JMG – Department of
Journalism, Media and
Communication
Effect of Populist Campaign
Communication on Citizen Engagement
in Social Media
Manjurul Mahmud
Master’s Thesis in Media and Communication
Thesis: 30 hp
Program and/or course: Master’s Programme in Political Communication/MK2502
Level: Second Cycle
Semester/year: Spring 2024
Supervisor: Isabella Glogger
Abstract
Objective: The aim of this research is to investigate the effects of populist campaign com-
munication by political actors on social media, particularly focusing on the period before an
election. The study aims to understand how these communications on platforms like Face-
book and Twitter influence citizen engagement, examining both populist and non-populist
messages.
Theory: This study integrates theories from political science, sociology, and communication
to examine the impact of populist campaign communication on citizen engagement on social
media platforms. It employs perspectives from political communication, such as the populist
dichotomy between "the people" and "the elite" , psychological frameworks like Social Iden-
tity Theory to explain group dynamics and attraction to populist messages. Additionally,
the Civic Voluntarism Model are utilized to understand people are attracted to populist
messages.
Method: The research employs a quantitative content analysis of 918 social media posts
from four political actors—two each from the United States and the United Kingdom—during
their respective election campaigns. Engagement metrics such as likes, shares, and comments
were recorded for each post. Populist messages were identified using predefined categories,
allowing the study to measure the frequency and impact of populist versus non-populist
content across platforms.
Findings: The findings of the study indicate that populist campaign communication signifi-
cantly impacts social media engagement on platforms like Facebook and Twitter. Contrary to
initial expectations, Twitter proved more effective in generating engagement than Facebook.
Additionally, government leaders employed populist messaging more frequently and received
higher engagement than opposition leaders. Populist messages consistently outperformed
non-populist content in terms of user engagement.
Keywords: Populism, Populist Communication; Social Media; Facebook; Twitter; Political
engagement; Social Identity Theory; Civic Voluntarism Model; Election campaigns; Oppo-
sition versus Government.
Acknowledgement
First and foremost, I want to extend my deepest gratitude to Isabella Glogger. I feel incred-
ibly fortunate to have had such a remarkable mentor. She is, without question, the finest
teacher I’ve ever encountered—always encouraging me with strength and academic insight,
for which I am profoundly grateful.
I also want to thank our master’s program coordinator, Nicklas Håkansson. His constant
support, whether academic or administrative, has been invaluable. I knew I could approach
him with any issue, large or small, and he would always be there with a solution. This re-
search would not have been possible without the insightful recommendations and thoughtful
guidance of Jesper Strömbäck, who helped shape my initial ideas and direction. My sincere
thanks to all the mentors at JMG, who have together crafted an academic journey I will
cherish.
When I think about my classmates, I realize how fortunate I am to have shared this
experience with such open-hearted, inspiring individuals. They broadened my perspective
at every turn, and I especially want to mention Karo, Luna, Mamiko, Laura, Tayná, and
Filip—you’re all incredible!
The journey of this thesis was brightened by countless hours spent wandering Gothen-
burg, chatting and laughing with Sabrina, Imran, and Mrinmoya. I’m so thankful for these
cherished memories. From across the sea in Ireland, Manab, and from Bangladesh, Abid,
provided constant mental support. Without you two, life would have lost its pun!
My deepest gratitude belongs to my family. They have supported me unconditionally
since childhood, and it is their love and guidance that brought me here today. Thank you to
my mother, father, and sister, for motivating me across thousands of miles. Abba, Amma,
and Bubu—I love you beyond words!
I must give special thanks to my life partner, my wife, Nafisa. Her unwavering support
is beyond description. She has been my rock through every moment of this journey, and
during the most challenging times, her insight and encouragement gave me the strength and
ideas I needed to move forward.
This thesis is also a dedication to the more than thousands brave souls in Bangladesh
who lost their lives fighting for freedom and democracy during this period. Their courage is
an inspiration to us all. You are the true heroes, and this work is dedicated to you.
Contents
Contents
1 Introduction 1
1.1 Background and Societal Relevance . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Academic Relevance and Knowledge Contribution . . . . . . . . . . . . . . . 2
1.3 Aim and Research question . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.4 Disposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2 Theoretical Framework and Literature Review 5
2.1 The Pulse of Populism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.1.1 Triangle of Populism . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.1.2 Populist Communication Strategy . . . . . . . . . . . . . . . . . . . . 8
2.2 Social Identity Theory: Why are "People" attracted? . . . . . . . . . . . . . 10
2.3 Civic Voluntarism Model: Why do "People" participate? . . . . . . . . . . . 10
2.4 Platform Differences in Populist Messaging: Facebook vs. Twitter . . . . . . 12
2.5 Populist Rhetoric in Opposition and Government . . . . . . . . . . . . . . . 13
2.6 Engagement with Populist vs. Non-Populist Messages . . . . . . . . . . . . . 14
2.7 Engagement Dynamics of Government vs. Opposition Leaders . . . . . . . . 15
2.8 Research Gap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.9 Hypothesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3 Research Method and Materials 18
3.1 Research Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.2 Choice of Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.3 Context: Two Nations Divided by a Common Language . . . . . . . . . . . . 19
3.4 Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.4.1 Sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.4.2 Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.4.3 Coding Scheme and Procedure . . . . . . . . . . . . . . . . . . . . . . 24
3.5 Reliability, Validity and Limitation . . . . . . . . . . . . . . . . . . . . . . . 25
3.6 Data Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.7 Ethical Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
4 Results 27
4.1 Descriptive and Patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
4.2 Hypothesis 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
4.3 Hypothesis 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
4.4 Hypothesis 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
4.5 Hypothesis 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
4.6 Hypothesis 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
5 Discussion and Conclusion 36
5.1 Concluding Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
5.2 Strength and Limitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
5.2.1 Strengths of the Study . . . . . . . . . . . . . . . . . . . . . . . . . . 39
5.2.2 Limitation of the Study . . . . . . . . . . . . . . . . . . . . . . . . . 40
5.3 Future Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
Bibliography 49
A Literature Search Process 50
B Codebook 51
C Inter-coder Reliablity Test and Cohen’s Kappa 57
Chapter 1
Introduction
1.1 Background and Societal Relevance
"One of the dangers of the internet is that people can have entirely different realities. They
can be cocooned in information that reinforces their current biases" Barack Obama, the
former President of the United States, expressed concerns in an interview taken by Prince
Harry of the U.K. royal family on the BBC’s Today Program on September 2017 (“Obama
warns against irresponsible social media use”, 2017). In the age of social media, this concern
is increasingly becoming an established reality (Bakshy et al., 2015). Platforms like social
media often give space to such biases of citizens, which indirectly supports ideas like populism
(Tucker et al., 2018).
One notable event to mention is the outcome of the recent election in the Nether-
lands. Geert Wilders, a prominent Dutch right-wing politician renowned for his anti-Muslim
rhetoric, posted in X, "I would like a right-wing cabinet. Less asylum and immigration.
Dutch people first. The love for my country and voters is great and more important than
my own position" as he was withdrawing from his desire to become Prime Minister. This
decision was made due to his inability to garner the requisite support from all coalition par-
ties, even though the Freedom Party (PVV) which Wilders led secured the majority of votes
in the 2023 Dutch general elections (“Geert Wilders abandons bid to become Dutch prime
minister”, 2023). This kind of populist rhetoric made him popular in the Netherlands for a
long time Vossen, 2011.
The rise of populist parties is not confined to the Netherlands. In the most recent national
election in Sweden, there has been a notable increase in support for populist parties (Aylott
and Bolin, 2023). Similar trends have been observed in Italy, Hungary, and Poland according
to Pew Research Center. In these countries, populist parties have gained significant traction
(Silver, 2022).
Although social media is a phenomenon of the new era, populism is not a recent issue at
all. Since the 1990s, populism has steadily increased in Europe (Kaltwasser, 2012). Mudde
(2004) believes that populism has become so common in Western democracies that it now
represents a ’populist spirit’. Two examples in this regard are particularly noteworthy. The
first is UK’s exit from the European Union in 2016, known as Brexit, and the second is
the Capitol attack in US by Donald Trump’s supporters in 2020. D. Smith et al. (2021)
explains that during the 2016 UK EU Referendum campaign, top politicians used populist
1
language to gain political advantage by presenting themselves as rebels. Furthermore, the
Capitol attack was also caused by "populism," which was an extraordinary attack on Ameri-
can democracy (Crothers and Burgener, 2021). These activities are a threat to a democratic
society. Since populism is popular and people-centric, it gives an impression of being demo-
cratic. Ruth-Lovell and Grahn (2022) mentioned that while both populism and democracy
involve the people, the connection between populist ideas and electoral democracy is some-
what ambiguous. Empirical evidence presented by them indicates that populist governments
tend to undermine electoral, liberal, and deliberative democracy (Tucker et al., 2018). To
understand society and fully grasp the current political situation in the world, it is crucial
to know about populism and its impact on the people. Since social media is a one of the
major media of communication in 21st century between citizens and politicians (Chadwick,
2013), such research should be conducted on social media.
