitudes about politics
on social media, and journalists use social media to report on public opinion
(McGregor 2019).
While less obvious than instrumental uses, opinion data also serve symbolic
purposes for campaigns. The symbolic use of public opinion data is centered
on emotional appeals (Herbst 1993), such as when it is used to evoke partisan
identities. While public opinion measures may be designed in service of ra-
tionality, as political discourse develops around them, they emerge as sym-
bols of public opinion, utilized for affective appeals or to convey authority.
Quantified public opinion can be deployed rhetorically, for example to indi-
cate that a political opponent is “out of touch” with the majority of the public.
The concept of public opinion itself is a strong rhetorical device, which can be
partially attributed to its personification;1 for example, “public opinion favors

1. Krippendorff (2005, p. 130) asserts that personification is “the most pervasive metaphorical
root of the social construction of public opinion.”
Campaign Use of Social Media as Public Opinion 239

one candidate” (Krippendorff 2005). Polls, empowered by their scientific-ness

and representativeness, are themselves powerful symbols, their very release
capable of driving news coverage (Herbst 1993). Likewise, the release of so-
cial media metrics around events like debates drives news coverage—journal-
ists’ pen stories declaring political winners and losers based on these metrics
(McGregor 2019). Far-right groups target journalists specifically, in part
through manipulation of social media metrics as a marker of newsworthiness,

Downloaded from https://academic.oup.com/poq/article/84/S1/236/5871904 by guest on 10 June 2025

driving conspiracies into mainstream news (Marwick and Lewis 2017). Social
media, as an emergent representation of public opinion, offers powerful means
to symbolically communicate opinion.
Examinations of instrumental and symbolic uses of public opinion center
on quantitative methods. Campaigns have a long history of quantitatively as-
sessing public opinion to identify supporters, predict which of them will vote,
ascertain who among the public is persuadable, and improve the effectiveness
of campaign communication. Early studies on persuasion and public opinion
“set the stage” (Baldwin-Philippi 2016) for campaigns’ use of quantitative
measures to test which messages might best persuade and mobilize (Berelson,
Lazarsfeld, and McPhee 1954; Berelson, Gaudet, and Lazarsfeld 1968). As
polling continued to rise in popularity and legitimacy (Igo 2007), campaigns
leveraged increasingly fine-grained public opinion data to create subgroups
of the public. Scholarly evidence of the utility of strategies derived from
microtargeting (Gerber and Green 2000, 2001; Bergan et al. 2005) prompted
a shift in the role of polling in campaigns. Campaigns pressed on constructing
individualized perceptions of the public, building multilayered databases of
public records, purchase histories, digital trace data, polling data, and more
(Kreiss 2012, 2016a; Nielsen 2012; Baldwin-Philippi 2015, 2016; Hersh 2015;
Karpf 2016). Accounts of data-driven or “technology-intensive campaigning”
(Kreiss 2016a) demonstrate the shifting ways political professionals seek to
understand the public. Social media data take quantitative forms in aggregated
metrics such as counts of followers or shares as well as the use of quantitative
methods to assess engagement metrics such as mentions or likes.
Campaign professionals also turn to more qualitative methods to construct
representations of public opinion. Fenno (1978) documents how members of
Congress visit their home districts, interacting with their constituencies, and
thereby constructing a sense of the public. Through these informal interactions,
politicians intuit how they should present themselves, which voters are sup-
porters or opponents, and what issues they care about (Fenno 1978). Herbst
(1998) finds that political staffers relied on conversations with lobbyists to
understand the public. Political professionals understand and seek out social
media in real time as a form of instantaneous public opinion that they use to
assess reactions to events on the campaign trail. Previous work (Kreiss 2016b;
Baldwin-Philippi 2015), without explicitly mentioning public opinion, reveals
that this happens informally in elections, when campaign staffers gauge their
own work through the lens of qualitative readings of social media. Journalists
240 McGregor

also turn to social media, especially Twitter, to gather and report vox populi
quotes, reading and reporting individual posts as qualitative expressions of
public opinion (Anstead and O’Loughlin 2015; McGregor 2019; Lukito et al.
2020). After decades of a public opinion informed primarily by quantitative
measures, interrogating the ways political professionals navigate qualitative
impressions of public opinion reveals one of the many unique aspects of social
media: the availability of qualitative measures of public opinion, at a previ-

Downloaded from https://academic.oup.com/poq/article/84/S1/236/5871904 by guest on 10 June 2025

ously unthinkable scale, in the form of individual posts and replies.
This study aims to understand the role of social media in how campaigns
assess, conceive of, and communicate public opinion. Social media data may
be used in quantitative and/or qualitative forms for instrumental and/or sym-
bolic purposes by campaigns. I propose this as a theoretical model, as shown
in figure 1. I ask:
How do campaign professionals use quantitative and qualitative meas-
urements of social media as public opinion for instrumental and symbolic
purposes?