In this thesis, it will be analyzed how politicians do their populist communication through
social media and how citizen engage those populist rhetoric.
1.2 Academic Relevance and Knowledge Contribution
Researchers across diverse fields, such as sociology, psychology, communication studies, and
political science, are working to examine how populism influences social media dynamics.
An important point here is how citizens respond to the populist messages of political leaders
which leads them to engage in various activities on social media. Therefore, there are op-
portunities for new ideas in various research fields. This approach allows to explore different
aspects of populism and populist communication. If it is possible to understand the rela-
tionship between citizens and populist political actors on social media, then it will explore
different avenue of mentioned research field.
Firstly, understanding political mobilization is crucial in political science because it re-
veals how political actors and organizations engage and activate voters, influencing election
outcomes and shaping public policy (Rosenstone and Hansen, 1993). Mobilization strategies
can significantly impact voter turnout and democratic participation, providing insights into
the functioning and health of democratic systems (Verba et al., 1995).
Secondly, studying social media platforms is essential in media and communication re-
search because these platforms have transformed how information is disseminated and con-
sumed (Hermida, 2010). They play a critical role in shaping public discourse as well as
influencing political opinions, and mobilizing social movements, thereby altering traditional
media landscapes (van Dijck, 2013).
Thirdly, understanding political campaigns is vital as they are the primary means through
which political candidates communicate their messages, mobilize support, and ultimately
influence voter behavior (Norris et al., 1999). Campaign strategies and communication
methods can determine the success of political actors and affect the democratic process
(Iyengar and Simon, 2000).
Fourthly, understanding citizen behavior is crucial for identifying the factors that mo-
tivate individuals to participate in political processes. This knowledge helps in developing
strategies to enhance civic engagement and democratic participation, ensuring that diverse
voices are heard in the political arena (Tajfel and Turner, 1979).
2
Finaly, it is very crucial how populism, social identity theory and civic voluntarism model
are intertwined. Studying populism and populist communication is necessary to grasp how
these movements influence political landscapes and voter sentiments. Populist rhetoric often
appeals to emotions and identities, making it a powerful tool for mobilizing support and
challenging established political norms (Mudde, 2004). Additionally, Social Identity Theory
is essential for understanding how group memberships influence individual behaviors and
attitudes. This theory explains the psychological mechanisms behind group dynamics, prej-
udice, and intergroup conflict, which are crucial for addressing social and political issues
(Tajfel and Turner, 1979). The Civic Voluntarism Model is important for identifying the
resources, engagement levels, and networks that drive political participation. It highlights
how resources and social influences affect individuals’ likelihood of participating in civic
activities, providing a framework for enhancing democratic engagement (Verba et al., 1995).
1.3 Aim and Research question
The aim of this research is to investigate the effects of populist campaign communication by
political actors on social media, particularly focusing on the period before an election. The
study aims to understand how these communications on platforms like Facebook and Twitter
influence citizen engagement, examining both populist and non-populist messages. This
research seeks to explore the strategies used by political actors during election campaigns
and how these strategies impact voter behavior and engagement.
This study aims to provide a comprehensive understanding of the dynamics between
social media populist communication and citizen engagement by addressing the following
research questions:
Research Question 1
Which social media platform is most effective for populist messages in terms of user engage-
ment?
Sub-questions:
• What is the predominant social media platform utilized by political actors for sharing
populist messages?
• Which platform, Facebook or Twitter, facilitates greater user engagement for political
messages?
Research Question 2
How do political actors’ roles influence the frequency and engagement of populist messaging
on social media platforms?
3
Sub-questions:
• How do opposition political actors compare to those in government regarding the fre-
quency of populist messaging on social media?
• Do populist messages on social media receive higher user engagement (e.g., likes, shares,
and comments) than non-populist messages?
• Does social media activity by government leaders lead to higher user engagement com-
pared to that of opposition leaders?
1.4 Disposition
In the second chapter, the theoretical framework and literature review of the thesis will
be outlined. The discussion will start with an exposition of populism and populist com-
munication. Then it will investigated why people are attracted and participate to populist
communication which will connect the Social Identity Theory (SIT) and Civic Voluntarism
Model (CVM). Overall, the theoretical crossword with social media’s role in shaping the
research objective will be explained in this chapter.
The focus of the third chapter will be on the methodology employed in conducting the
research. Detailed attention will be given to the materials, variable, measures, and statis-
tical techniques utilized in the study. The chapter will also articulate the epistemological
perspective and reflections on the chosen methods along with the quality criteria that guided
the research process.
In the fourth chapter, the findings of the study will be presented and interpreted. The
results will be discussed in depth, highlighting significant patterns and trends that emerged
from the data. The final chapter will synthesize the insights gained from the research. The
results will be concisely summarized, followed by a discussion of the study’s limitations and
contributions to the field. The chapter will conclude with suggestions for future research
and identifying potential avenues for further exploration based on the findings of this study.
4
Chapter 2
Theoretical Framework and Literature
Review
To achieve a theoretical understanding of the effect of populist campaign communication on
citizen engagement, this chapter will present a theoretical framework where central concepts
will be discussed and described. See appendix A to understand these Literature has been
chosen.
2.1 The Pulse of Populism
Contemporary populism can trace its origins back to the United States’ Farmer’s Alliance,
an economic movement initiated in the 1870s (Goodwyn, 1976). This movement aimed to
improve the precarious conditions faced by farmers following the American Civil War but
following the failure to its economic objectives, the alliance transitioned into a political
entity (Goodwyn, 1976). By advocating for radical agrarianism and anti-elitism, this party
ostensibly represents the inaugural populist party of modern-days (Wettstein et al., 2016).
However, it is very hard to define the term populism (Priester, 2011) because populism
changes with context, leading to ambiguity in its definition (Wettstein et al., 2016).
There are various opinions on the definition of populism, but the most significant ap-
proach has been presented by Mudde (2004) in his published research paper, "The Populist
Zeitgeist." There was a time when it was thought that populism was perhaps associated only
with the right-wing. However, Mudde (2004) has shown that, while historically associated
with the right, in recent decades, populism has been embraced by both the radical right
and the left. This demonstrates the adaptability of populist rhetoric across the political
spectrum. In this regard, another assumption was that populism might exist only among
marginal parties. According to Bartels and Remke (2023), marginal parties are political
groups that operate on the periphery of mainstream parliamentary activities. But today,
populism is not just found in these marginal parties but has become a part of mainstream
political discourse (Mudde, 2004). Political leaders from mainstream parties often adopt
populist tactics and language to appeal to broader electorates, reflecting populism’s deep
infiltration into mainstream politics (Akkerman et al., 2016) . Mudde (2004) has also shown
that the influence of populism is not restricted to Europe but is a global phenomenon seen
in various forms in North America, Australia, and New Zealand as well. This widespread
5
presence underscores its relevance and adaptability to different political and social contexts.
According to Mudde (2004), the rise of populist movements has been facilitated by
changes in media landscapes, including the decline of traditional party-controlled media
and the rise of more independent and commercialized media platforms that provide a stage
for populist messages. He defines populism as an ideology that perceives society to be ul-
timately separated into two homogeneous and antagonistic groups: ’the pure people’ and
’the corrupt elite.’ It argues that politics should be an expression of the volonté générale
(general will) of the people. This definition emphasizes the dichotomous view populists hold
regarding society and their moralistic framing of politics as a battle between good (the peo-
ple) and evil (the elite). According to him, it is a thin centered ideology. The thin-centered
ideology is a fragmentation of ideology compare to thick ideology which is not comprehen-
sive or fully developed in its own right but instead relies on a limited set of core concepts
(Freeden, 1998). As flexibility of populist rhetoric, allowing it to adapt to various political
contexts while maintaining a core ideological simplicity it can be defined as thin centered
ideology (Stanley, 2008).
Five different approaches of populism have been identified by Moffitt and Tormey (2014).