Methods
I examine the ways in which campaigns use social media to understand public
opinion, taking up the 2016 US presidential election as a case study. There are
distinct differences between races at the presidential level and down-ballot
races, especially in regard to their use of data analytics (Baldwin-Philippi
2016). But, by focusing on professionals working on the top tier of campaigns,

Figure 1. Conceptual model of campaign methods and uses of social

media to represent public opinion.
Campaign Use of Social Media as Public Opinion 241

I drew on the combined knowledge of decades of work in politics across a

multitude of elections. I focused on professionals at the digital or social media
director level. Interrogating professionals who worked directly in this area,
and were therefore responsible for communicating insights up the organiza-
tional hierarchy, offered the greatest potential for insight.
In 2016 and 2017, I conducted interviews with digital and social media dir-
ectors, or people in similarly relevant positions, about their work on the 2016

Downloaded from https://academic.oup.com/poq/article/84/S1/236/5871904 by guest on 10 June 2025

primary and general election campaigns. A list of these individuals, their roles,
and the date(s) on which I interviewed them are available in Appendix A. All
interviews were conducted on the record, although participants could declare
any statement not for attribution, on background, or off the record at their dis-
cretion. On average, the interviews lasted about an hour. See Appendix B for
the interview protocol, though note that additional questions were asked, given
the nature and flow of each interview. At the end of each interview, I asked if
there were other people with whom the subject thought I should speak for my
research. As such, additional subjects were added throughout the data collec-
tion process.
An analysis of these interviews (recorded and then transcribed) allowed me
to understand how campaign professionals make sense of their work (Lindlof
and Taylor 2010). In reading and analyzing the interviews, I inductively devel-
oped categories and concepts to document how campaign professionals used
social media to understand and represent public opinion, looking for similar-
ities and differences in their accounts. I followed an integrated approach to de-
velop these categories—bridging inductive categories emerging from a textual
analysis while integrating existing literature on theories of public opinion and
political campaigning (Luker 2008). In developing these categories, I focused
on determining what specific patterns of social media usage by campaigns
demonstrate qualitative and quantitative measurements of public opinion for
instrumental and symbolic purposes.

The Pulse of Public Opinion: Representing the Public

Through Social Media in Campaigns
Throughout the course of the 2016 presidential election, campaigns used so-
cial media in a variety of ways to understand and communicate public opinion.
These emergent forms of public opinion informed strategy and outreach,
shaped and became candidate communication, and were deployed to influ-
ence press coverage.

QUANTITATIVE DATA, INSTRUMENTAL USES

No matter how well funded, campaigns are almost always strapped for re-
sources—this means it is impossible to run costly polls to gauge public
242 McGregor

reception to most campaign messages. Through their organic and paid

accounts, campaigns accessed engagement metrics from social media plat-
forms, which offered cost-effective alternatives to assessing public opinion.

Assess reaction to key moments: Campaigns looked to engagement metrics on

social media to understand public reaction to media events like debates or cam-
paign speeches. The news media have been criticized for their use of opt-in on-

Downloaded from https://academic.oup.com/poq/article/84/S1/236/5871904 by guest on 10 June 2025

line polls to determine whom the public thinks “won” a debate, but even these
ill-conceived measures rarely shed light on what particular moments in an
hours-long debate resonate with the public. In this context, campaigns learned
from social media what notes struck a chord with the public—and adapted
their campaign communication and strategy accordingly. Jack Minor, deputy
digital director for the Ted Cruz campaign, noted:

One of the debates where Cruz had kind of been quiet … and then just
came out of the gun. It’s when he turned and—it was the CNBC de-
bate. I think it’s [Chris] Harwood, and [Cruz] just turned to him and be
like “These questions are ridiculous. Ben Carson, are you a homophobe?
Donald Trump make fun of Jeb Bush.” When he went on that rant, that
was like our pop of it for the campaign. That showed to Facebook …
I think that was a good measure of, okay, bashing the media works. It was
certainly a pivot for our campaign.

This debate moment—clipped and shared “almost immediately” on

Facebook—garnered engagement numbers that helped convince the Cruz cam-
paign to double down on attacking the media. Information like this is helpful
for campaigns, but it offers not just temporal specificity but also a particular
audience. Facebook users are not representative of the voting public—yet
quantified reactions from engaged constituents on social media did serve as
tools to gauge public opinion for campaigns, especially during media events.

A/B message testing on social media: Almost all the campaigns took to social
media to engage in A/B message testing, where they tested multiple versions
of messages on segmented audiences to determine the most effective variation.
In particular, campaigns saw Facebook as a platform on which they could get
more reliable measures of public opinion, as compared to sites with smaller
and more niche users like Twitter or Reddit. As Gary Coby, director of digital
advertising and fund-raising for Donald Trump’s general election campaign,
recounted:

There was a time at the end of summer where the president [then candi-
date Trump] did four policy speeches in like a week and a half period,
very short period. That was actually right after we just proved through
Campaign Use of Social Media as Public Opinion 243

our [Facebook] measurements that Trump talking about issues and vi-
sion and [how] he thinks these things are going, the whole message and
positive approach, moved every single demo [demographic] in our dir-
ection versus talking about Hillary and talking negative. That immedi-
ately spurred the campaign to act and he went and did a bunch of policy
speeches. I’m sure they were already working on those things. I don’t
think this was the sole decision to get that to happen, but I’m just saying

Downloaded from https://academic.oup.com/poq/article/84/S1/236/5871904 by guest on 10 June 2025

that’s one way that we saw the data was saying X and then we acted.

As Coby points out, assessments of public opinion via Facebook were not
limited to the blunt subject of that site’s broad user base. Because of the data
available about users of the site, campaigns were able to see how messages
resonated with specific audiences, allowing them to create ads designed to
appeal to particular groups of users. Campaigns then targeted these groups
of users and assessed engagement metrics from these segmented groups on
Facebook.

Examining opponents’ metrics: Campaigns also monitored what worked for

other campaigns by assessing opponents’ social media metrics. Echoing what
professionals from other campaigns said, Hector Sigala, social media director
of the Bernie Sanders campaign, described auditing competitors’ social media
metrics:

So their profile on Facebook, on Twitter—we were just able to ana-

lyze dozens of people from the Democratic side, the Republican side
and see what other people were doing that was working well that we
weren’t [doing] or what we were doing that was working well and try
to find out why it wasn’t working for other people so we can kind of
stick with that.