First one is about Muddes definition. Second, Laclau (2005) argued that it a “structuring
logic of political life.” In the next definition, populism is seen as discourse. It is based on
two assumptions. The first assumption indicates that society is divided into two parts peo-
ple by the discourse created by populism. It is relatively similar with the concept which
identifies populism as an ideology. The second assumption, as studied by Jagers and Wal-
grave (2007), looks at how we can use qualitative methods to measure how much populism
is expressed in written texts. Furthermore, populism is seen as strategy or an organization
where political actors executes power based on direct support from followers. Jagers and
Walgrave (2007) have defined it as a political communication style. According to Moffitt
and Tormey (2014) understanding populism as a political style requires examining how the
communication strategies of populist leaders influence their supporters.
To classify populism, all of these five approaches can be used. However, in this thesis,
populism is used as a political communication style which is based on some populist com-
munication strategies. The various aspects of populist rhetoric discussed in the following
chapters support the idea that populism views society as divided into two distinct groups.
These two groups are mainly divided by political actors. So, three different group are related
to populist communication. The people and its counter part the elites as well as the political
actors. In this paper, this three groups are altogether seen as triangle of populism. So, it is
needed to understand every angle of this triangle.
2.1.1 Triangle of Populism
First angle is about the people which will eventually be denoted as citizen in following
chapters. In discussions of populism, the notion of "the people" is often considered central
concept as it aligns with the the fundamental principle of sovereignty that populist ideologies
frequently use (Taggart, 2000). Jagers and Walgrave (2007) argue that populism inherently
connects with "the people". The core of this ideology is the belief that the people hold the
right to ultimate sovereignty which is a concept at the heart of most populist frameworks
(Stanley, 2008). This emphasis on popular sovereignty allows populism to invoke the au-
6
thority of "the people" as the legitimate force in a democracy (Canovan, 1999). Typically,
populist rhetoric portrays "the people" as a unified and homogeneous group (Hameleers
et al., 2017). It is often described as inherently virtuous and morally superior (Stanley,
2008). This group is imagined as a cohesive social body, united by shared values and per-
spectives, and often juxtaposed with an elite or a minority (Taggart, 2000). He suggests that
this "silent majority" possesses a common-sense wisdom which is representing the collective
voice that speaks above that of an insistent but unrepresentative minority.
However, Mudde (2004) points out that this idea of "the people" is often an artificial
or symbolic construct—a selective subset that excludes others rather than representing a
complete or realistic view of the population. The term "the people" takes on different mean-
ings depending on the specific populist context. Following the work of Canovan (1999) and
Kriesi (2013), who build on Mény and Surel (2002) classifications, three main interpretations
emerge. The first is a political conception of "the people" as a sovereign body, often ref-
erenced as the "united people" (Canovan, 1999), embodying a unified national demos that
opposes internal division (Kriesi, 2013).
The second perspective is cultural, viewing "the people" as a national or ethnic com-
munity and emphasizing exclusion based on cultural or ethnic criteria (Kriesi, 2013). Here,
populism can be seen as defending "the ethnos"—a concept focused on excluding those
viewed as foreign or marginal, such as immigrants or other cultural minorities.
The third interpretation is economic, which frames "the people" as a socio-economic class
in opposition to the elite. In this view,"ordinary people" are positioned against a privileged,
often cosmopolitan elite who are seen as profiting unfairly at their expense (Kriesi, 2013).
This dynamic contrasts the interests of the common people with those of powerful and
educated elites (Canovan, 1999).
On the other hand, "the elite", which is another angle, represents the primary antagonist
to "the people" (Mudde, 2004). This dichotomy is central to populism and some scholars
identifying it as the defining feature of the ideology (Stanley, 2008). The relationship between
the elite and the people is often described in Manichean terms by framing them as forces of
light and darkness where the elite are portrayed as corrupt forces encroaching on the purity
of the people (Mudde, 2004). Within this framework, the elite are deemed "enemies of the
people" (Albertazzi and McDonnell, 2008).
The specific targets of this antagonism can vary on different factors. Depending on con-
text, the elite might encompass political figures, economic powerhouses, cultural influencers,
intellectuals, or members of the judiciary (Jagers and Walgrave, 2007). Despite the diver-
sity within these groups, populist rhetoric consistently portrays them with overwhelmingly
negative traits, labeling them as corrupt, exploitative, immoral, and even conspiring (Al-
bertazzi and McDonnell, 2008). They are accused of being selfish, arrogant, unaccountable,
and incompetent (Rooduijn, 2013). An identity that stands in direct contrast to the virtu-
ous and unified people is imposed on the elites (Rooduijn, 2013). This portrayal effectively
homogenizes the elite, much like "the people," as a cohesive group bound by undesirable
characteristics (Stanley, 2008).
The elite are often accused of betraying the public trust and hoarding power unjustly as
well as they are seen as responsible for the difficulties faced by the people(Mény and Surel,
2002). Furthermore, the elite are accused of having lost touch with the public’s interests,
understanding only their own perspectives and priorities (Stanley, 2008). Through corrup-
7
tion, these elites are perceived as having distorted democracy to serve their own interests,
weakening the people’s sovereignty in the process (Albertazzi and McDonnell, 2008). In
populist narratives, the people have effectively lost control over democratic processes and
left in a dependent as well as powerless position against an elites who prioritize their own
gains over the well-being of the broader public (Rooduijn, 2013).
In the populist framework, the third angle the populist actors serves as a critical inter-
mediary who challenges the elite for obstructing the will and prominence of "the people"
(Rooduijn, 2013). This figure claims to restore the voice and power of the people, aligning
closely with their interests and aspirations (Albertazzi and McDonnell, 2008). Populists may
emerge as movements, political parties, or individual leaders, the latter often characterized
by their charismatic appeal (Mudde, 2004). Many scholars view the charismatic leader as
a central figure in populism (e.g., Canovan, 1999; Albertazzi and McDonnell, 2008), while
others consider charisma an optional but frequent element of populism (Mudde, 2004 ).
Charismatic leaders tend to be political outsiders who articulate the will of the people and
act as their advocates and they often instill a unique "populist mood" or emotional fervor in
their supporters, transforming politics from ordinary administration into an urgent mission
to "save" the nation (Canovan, 1999).
One hallmark of populist ideology is a rejection of lack of transparency and procedu-
ral complexity within government. Populists criticize backroom negotiations, bureaucratic
intricacies, and policy details that require specialized knowledge, advocating instead for
straightforward and transparent governance (Canovan, 1999). Additionally, populists often
favor direct democracy, with policies that minimize intermediary bodies such as parliaments
and emphasize directly elected leaders (Albertazzi and McDonnell, 2008). This direct con-
nection between the people and leadership is seen as essential to honoring popular sovereignty
(Mény and Surel, 2002).
Populists also oppose established political parties criticizing these organizations dilute
the unity of the people and prioritize self-interest over public good (Mudde, 2004). They
assert that they alone are truly aligned with the people’s desires, intuitively understanding
and embodying their needs and demands (Mudde, 2004). They promote themselves as the
authentic voice of the people as well as standing as guardians of popular sovereignty and
emphasizing an unmediated relationship between the people and their leaders (Rooduijn,
2013).
2.1.2 Populist Communication Strategy
In this triangle of populism, populist leaders are using some communication strategies by
which they attract the people. To develop effective populist rhetoric and political commu-
nication strategies, it is essential to identify the core dimensions of populism. Previously,
researchers identified three core dimension of populist communication which are basically
the communication strategy.
The dimension of anti-elitism is related to the word "elite" which is previously described.
Anti-elitism is characterized by negative references to "the elite” (Fernández-García and Lu-
engo, 2020). This anti-elitism strategy can be divided into three different categories. Firstly,
discrediting the elite which aims to undermine the legitimacy of established institutions and
leader. It delves with portraying the elite as corrupt and out of touch with common people
8
(Cassell, 2021).
Secondly, most common segment of anti elitism strategy is blaming the elitee. It is
basically happened when populist frequently attribute societal problems to the failure and
actions of the elite. This blame game helps in rallying the masses against the masses against
a common adversary (De Bruycker and Rooduijn, 2021). Hameleers et al. (2023) have de-
scribed it through an example of a message that portrays that the government is responsible
for not clearly informing the public on Covid-19 health regulations assigns blame to polit-
ical elites but does not refer to a moral antagonism between “pure” or “ordinary” citizens
and “corrupt” elite actors. In 2020, opposition leader Joe Biden blamed President Donald
Trump accusing for the same reason (The Guardian, 2020). And the last segment of the anti
elitism dimension is to emphasize “us versus them” for establishing the disconnection or a
detachment between the elite and the common people by the populist actor (Wirth et al.,
2016). For example, a tweet, "The Government no longer knows, where it discriminates and
even if, not why" by Volker Beck, German politician indicate the detachment (Ernst et al.,
2017).