In looking at “what worked” for opponents vis-à-vis social media data, cam-
paigns extended their measurement of public opinion beyond the reach of their
own following.

QUALITATIVE DATA, INSTRUMENTAL USES

What campaigns learned about the public through metrics was also shaped
by qualitative readings of opponents’ social media posts. The above example
of campaigns using quantitative metrics to infer public reception to messages
from other campaigns was paired with qualitative impressions about how
these different strategies worked. “It’s like, if Ted Cruz had a really good day,
we would go and see … look at all of social media and see what he was doing
that day,” said Sigala, of the Sanders’s campaign.
244 McGregor

“Taking the temperature of the room”: For all the metrics available, qualita-
tive and routine readings of social media also shaped how campaigns under-
stood public opinion. One theme, repeated verbatim across interviews, was
that social media allowed campaigns to “take the temperature of the room.”
Some of this was informed by quantitative metrics, but campaign profes-
sionals also regularly read things like Facebook comments and Twitter replies
to take the pulse of public opinion. “We would use social media, obviously,

Downloaded from https://academic.oup.com/poq/article/84/S1/236/5871904 by guest on 10 June 2025

to take the temperature of the room,” said Chris Maiorana, digital strategist
and chief technology officer on the Mike Huckabee campaign. The campaign
professionals I spoke with described routine reading of social media posts. In
some cases, these tasks were assigned to specific members of the digital team,
who produced reports that were shared with other teams on the campaign. As
Vince Harris, digital director of the Rand Paul campaign, described:

Most campaigns spend a lot of time looking and replying, and they will
use their digital firms like my company to be replying back and com-
municating with people. I think a lot of times that the politicians get
obsessed with the comments too, and they will disproportionately put
emphasis in the campaign on how people are commenting back and what
people are saying.
Now that said, if you’re actually bringing people onto your Facebook
page who are tied to the voter file and they’re your type of voters, and
they’re the ones commenting then you should care about what they’re
saying. If you built a Facebook page properly, and you know that the
people commenting are real voters, and they’re not trolls or something,
then you should pay attention to what those people are saying as a form
of public opinion.

The first part of this quote echoed what I heard from other professionals: Often
candidates themselves read comments and replies, which shaped their under-
standing of public opinion. The second part reveals not only the sophistication
with which some campaigns attempted to build their social media bases,2 but
also that any feedback from these pages is necessarily shaped by the cam-
paigns’ own efforts to fill “the room” with targeted voters, all while appearing
to the public to be a site of grassroots support for the candidate.
Campaigns likened social media “listening” to the sorts of impression-
istic interactions with constituents that have long shaped what politicians
view as public opinion. Campaign professionals read comments and replies
daily, which shaped instrumental work like converting people to volunteers
or garnering donations, but also overall impressionistic views of how their

2. It should be noted that anyone can “like” and “follow,” and thus interact with, the Facebook
page of a public official.
Campaign Use of Social Media as Public Opinion 245

candidate was doing in the race. Many also mentioned that these readings in-
formed, or reinforced, digital strategy. As Coby, of Trump’s general election
campaign, said:

So, [to find out] what [people] care about, honestly, one, you listen to the
conversation and social plays a role in there. It’s helpful to read through
comments … That data is extremely helpful to inform what people care

Downloaded from https://academic.oup.com/poq/article/84/S1/236/5871904 by guest on 10 June 2025

about, what they’re worried about. We have staff where that’s their whole
job, to a) respond to those, b) receive those, and c) put together reports
to help inform us.

Understandings of public opinion gleaned from reading posts, comments, and

replies were shared alongside metrics up the campaign command chain in
meetings.

Listening to supporters: Like the listening tours or small meetings described

in Herbst’s and Fenno’s accounts, my interviews reveal that social media
provided candidates and campaigns access in particular to their supporters’
opinions, which informed messaging strategies. As Sigala, of the Sanders
campaign, said:

You could [check in] on Facebook and Instagram and everywhere else
and Reddit too, you can kind of gauge how your supporters are feeling
about stuff … And that was really one of the best gauges, now that I think
about it. When we put something out, or we’re getting feedback or even
ideas or planning a big roll out of something.

Several campaigns reached out to groups of “hardcore” supporters on social

media, via Reddit pages or Facebook pages or lists maintained by the cam-
paign, to test messaging among their base. Campaigns also paid particular at-
tention to conversations in those groups for indications as to what issues their
supporters cared about deeply.
Campaigns perceived that by using social media to identify issues that
were important to supporters, and then acting on them or communicating
support, they could improve their standing in the race. As Maiorana of the
Huckabee campaign recalled, social media readings were used “… to find the
hot button issue that will allow you to move up in the field, really, on a daily
basis.” Reading social media messages from supporters informed campaign
messaging, as well as strategy more broadly.

QUALITATIVE DATA, SYMBOLIC USES

Public opinion is a powerful rhetorical force. Qualitative forms of social media

data—individual posts, comments, or replies—were shared by campaigns to
signify their popularity.
246 McGregor

Communicating support and enthusiasm: As noted earlier, campaigns were

well aware of their ability to reach loyal supporters through various social
media outlets. From these groups of supporters, campaigns identified “influ-
encers”—those supporters with a large social media following and/or influen-
tial network—to amplify messages favorable to the campaign. Chris Georgia,
digital director of the Jeb Bush campaign, spoke about the campaign’s de-
ployment of this form of public opinion, noting its particular utility in the

Downloaded from https://academic.oup.com/poq/article/84/S1/236/5871904 by guest on 10 June 2025

absence of adequate funding to purchase social media ads: “We had a guy
during the debates that was his only job. He would pull the best social media
posts whether it was from us or others about us, and then we’d send it to our
team, our digital volunteers to have them echo it, retweet it, share it, all that.”
Campaigns routinely re-shared, or signal-boosted, these positive posts from
the public, gathered across various social media, to symbolically evoke their
candidate’s popularity.
But in the same way, campaigns benefited from and had to push back against
news coverage that was based on social media posts. Journalists use social
media to represent public opinion in their reporting (McGregor 2019; Lukito
et al. 2020), including bolstering claims that Sanders—whose campaign re-
ceived the most positive coverage among primary candidates (Patterson
2016)—had a particularly enthusiastic base. But the Sanders campaign ac-
tively worked to influence their supporters’ social media posts, which were
amplified in press coverage. As journalists used social media to bolster claims
of voter enthusiasm for Sanders in the primary period, Hillary Clinton staffers
perceived general election coverage to then convey that voters were not enthu-
siastic for their candidate. Christina Reynolds, deputy communications dir-
ector for the Clinton campaign, describes the dynamic:

Donald Trump supporters were thought to be enthusiastic. Ours were not.

I always found that frustrating, and I think some of that comes from so-
cial media ... I don’t think it comes from being on the ground.
Yes, our crowd size was not always the same … We had targeted some
things where we weren’t always looking to do big rallies, but they [jour-
nalists] used anecdotal stuff and their view of what was happening online
to throw this at our supporters that I always just felt was sad and not fair.

Reynolds goes on to say that the Clinton campaign actively worked on social
media to push back against what they perceived as journalists’ rhetoric about
the lack of enthusiasm among Clinton supporters, gleaned from social media.
These tactics involved sharing posts from supporters, but in the end, Clinton
staffers expressed frustration that they were not able to convey this effectively.

Managing public comments: Campaigns actively worked to manage what

replies or comments looked like across social media platforms. For some
Campaign Use of Social Media as Public Opinion 247

campaigns, this meant deleting comments—some for outright incivility, but

also sometimes for not being supportive. Georgia, who worked on the Bush
campaign, explained: “Whenever you do that [like a comment from a sup-
porter] it pops that comment up to the first one at the top of the comments so
that way there was no negative comment that was going up there. There was
something positive.”
Campaigns engaged in replies or comments to shape the nature of social

Downloaded from https://academic.oup.com/poq/article/84/S1/236/5871904 by guest on 10 June 2025

discussions about the election. Campaigns inserted counterpoints, especially
to negative comments, in attempts to steer the conversation substantively, but
also because they perceived that a comment with a reply from the candidate’s
own Facebook page would be prioritized by the Facebook algorithm, such
that more people would then see it. As Coby, of the Trump campaign, said:
“…social media is just where the conversation is happening now. Being a
part of that, listening to it, and helping to stimulate it are just must-haves for
a campaign.”

QUANTITATIVE DATA, SYMBOLIC USES

Shaping public opinion: In part because of the extent to which journalists rely
on social media to report public opinion (McGregor 2019; Lukito et al. 2020),
campaigns used lists of influential supporters to manipulate what public
opinion looked like. For example, the Sanders campaign had a strong working
relationship with a large sub-Reddit group of Sanders’s supporters (see also
Penney 2017). The campaign would reach out directly to these influential and
active supporters in order to shape social media metrics, such as to get a par-
ticular hashtag trending. As Sigala, from the Sanders campaign, noted: “And
that’s one of the instances where we were able to tap into the network of
supporters and be like, ‘Hey guys, I know this really big thing happened and
there’s a lot of hashtags that are trending right now about Bernie and how
we’re are still with him, but let’s get back to the issues and let’s talk about this
Cost of Living Adjustment.’”
The Trump campaign adopted a similar method during the general election
period. From their extensive text message list, people’s phone numbers were
matched with their social media profiles. The campaign then identified the top
10 percent of users with the highest influence in terms of follower counts and
their ability to drive social media conversations. These users were invited to
join the campaign’s rapid response team, “The Big-League Trump Team.” As
Coby, of the Trump campaign, said:

Basically, during the debate we were sending a text out to this small
group every three to five minutes. Those pieces of content we sent out,
were getting up to 500 percent additional increase in reach [on social
media] relative to everything else. Trump had a big footprint, but then we
were behind the scenes kind of putting gasoline on all of that.
248 McGregor

This shaping of public opinion is particularly important, as the press relies

on social media metrics—like a hashtag being trending—as an indicator that
a particular issue was important to the public (Freelon et al. 2018). Hashtags
motivated news coverage, yet these metrics were often directly influenced by
campaigns and did not organically materialize among groups of voters, as
press coverage intimates (McGregor 2019).

Downloaded from https://academic.oup.com/poq/article/84/S1/236/5871904 by guest on 10 June 2025

Communicating popularity: Campaigns communicated social media metrics
as indicators of popularity, especially in the wake of media events like debates,
when campaigns perceived that more of the public was engaged, lending
added symbolic weight to metrics. Tim Miller, director of communication for
the Bush campaign, explained:

Around the debates, there also was a kind of interest in this data, and
pushing positive data points out to the media, social media data points
out to the media, impacted the post-debate who won, which was just as
important as the actual debate itself. … We were able to quickly turn
around information that said, hey, the biggest moment of the night on
Google was when Jeb said this. That gave these pundits, who were just
looking for anything to justify their opinion and sound smart, a data point
that they could use. That was another way that we used Google, and
Facebook, and Twitter.