People centrism can be argued to be the most important dimension of the populist
rhetoric. Populist actors want to first and foremost refer to the people as well as identify
with them. However, the people are seen as a monolithic group that does not have any inter-
nal differences (Jagers and Walgrave, 2007). According to different scholars, this dimension
can be segmented into four categories. Sometimes the populist leader highlight the virtues of
the common people and portraying them inherently good and morally superior to elite (Block
and Negrine, 2017). It can be defined as stressing the people’s virtues. Secondly, praising the
peoples achievements is another example of people-centrism. This involves celebrating the
achievements and potential of the people and reinforcing their collective identity and pride
(Sintes-Olivella et al., 2020). An example of celebrating the achievements and potential of
"the people" while reinforcing collective identity and pride can be found in Hugo Chávez’s
speeches and rhetoric during his presidency in Venezuela (Hawkins, 2009). Chávez often
framed his government as a movement for the Venezuelan people and promoting a vision of
pride and empowerment rooted in national identity. Thirdly, it is already mentioned before
that is populist often describe people as people as homogeneous group (Hameleers et al.,
2017). That means people has common interests and values. By stating a monolithic group
people populist often exclude those who do not fit in this category. For example, Recep
Tayyip Erdoğan’s rhetoric often frames his supporters as the "authentic" people of Turkey,
marginalizing secularists and liberals as outsiders and fostering a polarized national iden-
tity (Yabanci, 2018). Furthermore, populist leaders seek to show their connection to the
ordinary people through rhetoric and actions that suggest they understand the people’s con-
cern (Cassell, 2021). For instance, Mexican populist leader Andrés Manuel López Obrador
demonstrates his connection to ordinary Mexicans by emphasizing his humble roots and
rejecting presidential privileges, positioning himself as a true “servant of the people.”
Finally, another prominent populist communication strategy is Restoring Sovereignty.
A key strategy is to call for the return of power to the people and advocating for direct
democracy as well as citizen participation in governance. It is basically demanding pop-
ular sovereignty by populist leader (Canovan, 1999). It is basically demanding popular
sovereignty by populist leader and denying elite sovereignty to curtail power of the elites. An
example of a populist leader demanding popular sovereignty and rejecting elite sovereignty
9
is Viktor Orbán in Hungary. He frequently criticizes the "liberal elite" and frames his lead-
ership as a restoration of power to the Hungarian people (Bugaric, 2019). He advocates for
"illiberal democracy," positioning himself against what he describes as foreign influence and
elite-dominated institutions (Bugaric, 2019).
These are the communication strategies which populist leaders often use in both tradi-
tional and social media. In this thesis context, this strategies are used in social media posts
by the political leader. Later it will be discussed in the method section how can we measure
those communication strategies in social media.
2.2 Social Identity Theory: Why are "People" attracted?
Now the question is why the people are attracted to these communication strategies. It can
be described by Social Identity Theory which was developed by Henri Tajfel and John Turner
(Harwood, 2020). The theory explores how individuals self concepts are derived from their
membership in social groups and it is also instrumental in understanding group dynamics,
inter group relations as well as the effects of social identification on behavior (Harwood, 2020).
So it can be easier to understand individual’s group behavior in social media through this
theory. Scholars segmented four core concepts of SIT. Firstly, the social categorization, which
involves classifying people, simplifies the social world helps individuals define their place
within it (Tajfel and Turner, 1979). For instance, people may categorize themselves based
on nationality, religion, or political affiliation (Tajfel and Turner, 1979). So, it is obvious if
a political leader’s ideology or any of the mentioned facts match with an individual, they
can be attracted by these populist communication. Secondly, once categorized, individuals
adopt the identity of the group. According to Tajfel and Turner (1986), this identification
provides a sense of belonging. For example, being identified as a member of a political party
can influence their political views and participation. Thirdly, Turner (1982) mentioned the
individual compare their in group to out group which serves to maintain self esteem. He
added this comparison often leads to in group favoritism and out group discrimination.
Lastly, Tajfel (1981) described that individuals have a tendency to favor one’s own group to
others. He added also this favoritism can lead to prejudice and conflict.
These four criteria of social identity theory fit with a citizen, who represents the people in
the context of populism, to be attracted to populist messages in social media because in the
realm of social media people can classify themselves into a particular political ideology. As
political actor has a huge fan base in social media (G. Enli and Moe, 2013) and the supporter
have their own social media groups or channels as well as they can share their thoughts on
those places (Katz et al., 2013). Eventually, people can adopt the identity from that group
of people. Furthermore, it also create in group favoritism and out group discrimination
(Reicher et al., 1995). Citizen can easily compare other people or group of people in social
media (Vogel et al., 2014). That also matches one of the criteria of social identity theory.
2.3 Civic Voluntarism Model: Why do "People" participate?
Social Identity theory explain the reason behind citizen’s attraction to populist communica-
tion. But still it is needed to clarify that why people do participate in populist messages as
10
participation is something about taking action (Verba et al., 1995). The Civic Voluntarism
Model (CVM), introduced by Verba, Schlozman, and Brady in Voice and Equality (1995),
provides a lens through which we can understand why some people participate actively in
civic and political life while others do not (Verba et al., 1995). Scholars have used this model
to explore disparities in participation across society and most importantly they can describe
not only who participate but also why some individuals are excluded from civic life (Brady
et al., 1999;Putnam, 2000 ;Rosenstone and Hansen, 1993) .
One of the main ideas of the CVM is that participation depends on whether people have
the necessary resources. Verba et al. (1995) mentioned that these resources include time,
money, and the skills and without enough time, individuals may find it difficult to attend
meetings or volunteer. They added, financial resources are often needed to make donations
or participate in events. The model thus emphasizes that the lack of these resources can be
a key barrier to civic involvement (Verba et al., 1995). On social media, resources like time,
digital skills, and sometimes financial means are crucial for active participation. It is im-
possible to access social media without device which required financial ability. Furthermore,
individuals with limited time may struggle to engage with social media platforms such as on-
line discussions, creating content, or following issues consistently. This is comparable to the
offline challenge of attending meetings or volunteering. Digital skills also become essential in
the social media context. The ability to navigate platforms and understand content as well
as create posts influences how effectively someone can participate in civic discourse online.
So, the CVM suggests that lacking these online resources (time, digital skills, and sometimes
financial means) may prevent people from actively participating in civic discussions.
Another factor the CVM examines is the individual’s psychological engagement or the
level of interest. This includes political efficacy which is the belief that one’s voice can make
a difference in the political arena (Verba et al., 1995) and without this sense of efficacy
people may feel that their efforts are pointless (Verba et al., 1995). That means , they
may lack of motivation to participate. Applying the Civic Voluntarism Model’s (CVM)
concept of psychological engagement to social media it can be reasonable too. On social
media, psychological engagement comparable to users’ willingness to actively contribute to
discussions, share political content, and advocate for causes they believe in. And previously
it is mentioned that people feel attracted to those content because of social identification.
The people who have high political efficacy are more likely to participate actively by posting,
commenting, or sharing content because they believe these actions can make an impact on
their audience. Research has shown that political efficacy is linked to higher levels of civic
engagement (Bandura, 1997; Delli Carpini, 2004). On social media, a sense of influence
and validate individual voices can encourage users to feel more politically efficacious and
make them more likely to participate actively (Velasquez and LaRose, 2015). Features like
likes, shares, and comments can reinforce a user’s sense of impact by providing immediate
feedback and recognition. Consequently, it can enhance the feelings of political efficacy and
motivating further engagement.
The third component of the CVM involves recruitment networks. It is basically social
channels that invite or encourage people to participate. According to Verba et al. (1995)
these network can be families, friends, colleagues, and community. As per studies, people are
more interested to engage civically when they are invited by known and trustworthy person
(Brady et al., 1999). In the context of social media, the Civic Voluntarism Model’s (CVM)
11
concept of networks can be observed through online communities, friends, followers, and
influential figures. These people encourage participation and foster engagement. On social
media, networks operate through like tags, shares, and mentions. The users can directly
involve others in a conversation or campaign through those network.
These are the components explain why people are willing to participate in populist ac-
tivities. In later section, it will discover platform specific observation, the difference between
government and opposition leader’s messages and most importantly the engagement with
populist message in social media.
2.4 Platform Differences in Populist Messaging: Facebook vs. Twitter
Social media has fundamentally changed how political leaders interact with citizens and
it is also offering them direct access to public attention without the need for traditional
media (Chadwick, 2013). Facebook and Twitter are widely used by political actors to reach
their audiences among many platforms available (Bossetta, 2018). However, these platforms
have unique characteristics that influence how well they support different types of political
messages as well as especially populist content aimed at mobilizing “the people” against
perceived elites (Engesser et al., 2017).