These metrics shaped how journalists evaluated candidate performances

(McGregor 2019), and campaigns sought to influence that by highlighting fa-
vorable metrics. Sigala, of the Sanders campaign, said, “I remember our first
debate, and it turned out that everyone was like, Bernie didn’t exactly win
this debate, but he definitely won the internet … We had the most Twitter
retweets and most Twitter questions and most Twitter mentions and Facebook
and our website just took off—this happened every single debate.” But public
opinion on social media did not match electoral preferences—Bernie Sanders
may have “won” the internet, but Hillary Clinton won the Democratic pri-
mary. As Reynolds, of the Clinton campaign, noted, “You use it [social media
metrics] as sort of an indicator. It’s still a bit of an artificial indicator, and it
was sometimes frustrating for us, because certainly during the primary Bernie
supporters were generally speaking more active on Twitter than some of our
supporters. I think it sometimes skewed the views, or at least cooked the books
a little bit.”
Campaign professionals described the work that went into building a strong
following for their candidate on social media. Even though they found it more
helpful to target whom they perceived to be reliable or persuadable voters,
the raw follower counts on social media were important for their ability to
symbolically communicate popularity. Journalists report quantitative metrics
Campaign Use of Social Media as Public Opinion 249

such as the number of people who like a candidate’s Facebook page and the
number of people who follow a candidate on Twitter as indicators of popu-
larity (McGregor 2019). But, as Minor, of the Cruz campaign, noted, these
numbers are subject to manipulation:

For the candidates, [having a large social media following] is a huge

messaging point. It’s something to be boastful about. They used that as a

Downloaded from https://academic.oup.com/poq/article/84/S1/236/5871904 by guest on 10 June 2025

count of “their movement,” their universe, their audience when they were
doing media hits. I think it certainly is about the quantity, but it’s also
about the quality of those folks. You can fudge those numbers quite a bit,
especially on Twitter.

No one I spoke with talked about artificially swelling their counts, though it is
possible to “buy” followers, made up of fake accounts. In campaign speeches
and during debates, Trump often touted his Twitter following, though reports
have indicated that a significant portion of those are fake, or bots (Bialik
2016). Real or artificial, follower counts do not accurately represent political
popularity—both Trump and Ben Carson had healthy large social media fol-
lowings long before they began their political careers.

Discussion
Throughout the course of the 2016 US presidential election, campaigns used
social media to understand and communicate public opinion. I mapped these
uses onto a theoretical model (see figure 2) that accounts for the types of data
and methodologies used, as well as for what purpose the data was used.
Using social media to do things like check in with supporters and deploy in-
fluencers to shape what social public opinion looks like is a new development
in the relationship between campaigns and supporters. On the one hand, cam-
paigns’ use of social media to understand and communicate public opinion
means that at least part of the public stands to play a more direct role in shaping
campaigns’ perceptions of the public. Some might see this as more evidence
of the role that digitally enabled citizens play in politics, “fostering openness
and grassroots participation” (Chadwick and Stromer-Galley 2016, p. 285).
But this study finds less interaction and more top-down manipulation of social
media posts and metrics for symbolic ends—taking up the work of supporters
to symbolically speak for the public on behalf of the candidate. From the citi-
zens’ perspective, they are encouraged into “structured interactivity” (Kreiss
2012) or “controlled interactivity” (Stromer-Galley 2019), their social media
posts and profiles deployed strategically mostly on the campaigns’ terms.
The use of individuals’ posts and interactions on social media as powerful
symbols of public opinion further complicates the relationship between
elites and everyday citizens. As campaigns deploy confederate influencers to
250 McGregor

Downloaded from https://academic.oup.com/poq/article/84/S1/236/5871904 by guest on 10 June 2025

Figure 2. Campaign uses of social media as public opinion in the 2016 US
presidential election.

symbolically communicate public opinion, then even true grassroots struc-

tures appear—to the consuming public—to be one and the same.
The use of social media as a symbol of public opinion further highlights the
hybrid logic of the political landscape. Campaigns not only instrumentally use
social media metrics and posts gleaned from the public, but actively work to
shape them in their favor. This powerful new symbol of public opinion then
drives and informs press coverage. Journalists use social media posts—some
shaped or amplified by campaigns—as a form of new vox pop in election
coverage (McGregor 2019). The press also uses social media metrics—
shaped purposively by campaigns—to position candidates in the horse race
(McGregor 2019). Campaigns’ use of social media to represent public opinion
is not designed solely to “go around” the press, but also to influence the press,
taking advantage of journalists’ newfound source of public opinion fodder.
Opinions on social media are widely manipulable. This study describes
how political professionals attempt to shape public opinion on various social
media sites, including through the use of volunteer supporters. In the wake
of the 2016 election, we’ve learned about the scale of the Russian campaign,
through social media, to influence the election (Jamieson 2018; Kim et al.
2018). Social media accounts, set up by Russian trolls, were counted among
the quantified measures of public opinion campaigns used to understand and
represent American public opinion, also appearing in news stories that used
social media to report public opinion (Lukito et al. 2020). Homegrown ma-
nipulation of social media metrics abounds, as part of the far-right tactic of
“attention hacking”—and campaigns’ (and journalists’) reliance on these
Campaign Use of Social Media as Public Opinion 251

metrics for instrumental and symbolic purposes make them vulnerable to ma-
nipulation (Marwick and Lewis 2017).
Even if it were possible to weed out direct manipulation and bad actors
from social media, the universe from which this new form of public opinion is
drawn would still not mirror the population. Campaigns used social media to
“take the temperature of the room”—but the room only consists of the small
number of Americans who are vocal about politics on social media. This

Downloaded from https://academic.oup.com/poq/article/84/S1/236/5871904 by guest on 10 June 2025

smaller chorus of voices tends to be more polarized: Democrats who post
on social media are more likely to be progressive and liberal, as compared
to the larger share of moderate Democrats who don’t post on social media
(Cohn and Quealy 2019). In general, Twitter users in the United States are
younger, more likely to identify as Democrats, more educated, and wealthier
than US adults overall (Wojcik and Hughes 2019). There is also significant
variation across social media sites, not only in their user bases, but in the
genres of communication present on them (Kreiss, Lawrence, and McGregor
2018). Social media may present a more polarized sense of the public, where
incivility is rampant (e.g., Kushin and Kitchener 2009; Groshek and Koc-
Michalska 2017; Sydnor 2018).
Of course, as I show here, a public opinion constructed from social media
can and likely always will be manipulated by those in power—and with louder
voices—even though they may not represent the average American opinion.
This does not mean we should dismiss social media as an important arena to
understand public opinion. In fact, unlike private polls, social media offers a
window into expressive hierarchies in society—“their louder voice is simply a
superstructural indicator of what lies beneath” (Herbst 2011, p. 93).
For all its downsides, the use of social media to understand the public makes
available—at scale—qualitative data stemming from individuals’ conversa-
tions about politics. As Doris Graber (2006) suggests, closed-ended questions
about specific, contrived, and narrowly framed questions prevent respondents
from selecting and expounding upon their views of the issues that they deem
important. Social media platforms loosen constraints around deliberation, pro-
viding relatively open spaces where members of the public may express their
views and enter into dialogue with one another largely on the basis of their
own priorities, bringing to the table the information, arguments, and opinions
they believe are relevant. In this sense, social media may provide particularly
fertile ground for gathering and gauging fuller, richer, and less constricted
citizen input. Campaigns looked to social media posts to “take the tempera-
ture of the room”—these readings bring qualitative and less-structured opin-
ions into the practice of studying public opinion. But campaigns did so with
relatively little consideration for the significant caveats of these data—of the
particularities of individuals who publicly engage with politicians and politics
on social media.
Elite use of social media to represent public opinion may bring political
elites closer to understanding the public as engaged, and thus prizing (social
252 McGregor

media) conversations in a Deweyan sense (Dewey 1954), forefronting public

deliberation. Social media offers the public a means to connect and converse
about politics, as well as to communicate their needs and concerns to elites.
This may be particularly true during media events like debates, where cam-
paign professionals saw increased value in learning about the public vis-à-vis
social media, as they perceived that more of the public was more engaged.
Campaigns’ use of social media to understand public opinion prizes certain

Downloaded from https://academic.oup.com/poq/article/84/S1/236/5871904 by guest on 10 June 2025

types of people and certain types of engagement. For a multitude of reasons,
many people—especially those from historically marginalized groups—do
not feel comfortable or even safe expressing their political views online or
offline (Van Duyn 2018).
This study lays out a model for how campaigns use social media as public
opinion, providing scholars an organizing mechanism to examine the conse-
quences of such work. For example, campaigns using qualitative readings of
social media posts for instrumental purposes to infer broader public opinion
are at risk of mischaracterizing the public. The model brings much-needed at-
tention to qualitative data, a novel aspect of social media in understanding and
representing public opinion. For the better part of a century, studies of public
opinion have focused predominantly on surveys. The use of social media data
to understand the public, with all its problems of representativeness, may
provide a retort to long-standing criticisms of surveys—specifically that sur-
veys do not reveal the hierarchical, social, public, or conversational (Blumer
1948; Herbst 1998; Cramer 2016) aspects of public opinion formation. This
model highlights a need to explicate what can—and cannot—be understood
about public opinion. Scholars of polling have documented the many forces
that shape survey answers: question form and order, interviewer effects, and
mode effects. Examining the uses laid out in this model highlights the need for
scholars of public opinion to engage with communication and media studies,
which offer much-needed context about how the structures, affordances, and
cultures of social media sites shape public political expression on them. This
study is limited in that it relies on interviews structured around a single case
study—the 2016 US presidential election. It is my hope that this study is gen-
erative—that this model may be the starting point for more in-depth exam-
inations of particular ways in which political elites use new forms of data to
understand and represent the public.
Chadwick (2017, p. 45) describes the shifting dynamics of power in the
hybrid media system, in which “actors create, tap, or steer information flows
in ways that suit their goals and in ways that modify, enable, or disable the
agency of others, across and between a range of older and news media set-
tings.” This study shows how social media posts by the public are consumed
and rebroadcast by political elites to understand and communicate public
opinion, exemplifying the hybrid media system. These practices may appear
to bring the public closer to politics, but instead allow elites more control over
the substance of public opinion to their advantage.
Campaign Use of Social Media as Public Opinion 253

Appendix A. Campaign Professionals Interviewed

Name Title Campaign Date(s)

Chris Georgia Digital director Bush 7/13/2016
6/14/2017

Downloaded from https://academic.oup.com/poq/article/84/S1/236/5871904 by guest on 10 June 2025