Facebook does basically emphasize visuals, community-building, and long-form story-
telling which all play into the strengths of populist messaging(Gandini et al., 2022). Pop-
ulism, as Mudde (2004) describes, relies on appeals to “the people” while often positioning an
“elite” as an opponent. Algorithms of Facebook are designed to prioritize posts that encour-
age engagement, like shares, comments, and reaction as well as it aligns well with this style
of rhetoric which often provokes emotional responses and lively discussion (Cinelli et al.,
2021). Furthermore, large and diverse user base of Facebook also provides a fertile ground
for populist messages to reach and resonate with a wide audience (Bossetta, 2018). The
features like comment thread and reactions of this platform encourages users to participate
in conversations which helps spread populist narratives and increase their visibility (Kreis,
2017).
On the other hand, the structure of Twitter depends heavily on short-form content and
real time interaction (Bruns, 2020). These features may limit its effectiveness for the de-
tailed storytelling that populist messages sometimes require. The success rate of Twitter
is really high in news and brief updates but its character limits it challenging to build the
nuanced narratives that populist messages often involve (Ott, 2017). Moreover, Twitter’s
user base is more concentrated among journalists, politicians, and other professionals (Woj-
cik and Hughes, 2019). So, these factors are making Twitter less diverse than Facebook and
potentially less receptive to populist messaging.
Studies suggest that political content with populist themes often achieves higher engage-
ment on Facebook than on Twitter (Engesser et al., 2017). Cinelli et al. (2021) investigated
that Facebook’s algorithms help contents that generate significant interaction and creating a
feedback loop. In this mechanism, populist posts with high engagement become more visible
and gain even more engagement (Cinelli et al., 2021). Facebook also allows for longer cap-
tions, multimedia posts, and detailed commenting. These features provide a richer context
for populist messages and can lead to prolonged user interaction (Bobba, 2019).
12
In contrast, Kalsnes (2019) found that algorithm of Twitter prioritizes recency. So tweets
have a shorter lifespan and are more likely to reach audiences only in the immediate hours
after posting (Kalsnes, 2019). While Twitter fosters real-time conversation but its platform
design does not encourage extended engagement with individual posts in the same way
Facebook does. Furthermore, political actors might find it more challenging to sustain high
engagement on Twitter for populist content compared to Facebook where both the content
and the platform’s features align well with populist engagement (Stier et al., 2018)
2.5 Populist Rhetoric in Opposition and Government
Populist messaging often serves as a powerful tool for political actors because they position
themselves as the voice of "the people" against an entrenched "elite" (Mudde, 2004). This
rhetoric can be especially opportunities for opposition political leaders who seek to challenge
those in government by portraying themselves as advocates for the common citizen (Hawkins,
2009). Research consistently shows that opposition leaders tend to adopt populist messages
more frequently than their counterparts in government (Pauwels, 2011) . They use this
approach to mobilize support by tapping into public frustrations and presenting themselves
as a corrective force to the status quo (Mudde, 2004).
Studies suggest that opposition figures leverage populist rhetoric to frame the government
as disconnected from the needs and values of the people and they create a clear “us versus
them” narrative that resonates with voters who feel marginalized or undeserved (Bos and
Brants, 2014) . This strategic use of populist messages allows opposition leaders to establish
a sense of solidarity with their audience (Hameleers et al., 2017). Engesser et al. (2017) found
that opposition politicians frequently employ simple and emotive language to heighten this
contrast. They also added that, making their messages accessible and emotionally compelling
to a broad base of potential supporters.
In contrast, research indicates that government actors often adopt a more moderate tone
when it comes to populist messaging. As representatives of the establishment they once
criticized, government leaders typically face constraints that limit their ability to position
themselves as outsiders (Moffitt, 2016). This shift can be explained by a need to maintain a
more diplomatic and balanced image and often due to their responsibility to represent not
only their supporters but the entire nation or constituency (de Vreese et al., 2018). In their
study on European populist leaders, Waisbord and Amado (2017) highlight how government
officials tend to moderate their language once in power and move away from the strong
populist tones they may have used in opposition.
Power dynamics play a crucial role in determining how and when populist rhetoric is used
(Bonikowski and Gidron, 2016). The opposition leaders benefit from positioning themselves
as challengers to an established system but those within the government are often pressured
to adopt a more measured approach to retain public confidence and legitimacy (Bonikowski
and Gidron, 2016), Moffitt and Tormey, 2014). This is especially true in democratic settings
where leaders are accountable to diverse groups of citizens and overly divisive messaging could
threaten their position (Krämer, 2014) . Thus, while populism can be a potent strategy for
gaining support, its use tends to diminish among political actors once they assume power
(Hawkins, 2009).
13
In summary, the literature suggests a clear divide in the use of populist messaging between
opposition and government political actors. Opposition leaders are more likely to employ
populist rhetoric to frame themselves as champions of the people against an unresponsive
elite. However, once in government, political actors often moderate their populist tone to
align with the broader demands of governance. This shift highlights how the strategic use of
populism is closely linked to a political actor’s position and the power dynamics that come
with it.
2.6 Engagement with Populist vs. Non-Populist Messages
Researchers discovered that divisive content tends to receive higher levels of engagement
than neutral or policy-focused messages (Bobba, 2019;Engesser et al., 2017). Engesser et al.
(2017) mentioned that populist messages are often characterized by conflict-driven narratives
that create “the people” against “the elite,” tend to attract more likes, shares, and comments.
This engagement not only extends the reach of populist content but also reinforces the power
of simplified, emotional appeals within digital platforms (Bos and Brants, 2014; Blassnig et
al., 2019).
One significant factor behind the high engagement of populist messages is their ability
to evoke strong emotions (Hameleers et al., 2017). Stieglitz and Dang-Xuan (2013) posits
that populist rhetoric commonly employs language that instigate anger, fear, or a sense
of injustice. In response, audiences eager to react, often through sharing or commenting
(Stieglitz and Dang-Xuan, 2013). Studies show that social media algorithms prioritize con-
tent that generates engagement and creating a feedback loop where high-response posts are
often populist (Cinelli et al., 2021). This dynamic has been shown in research by de Vreese
et al. (2018). He argued that platforms like Facebook encourage emotionally charged posts
and enhancing the visibility of populist messages.
Another reason populist messages resonate is their narrative simplicity (Jagers and Wal-
grave, 2007). By presenting issues in clear “us versus them” terms, populist rhetoric mini-
mizes complexity as well as making it highly shareable and accessible to a broad audience
(Krämer, 2017). Simplified messaging allows users to quickly grasp the content’s core mes-
sage and encourage immediate responses and sharing behaviors (Stieglitz and Dang-Xuan,
2013). Bobba and McDonnell (2016) demonstrated that such messaging is particularly ef-
fective in election context. They added that, simplified narratives are contrast with the
policy-heavy posts by mainstream politicians so that they struggle fighting their opponent.
Researcher found that populist messaging are outperforming non-populist content in en-
gagement. For instance, during the 2016 U.S. presidential election Blassnig et al. (2019)
examined social media activity. They found that posts populist themes consistently higher
than others in terms of user interaction metrics like likes and shares. Similarly, Engesser
et al. (2017) analyzed social media posts from European politicians. They noted that posts
containing populist elements consistently achieved higher engagement than non-populist mes-
sages. This finding aligns with findings by Bos and Brants (2014). They found that Dutch
political messages with populist themes were shared more widely when they targeted per-
ceived “elites”.
Furthermore, scholars highlight the role of algorithms in producing the popularity of
populist messages. Facebook prioritize content that ignite engagement, often placing posts
14
with strong emotional appeal at the top of users’ feeds (Cinelli et al., 2021). Underscore
this point, Stieglitz and Dang-Xuan (2013) showed that posts that evoke anger or outrage
are more likely to be shared. They added, this creates a self-reinforcing system where
populist content not only reaches more people but also gains credibility through widespread
interaction.
In summary, these litterateurs suggest that populist messages typically get higher en-
gagement than non-populist content on social media. This phenomenon is driven by several
factors: the emotionally evocative nature of populist rhetoric (Engesser et al., 2017; Stieglitz
and Dang-Xuan, 2013), the simplified narratives that make such messages easy to share
(Krämer, 2017; Bos and Brants, 2014), and the engagement-focused algorithms that pri-
oritize high-response content (Cinelli et al., 2021). These elements combine to create an
environment where populist messaging not only thrives but often outperforms more conven-
tional political communication.
2.7 Engagement Dynamics of Government vs. Opposition Leaders
Studies indicate that government leaders hold symbolic power that can attract substantial
attention and interaction on social media platforms (McGregor et al., 2017). This engage-
ment reflects the trust and interest to those in power and it facilitate government leaders
(Lalancette and Raynauld, 2018).
Research suggests that social media posts from government officials benefit. Government
leaders often represent stability and influence (Larsson and Moe, 2014). It resonates with
followers and drive engagement (Lalancette and Raynauld, 2018). In an analysis of social
media interactions with heads of state, Larsson and Ihlen (2015) found that government
political actors frequently received more likes, shares, and comments than opposition political
actor. This finding aligns with studies on the perceived authority of government political
actors , which highlights how their official capacity can amplify the public’s interest in their
communications (Kalsnes, 2016).
Further reinforcing this trend is the unique access that government leaders have to pol-
icy announcements, national events, and insider perspectives (D’heer and Verdegem, 2014;
Parmelee and Bichard, 2011). By doing this they can draw public interest and engagement
(Parmelee and Bichard, 2011). Social media users often turn to government profiles for up-
dates on policies that affect their lives directly (Bertot et al., 2012). This is leading to a
natural increase in engagement levels for government accounts. By comparison, opposition
leaders, although often active and vocal on social media but lack the same level of insider
authority and may not attract as much attention (G. S. Enli and Skogerbø, 2013).
Contrasting the engagement dynamics between government and opposition leaders reveals
how power dynamics influence social media engagement. Lalancette and Raynauld (2018)
noted that opposition leaders use populist tactics or critique government actions to mobilize
supporters but these messages do not always carry the same weight as official statements from
sitting leaders. This discrepancy in engagement may also create perception that government
leaders have direct influence over policy and decision-making (Lee and VanDyke, 2015),
whereas opposition leaders’ influence is often seen as more limited (Kriesi, 2014).
In summary, the literature suggests that government leaders generally experience higher
engagement on social media than opposition leaders. This engagement pattern highlights
15
how the power and position of political actors shape their social media interactions.
2.8 Research Gap
While there is a good amount of research on how populist communication shapes public
opinion and democratic engagement on social media ( Bos and Brants, 2014; Engesser et
al., 2017), we still know surprisingly little about how different platforms like Facebook and
Twitter affect the reach and impact of these messages. Most studies look at general patterns,
but they do not explore deep into how specific types of populist rhetoric encourage citizens
to engage, especially depending on whether the message comes from government leaders or
opposition figures (Cinelli et al., 2021; Stieglitz and Dang-Xuan, 2013). This study seeks to
address that gap by focusing on that populist messaging plays out across different platforms
and political roles. In doing so, it is assuming to provide a clearer picture of how social
media might be influencing civic participation in today’s digital democracy.
2.9 Hypothesis
Based on the integration of theoretical frameworks Populist communication, Social Identity
Theory (SIT), and the Civic Voluntarism Model (CVM) and informed by findings from
previous studies, the following hypotheses are proposed:
Drawing on Social Identity Theory (Tajfel and Turner, 1979) which explains how indi-
viduals identify with groups and attracted to populist messaging that resonates with “the
people” against “the elite” and on research by Engesser et al. (2017) and Cinelli et al. (2021)
regarding platform dynamics, the first hypothesis states:
H1: Facebook will be the most commonly used platform for populist messages by political
actors.
The features of Facebook align with SIT by nurturing strong group identities and commu-
nity building. Then it becomes a favored platform for engaging users with populist messages.
On the basis of Civic Voluntarism Model (Verba et al., 1995) which highlights the role of
engagement resources such as time and digital skills, and constructing on studies by Bossetta
(2018) and Kalsnes (2019), the second hypothesis suggests:
H2: User engagement for populist messages is higher on Facebook than on Twitter.
Facebook’s emphasis on visuals and longer post formats aligns with CVM. It permits
users with different levels of digital skills and resources to engage more deeply with populist
content than they would on Twitter.
Based on positions opposition political actors as more likely to employ populist rhetoric to
challenge the government which is supported by findings from Pauwels (2011) and Bonikowski
and Gidron (2016), the third hypothesis proposes:
H3: Political actors in opposition will use populist messages more frequently than those
in government.
Opposition leaders use populist messages to mobilize people. Supported by Social Identity
Theory and findings from Engesser et al. (2017) and Blassnig et al. (2019), which suggest
that divisive, emotionally charged messages enhance group solidarity and interaction, the
fourth hypothesis is:
16
H4: Populist messages will receive higher engagement (likes, shares, comments) than
non-populist messages on social media.
Populist messages tap into the in-group vs. out-group dynamics of SIT, encouraging
individuals to engage more with content that resonates with their identity. Finally, the Civic
Voluntarism Model provide insights into how government political actors may experience
higher engagement. Studies by Lalancette and Raynauld (2018) and Larsson and Moe (2014)
indicate higher engagement for government leaders due to their influence, lead to the fifth
hypothesis:
H5: The social media activity of government leaders will result in higher levels of user
engagement compared to the activity of opposition leaders.
The CVM suggests that government leaders’ resources (e.g., access to policy announce-
ments) and influence contribute to greater engagement, even if they use less populist language
than opposition leaders.
In the next chapter research methodology will be discussed to investigate this hypothesis.
After getting the results from applying methods it can be possible to answer those research
questions previously mentioned.
17
Chapter 3
Research Method and Materials
3.1 Research Approach
In this study, a quantitative research approach will be used to explore how populist campaign
messages influence citizen engagement on social media. This approach allows us to gather
statistical data on how often political actors use populist messages in their social media cam-
paigns and to assess how people interact with these posts. By focusing on statistical analysis,
we aim to provide clear evidence of the relationship between populist communication and
user engagement on social media platforms. (Creswell, 2014).
Bryman (2016) mentioned that quantitative research is rooted in a positivist view of
knowledge. That means the objective truths can be discovered through systematic mea-
surement and analysis. Positivism is underpinned by an objectivist ontological perspective,
positing that empirical material is tangible and factual. This study will follow a deductive
approach, starting with hypotheses derived from existing theories and then testing these hy-
potheses through empirical data (Bryman, 2016). The hypotheses are informed by Populism,
Social Identity Theory, and the Civil Voluntarism Model.
The study will conduct a quantitative content analysis of social media posts from political
actors in the United States and the United Kingdom. Specifically, the analysis will focus
on the Facebook and Twitter posts of Donald Trump and Joe Biden from the US (2020
United States presidential election), and Nigel Farage and Boris Johnson from the UK (2019
United Kingdom general election), during the period leading up to the elections. This
approach allows for the systematic coding and quantification of textual data, enabling the
identification of patterns and trends in the use of populist messages.
By employing this research approach, the study aims to provide empirical evidence on the
prevalence and impact of populist messages in political campaigns on social media, thereby
contributing to our understanding of digital political communication.
3.2 Choice of Method
Content analysis involves systematically coding and quantifying textual data to identify pat-
terns and trends. This method is particularly suitable for analyzing social media content due
to its ability to handle large volumes of data and provide quantifiable results (Krippendorff,
2018). Additionally, content analysis allows researchers to objectively categorize and inter-
18
pret communication content, making it an effective tool for examining political discourse
(Neuendorf, 2017).
The coding scheme will categorize social media posts as either populist or non-populist
based on predefined criteria established in the literature. Engagement metrics such as likes,
shares, and comments will be recorded to measure user engagement with these posts. Quanti-
tative content analysis is well-suited for this type of research as it facilitates the identification
of patterns and trends across different social media platforms and political actors (Riffe et
al., 2019).
Using quantitative content analysis in this study has several advantages. First, it provides
a systematic and replicable means of analyzing communication content, which enhances the
reliability and validity of the findings (Weber, 1990). Second, it allows for the examination of
large datasets, which is essential for understanding the dynamics of social media interactions
in political campaigns (Bryman, 2016). Third, quantitative content analysis is flexible and
can be adapted to various research contexts, making it a versatile method for studying digital
communication (Hsieh and Shannon, 2005).
The quantitative approach enables the statistical testing of the hypotheses which are
formed in the previous chapter.
3.3 Context: Two Nations Divided by a Common Language
The famous playwright George Bernard Shaw humorously said that, England and America
are two countries separated by a common language (Milroy, 2000) to describe the similari-
ties and differences between United Kingdom and United States. These two countries have
a long history of fight and cooperation and most importantly, they are key player of in-
ternational politics. Recently, as previously mentioned, populism has an influence in their
national politics and policy making. Brexit in United Kingdom and Capitol Hill Attack in
United States are the significant event that can be mentioned as an example of populism.