Tim Miller Director of communication Bush 6/9/2017

Chris Maiorana Digital strategist & chief Huckabee 7/5/2016

technology officer 6/13/2017

Jordan Powell Deputy campaign manager Huckabee 7/12/2016

6/7/2017

Jack Minor Deputy digital director Cruz 7/5/2016

6/21/2017

Chris Wilson Director of research, Cruz 6/22/2017

analytics & digital strategy

Vincent Harris Digital director Paul 7/12/2016

6/27/2017

Matt Compton Deputy digital director Clinton 2/13/2017

6/27/2017

Christina Reynolds Deputy communications Clinton 2/7/2017

director 6/23/2017

Jenna Lowenstein Digital director Clinton 3/28/2017

Kevin Bingle Digital director Kasich 6/16/2017

Hector Sigala Social media director Sanders 7/13/2016

6/16/2017

Gary Coby Director of digital adver- Trump 6/20/2017

tising and fund-raising

Appendix B. Campaign Professional Interview Protocol

1. What position did you serve on the XX campaign?

2. What did the campaign rely on to understand and assess public attitudes?
3. How did the campaign gather data on public attitudes?
4. Did the campaign distinguish between voters, or likely voters, and the
larger public when assessing the role of social media in the campaign?
a. If so, how? And why?
254 McGregor

5. In your capacity as digital/social media director in the campaign, what

role did you play in communicating data about or understandings of the
public to other teams in the organization?
6. What was the campaign’s attitude toward social media?
7. What was your role with respect to social media?
8. What role, if any, did social media play in understanding the public or
voters?

Downloaded from https://academic.oup.com/poq/article/84/S1/236/5871904 by guest on 10 June 2025

a. Specifically, how was this incorporated with more traditional meas-
ures of public opinion like polling data?
9. Did you (or anyone on the campaign) read posts, tweets, comments, or
replies?
a. If so, for what purposes were those readings used?
10. Did the campaign use social media analytics?
a. If so, how were analytics used?
b.  What campaign tactics or communications were informed by
analytics?

References
Anstead, Nick, and Ben O’Loughlin. 2015. “Social Media Analysis and Public Opinion: The 2010
UK General Election.” Journal of Computer-Mediated Communication 20:204–20.
Baldwin-Philippi, Jessica. 2015. Using Technology, Building Democracy: Digital Campaigning
and the Construction of Citizenship. Oxford: Oxford University Press.
———. 2016. “The Cult(ure) of Analytics in 2014.” In Communication and Midterm Elections,
edited by John Allen Hendricks and Dan Schill, 25–42. New York: Palgrave Macmillan.
Berelson, Bernard, Hazel Gaudet, and Paul F. Lazarsfeld. 1968. The People’s Choice: How the
Voter Makes Up His Mind in a Presidential Campaign. New York: Columbia University Press.
Berelson, Bernard R., Paul F. Lazarsfeld, and William N. McPhee. 1954. Voting: A Study of
Opinion Formation in a Presidential Campaign. Chicago: University of Chicago Press.
Bergan, Daniel E., Alan S. Gerber, Donald P. Green, and Costas Panagopoulos. 2005. “Grassroots
Mobilization and Voter Turnout in 2004.” Public Opinion Quarterly 69:760–77.
Bialik, Carl. 2016. “Everyone Has Fake Twitter Followers, But Trump Has the Most. Sad!”
FiveThirtyEight (April 14). Available at https://fivethirtyeight.com/features/everyone-has-fake-
twitter-followers-but-trump-has-the-most-sad/. Accessed April 14, 2016.
Blumer, H. (1948). Public opinion and public opinion polling. American sociological review,
13(5), 542-549.
Chadwick, Andrew. 2017. The Hybrid Media System: Politics and Power. Oxford: Oxford
University Press.
Chadwick, Andrew, and Jennifer Stromer-Galley. 2016. “Digital Media, Power, and Democracy
in Parties and Election Campaigns: Party Decline or Party Renewal?” International Journal of
Press/Politics 21:283–93.
Cohn, Nate, and Kevin Quealy. 2019. “The Democratic Electorate on Twitter Is Not the Actual
Democratic Electorate.” New York Times (April 9). Available at https://www.nytimes.com/inter-
active/2019/04/08/upshot/democratic-electorate-twitter-real-life.html. Accessed April 9, 2019.
Cramer, K. J. (2016). The politics of resentment: Rural consciousness in Wisconsin and the rise of
Scott Walker. University of Chicago Press.
Dewey, John. 1954. The Public and Its Problems. Athens: Ohio University Press.
Fenno, Richard F. 1978. Home Style: House Members in Their Districts. New York: HarperCollins.
Campaign Use of Social Media as Public Opinion 255

Freelon, Deen, Lori Lopez, Meredith D. Clark, and Sarah J. Jackson. 2018. “How Black Twitter
and Other Social Media Communities Interact with Mainstream News.” Knight Foundation.
Available at https://kf-site-production.s3.amazonaws.com/media_elements/files/000/000/136/
original/TwitterMedia-final.pdf. Accessed February 28, 2018.
Gerber, Alan S., and Donald P. Green. 2000. “The Effects of Canvassing, Telephone Calls,
and Direct Mail on Voter Turnout: A Field Experiment.” American Political Science Review
94:653–63.
———. 2001. “Do Phone Calls Increase Voter Turnout? A Field Experiment.” Public Opinion

Downloaded from https://academic.oup.com/poq/article/84/S1/236/5871904 by guest on 10 June 2025