According to Piazza (2024), populism is responsible for political violence happened in United
States recent days. Similar results found by another researcher for United Kingdom (Pater-
son et al., 2023). Though these two countries are from different continent and their election
system are different, for both cases elections signify democracy but it is declining (Behr,
2024). Analyzing the the populist communication strategies of U.S. and U.K. political lead-
ers can shed light on how cultural and systemic factors shape the use and adaptation of
these techniques. The U.S. political system, with its strong focus on presidential power and
a two-party structure, may foster different rhetorical approaches compared to the U.K.’s
parliamentary system, where party leadership and coalition-building are essential. Under-
standing these nuances helps scholars, policymakers, and the public better comprehend how
populist communication functions in different democratic contexts and the potential risks it
poses to the resilience of democratic systems. Furthermore, exploring the populist commu-
nication strategies of political leaders from the U.S. and U.K. on social media is essential for
understanding the modern political landscape. It offers insight into the relationship between
technology, rhetoric, and democracy, helping us stay alert to both the opportunities and
challenges that come with this new era of political communication.
19
3.4 Materials
3.4.1 Sampling
The social media accounts of the politicians are chosen according to purposive sampling in
order to get two politicians from both United States and United Kingdom. A purposive
sampling is basically a non-probability sampling method where participants are selected
based on specific characteristics or qualities that align with the research objectives (Etikan
et al., 2016). The two politicians taken from the US are Donald Trump and Joe Biden.
They contested in the 2020 US Presidential Election. Trump was in power, and Joe Biden
was in the opposition. Similarly, in the 2019 General Election, Boris Johnson was in power,
and Nigel Farage was in the opposition. In line with the research objective, Facebook posts
and tweets have been collected from their Facebook and Twitter official accounts. Twitter
rebranded as X (Collins, 2023) but in this research "Twitter" is used. As Donald Trump’s
Twitter account has been suspended from Twitter, so that those data are collected from The
American Presidency Project by University of California, Santa Barbara (Project, 2024)
Table 3.1: Official Facebook and Twitter Accounts of Political Actors
Name Facebook Twitter
Nigel Farage Nigel Farage @Nigel_Farage
Boris Johnson Boris Johnson @BorisJohnson
Joe Biden Joe Biden @JoeBiden
Donald Trump Donald Trump @realDonaldTrump
The hottest period of pre-election activity on social media generally occurs during the
final weeks leading up to Election Day. Specifically, politicians are most active from weeks
6 weeks before an election. During this period, known as the "home stretch," candidates
ramp up their efforts to engage with voters, share their platforms, respond to opponents,
and mobilize supporters. This intense activity includes heightened advertising, frequent
posts, and live interactions across various social media platforms. Research indicates that
the volume and intensity of social media activity by politicians significantly increase during
these critical weeks as they seek to maximize voter turnout and solidify support (J. Smith,
2020). So, in this study 6 week pre election day has been counted. Additionally, it is decided
to make artificial weeks for the compile the data. For instance, if the study spans the total
period can be divided into several artificial weeks. This method allows for consistent and
equal sampling periods, ensuring that content is analyzed uniformly over the study duration
(Hester and Dougall, 2007). By using artificial weeks, researchers can capture temporal
patterns and trends in social media activity or media coverage more effectively (Hester and
Dougall, 2007). This approach also facilitates comparisons across different periods within the
research timeline, enhancing the robustness and reliability of the findings (Trochim, 2020).
The days are selected from each alternative weeks.
The data collection period spans six weeks leading up to both the 2020 United States
Presidential Election and the 2019 United Kingdom General Election. For the UK General
Election, data collection dates are set as follows: 1st November (Friday), 9th November
(Saturday), 17th November (Sunday), 25th November (Monday), 3rd December (Tuesday),
20
and 11th December (Wednesday), with the election taking place on 12th December. Sim-
ilarly, for the US Presidential Election, the dates selected are 22nd September (Tuesday),
30th September (Wednesday), 8th October (Thursday), 16th October (Friday), 24th Oc-
tober (Saturday), and 2nd November (Sunday), preceding the election on 5th November.
Throughout this period, a total of 918 tweets and Facebook posts from four politicians were
collected for analysis.
3.4.2 Variables
There are 15 variables in total. The first three items concerned formalities such as the
Political leader’s name, Identification number, date of publishing and which social media’s
post it is (either Facebook or Twitter.
The second part of the coding scheme is about the engagement variable. Here Share,
Comment and Like are listed as variable. The share and comment function on Twitter is
called a retweet and reply respectively. Users can like a tweet by clicking the heart icon. But
these terms are referred as Share Comment and Like in this thesis.
The third part of the coding scheme considered what type of populist message e.g. anti-
elitism, people-centrism and restoring sovereignty. These variables are also segmented. For,
anti-elitism, discrediting the elite, blaming the elite and detaching the elite from the people
are the variables. Similarly, people centrism is also segmented into four variables which are
stressing the people’s virtues, praising the people’s achievements and stating a monolithic
people and demonstrating closeness to the people. Furthermore, restoring sovereignty is
also divided into two variable, demanding popular sovereignty and denying elite sovereignty.
These are the key variable which are defined from operationalization of populist communi-
cation strategies by Ernst et al. (2017).
Table 3.2: Conceptualization and Operationalization of Populist Communication Strategies
Key Variable Underlying Ideology Categories
Discrediting the elite Elites are corrupt. Elites are accused of being
malevolent, criminal, lazy,
stupid, extremist, racist, un-
democratic, etc. They are called
names and denied morality,
charisma, credibility, intelli-
gence, competence, consistency,
etc.
Continued on next page
21
Key Variable Underlying Ideology Categories
Blaming the elite Elites are harmful. Elites are described as a threat
or burden, responsible for nega-
tive developments or situations,
or as having committed a mis-
take or crime. They are not con-
sidered a source of enrichment or
responsible for positive develop-
ments or situations.
Detaching the elite Elites do not represent the Elites are described as not be-
from the people people. longing to, being close to, know-
ing, speaking for, or caring
about the people, nor perform-
ing everyday actions.
Stressing the people’s The people are virtuous. The people are endowed with
virtues morality, charisma, credibility,
intelligence, competence, consis-
tency, etc., and exempt from be-
ing malevolent, criminal, lazy,
stupid, extremist, racist, un-
democratic, etc.
Praising the people’s The people are beneficial. The people are described as an
achievements enrichment or responsible for
positive developments or situa-
tions, not as a threat, burden, or
responsible for negative develop-
ments, nor as having committed
mistakes or crimes.
Stating a monolithic The people are homoge- People are described as shar-
people neous. ing common feelings, desires, or
opinions.
Demonstrating close- The populist represents the The speaker describes them-
ness to the people people. selves as belonging to, being
close to, knowing, speaking for,
caring for, agreeing with, or per-
forming everyday actions for the
people, claiming to represent or
embody them.
Continued on next page
22
Key Variable Underlying Ideology Categories
Demanding popular The people are the ultimate The speaker argues for general
sovereignty sovereign. institutional reforms to grant
the people more power (e.g.,
through direct-democratic ele-
ments or increased political par-
ticipation) or for empowering
the people on specific issues
(e.g., elections, immigration, se-
curity).
Denying elite The elites deprive the peo- The speaker argues for reducing
sovereignty ple of their sovereignty. elite power in specific contexts
(e.g., elections, immigration, se-
curity).
In summary we have three different kind of variable. There are mainly nine key variables,
three engagement variable and three other variable. From the identification, it can be possible
to make new variable to separate opposition and government as well as the post is tweet
or not. From the nine key variable populist score will be formed for measurement. It will
described in later section.
Depending on the social media posts of each of the politicians, I have categorized the
nine variables as 1 or 0 which I have mentioned this in the codebook. 1 means here if the
posts suggest, for example, Praising the people’s achievements, then I have scored it 1 or
else 0. Then, I summed the score of each of the posts for each of the 9 variables and wrote
the populist score. The populist scores ranges from 0 to 9, 0 is base value and in my case
the highest populist score I got is 6.
I have categorized Facebook as 1 and Twitter as 0. For political affiliation, Opposition
is categorized as 1, and Government as 2. Here, Twitter is the reference category for social
media platform and Government is the reference category for political position.
Table 3.3: Descriptive Statistics
Variable Obs Mean Std. dev. Min Max
Share 918 9,281.394 19,763.9 1 318,027
Like 918 60,239.7 128,088.6 0 1,890,946
Comment 728 6,059.686 12,808.62 33 139,400
From the table 3.3 we can see the number of observations, their mean, standard deviation
(sd), and the maximum, minimum number for 12 variables.