Quarterly 65:75–85.
Graber, Doris A. 2006. “Government by the People, for the People—Twenty‐First-Century Style.”
Critical Review 18:167–78.
Groshek, Jacob, and Karolina Koc-Michalska. 2017. “Helping Populism Win? Social Media
Use, Filter Bubbles, and Support for Populist Presidential Candidates in the 2016 US Election
Campaign.” Information, Communication & Society 20:1389–1407.
Herbst, Susan. 1993. Numbered Voices: How Opinion Polling Has Shaped American Politics.
Chicago: University of Chicago Press.
———. 1998. Reading Public Opinion: How Political Actors View the Democratic Process.
Chicago: University of Chicago Press.
———. 2011. “(Un)Numbered Voices? Reconsidering the Meaning of Public Opinion in a Digital
Age.” In Political Polling in the Digital Age: The Challenge of Measuring and Understanding
Public Opinion, edited by Kirby Goidel. Baton Rouge, LA: LSU Press.
Hersh, Eitan D. 2015. Hacking the Electorate: How Campaigns Perceive Voters. Cambridge:
Cambridge University Press.
Igo, Sarah E. 2007. The Averaged American: Surveys, Citizens, and the Making of a Mass Public.
Cambridge, MA: Harvard University Press.
Jamieson, Kathleen Hall. 2018. Cyberwar: How Russian Hackers and Trolls Helped Elect a
President—What We Don’t, Can’t, and Do Know. Oxford: Oxford University Press.
Jungherr, Andreas. 2015. Analyzing Political Communication with Digital Trace Data: The Role
Twitter Messages in Social Science Research. Cham, Switzerland: Springer.
Karpf, David. 2016. Analytic Activism: Digital Listening and the New Political Strategy. Oxford:
Oxford University Press.
Kim, Young Mie, Jordan Hsu, David Neiman, Colin Kou, Levi Bankston, Soo Yun Kim,
Richard Heinrich, et al. 2018. “The Stealth Media? Groups and Targets Behind Divisive Issue
Campaigns on Facebook.” Political Communication 35:515–41.
Kreiss, Daniel. 2012. Taking Our Country Back: The Crafting of Networked Politics from Howard
Dean to Barack Obama. Oxford: Oxford University Press.
———. 2016a. Prototype Politics: Technology-Intensive Campaigning and the Data of
Democracy. Oxford: Oxford University Press.
———. 2016b. “Seizing the Moment: The Presidential Campaigns’ Use of Twitter During the
2012 Electoral Cycle.” New Media & Society 18:1473–90.
Kreiss, Daniel, Regina G. Lawrence, and Shannon C. McGregor. 2018. “In Their Own Words:
Political Practitioner Accounts of Candidates, Audiences, Affordances, Genres, and Timing in
Strategic Social Media Use.” Political Communication 35:8–31.
Krippendorff, Klaus. 2005. “The Social Construction of Public Opinion.” In Kommunikation
über Kommunikation: Theorie, Methoden und Praxis. Festschrift für Klaus Merten, edited
by Edith Weinand, Joachim Westerbarkey, and Armin Scholl (Hrsg.), 129–49. VS Verlag für
Sozialwissenschaften.
Kushin, Matthew J., and Kelin Kitchener. 2009. “Getting Political on Social Network Sites:
Exploring Online Political Discourse on Facebook.” First Monday 14.
Lindlof, Thomas R., and Bryan C. Taylor. 2017. Qualitative Communication Research Methods.
Thousand Oaks, CA: Sage Publications.
256 McGregor

Luker, Kristin. 2008. Salsa Dancing into the Social Sciences. Cambridge, MA: Harvard University
Press.
Lukito, Josephine, Jiyoun Suk, Yini Zhang, Larisa Doroshenko, Sang Jung Kim, Min-Hsin Su,
Yiping Xia, et al. 2020. “The Wolves in Sheep’s Clothing: How Russia’s Internet Research
Agency Tweets Appeared in U.S. News as Vox Populi.” International Journal of Press/Politics
25(2):196–216.
Marwick, Alice, and Rebecca Lewis. 2017. Media Manipulation and Disinformation Online. New
York: Data & Society Research Institute.

Downloaded from https://academic.oup.com/poq/article/84/S1/236/5871904 by guest on 10 June 2025

McGregor, Shannon C. 2019. “Social Media as Public Opinion: How Journalists Use Social
Media to Represent Public Opinion.” Journalism 20:1070–86.
Nielsen, Rasmus Kleis. 2012. Ground Wars: Personalized Communication in Political Campaigns.
Princeton, NJ: Princeton University Press.
Patterson, Thomas E. 2016. “News Coverage of the 2016 Presidential Primaries: Horse Race
Reporting Has Consequences.” HKS Working Paper No. 16–050, available at https://ssrn.com/
abstract=2884834.
Penney, Joel. 2017. “Social Media and Citizen Participation in ‘Official’ and ‘Unofficial’ Electoral
Promotion: A Structural Analysis of the 2016 Bernie Sanders Digital Campaign.” Journal of
Communication 67:402–23.
Schober, Michael F., Josh Pasek, Lauren Guggenheim, Cliff Lampe, and Frederick G. Conrad.
2016. “Social Media Analyses for Social Measurement.” Public Opinion Quarterly 80:180–211.
Stromer-Galley, Jennifer. 2019. Presidential Campaigning in the Internet Age. Oxford: Oxford
University Press.
Sydnor, Emily. 2018. “Platforms for Incivility: Examining Perceptions Across Different Media
Formats.” Political Communication 35:97–116.
Van Duyn, Emily. 2018. “Hidden Democracy: Political Dissent in Rural America.” Journal of
Communication 68:965–87.
Wojcik, Stefan, and Adam Hughes. 2019. “Sizing Up Twitter Users.” Pew Research Center Internet
& Technology (April 24). Available at https://www.pewresearch.org/internet/2019/04/24/
sizing-up-twitter-users/. Accessed April 26, 2019.