Here, we can see that on average each content were shared around 9281 times with a
sd 19,763.9 which is quite spread out indicating that there might be some posts which was
shared less and some posts were shared more. For ’Like’ we can see that on average every
content got approximately 60,240 likes and also in this case we can see much variability for
23
Variable Obs %1
Discrediting the elite 918 20.74%
Blaming the elite 918 22.98%
Detaching the elite from the people 918 14.92%
Stressing the people’s virtues 918 8.50%
Praising the people’s achievements 918 13.51%
Stating a monolithic people 918 19.47%
Demonstrating closeness to the people 918 13.74%
Demanding popular sovereignty 918 10.13%
Denying elite sovereignty 918 5.45%
(1) Note: Shares of post containing populist dimension in percent
the data set. For the ’Comment’ we can see that there are some missing data. Here, we can
see that per content there was approximately 6,060 comments with much variability in the
data which can be seen from the standard deviation.
We have previously mentioned that the variables Discrediting the elite, Blaming the elite,
Detaching the elite from the people, Stressing the people’s virtues, Praising the people’s
achievements, Stating a monolithic people, Demonstrating closeness to the people, Demand-
ing popular sovereignty, Denying elite sovereignty are binary variables. For the variable
’Discrediting the elite’ we see that for 20.74% of the data contains this behavior and the
sd expresses that although most of the content do not contain this behavior but still there
are significant number of contents which shows this. We can also that around 23% of the
content has the behavior ’Blaming the elite’. The sd for this variable also shows that there
are significant number of contents which has this behavior although not maximum contents.
Around 15% of the contents shows it has the behavior ’Detaching the elite from the people’
and the sd shows that there are not many posts which contains this behavior. Around 8.5%
of contents show that it has the element ’Stressing the people’s virtues’ which means that
most content do not poses this behavior. Approximately 13.5% of the contents have ’Prais-
ing the people’s achievements’. The sd shows that it is not very common in the content.
Also, approximately 19.5% of the content has ’Stating a monolithic people’ behavior and the
sd shows that the contents contain a moderate level of this behavior. Around 14% of the
contents shows ’Demonstrating closeness to the people’ and sd shows there is a moderate
level of variability. ’Demanding popular sovereignty’ is present around 10% of the contents.
The sd shows that it is less common in the contents. Also ’Denying elite sovereignty’ is not
common in the contents and only 5.5% of the contents contains this behavior.
3.4.3 Coding Scheme and Procedure
Coding Scheme
The coding process is based on Neuendorf (2017) recommendations for code training and
codebook revision. The codebook provided detailed instructions for the coder and was revised
24
continuously throughout the code training period to ensure clarity for all coding situations.
Afterward, a pilot coding session was conducted where I coded the same units twice to verify
that the instructions were detailed enough to prevent discrepancies, ensuring the codebook’s
reliability and consistency. The final coding was performed using STATA. For the detailed
codebook, see Appendix B.
3.5 Reliability, Validity and Limitation
In quantitative content analyses conducted manually, reliability can be interpreted as inter-
coder reliability or intra-coder reliability. Intercoder reliability refers to the uniformity of
coding and is determined by evaluating the "level of agreement among two or more coders”
(Neuendorf, 2017). The researcher chose to code 92 post in the interval of 15 days, which
is approximately 10 % of the actual data. The simple agreement was then calculated as
percent agreement (or "crude agreem(ent") using the following formula: )
Number of Agreements
Percent of Agreement = × 100
Total Number of Coding Decisions
We have calculated the percentage of agreement for the variables Discrediting the elite,
Blaming the elite, Detaching the elite from the people, Stressing the people’s virtues, Prais-
ing the people’s achievements, Stating a monolithic people, Demonstrating closeness to the
people, Demanding popular sovereignty, Denying elite sovereignty according to the Holsti’s
method. The Percentage of Agreement shows high percent of agreement which indicates the
inter-rater reliability. To see the numbers please go to the Appendix C.
A second reliability test is done which is called the Cohen’s Kappa κ which accounts for
the agreement due to chance. The Cohen’s Kappa ranges from 0, which says agreement by
chance to 1, perfect agreement. If the value of the coefficient is less than 0 then it indicates
that the agreement is less than the chance. I have calculated this coefficient for each of our
nine variables and I have found that apart from the variable ’Demanding popular sovereignty’
and ’Denying elite sovereignty’ which has a coefficient around 0.55 the other variables shows
a significant agreement. The two variables shows some limitation and we will discuss this in
our analysis.
External validity relates to the extent to which the results can be applied to other set-
tings. This concept, also known as generalizability, can be evaluated by looking at how
representative the sample is (Neuendorf, 2017). I argue that the sample social media post
used in this study is representative of the entire population of social media posts by the
specific political leaders within the analyzed period due to the use of a non-probability sam-
pling method and the relatively large sample size. Therefore, the findings can be generalized
to all articles published in those social media during that time frame, but they cannot be
generalized beyond this context.
According to Neuendorf (2017), a sample size of 384 is needed for a 95 percent confidence
level with a sampling error of +/- 5 percent. Increasing the sample size to 1,087 reduces
the sampling error to +/- 3 percent. In this study, the sample size is 918, which places the
sampling error between +/- 3 to 5 percent. This means we can be 95 percent confident that
the results of this content analysis can be generalized to the entire population, with a margin
of error of 3 to 5 percent.
25
Twitter has banned the account of Donald Trump on January 2021 (CNN, 2021), so
the data collected from The American Presidency Project by University of California, Santa
Barbara (Project, 2024). But the limitation of that archive is they collected only share and
like but comment is missing. So, for hypothesis testing only like and share will be selected
as engagement variable.
3.6 Data Analysis
The data analysis was done using STATA version 18.0. This section will outline the statistical
techniques used for analyzing the descriptive data and testing the hypotheses.
3.7 Ethical Considerations
All data used in this study was sourced from publicly accessible information from accounts
on Twitter (X) and Facebook. The researcher took care to follow all applicable guidelines,
specifically the Developer Policy and Terms of Service, ensuring compliance as of the time
of this research. This approach meant that any content with privacy settings or restrictions
was automatically excluded from our dataset. By only using publicly available data, this
research made sure that no private or sensitive user information was included in our analysis.
26
Chapter 4
Results
In this chapter, we will be discussing the five hypothesis that I have proposed. I am doing
the regression analysis to understand the relation between different variables. As part of the
exploratory data analysis (EDA), I have included some visualization to understand patterns
among different variables.
4.1 Descriptive and Patterns
Figure 4.1: User engagement over populist score by social media platforms
In this figure, I have considered the mean user engagement over populist score for different
social media platforms and political position and the standard error. The black vertical lines
shows the standard error for different categories. We can see that in Twitter the average
user engagement is higher especially for the opposition actors with the increment of the
populist score. For Facebook, we see that government actors receive higher user engagement,
especially for the populist score 0 to 2. We can also see that the error margin is lower for
post in Facebook by government actors, which means that there are more consistency in the
27
engagement for different populist scores. For the opposition’s post on Facebook there is less
variability in the data set which we only see that for the populist score 2 and 3 .
For Twitter, we can see that on average the opposition receives higher engagement than
the government actors with the increment of the populist score. For the populist score from
2 to 6 we can see the higher engagement in Twitter for the posts of opposition actors. We
also notice that there is a much higher variability in the data set for the populist score 5. For
the government actors we do not notice much variability in the data set as the error margin
is lower. For the populist score 6, we need more data from which we can give a concrete
interpretation.
The analysis from this graph suggests that average user engagement in Twitter is higher
for the opposition actors with the increment of the populist score and in Facebook the
government actors receive higher user engagement on average. We will delve more into it
when doing the regression analysis.
Figure 4.2: User engagement by populist score in Facebook
The figures 4.2 and 4.3 shows the mean user engagement in different social media plat-
forms for the populist score. From the figure 4.2 we can see that on average the user
engagement decreases with the increment of populist score. For Twitter, we can see the user
engagement increases with the increment of populist score, though for the populist score 6,
this trend is not maintained in 4.3. Overall, we can see that user engagement is higher in
Twitter than in Facebook.
I have also tried to plot the distribution of populist score by the four politicians and the
social media platform (figure 4.4). The box plot for Boris Johnson has low range of populist
score with a lower median in Twitter. For Facebook, the range of populist score goes up
to 2 but with a low median. This tells us that Boris Johnson uses less populist messages
in both the platforms. Donald Trump has the wider range of populist score for Facebook
and the median is also high which shows that the Donald Trump has the tendency to use
more populist messages on Facebook compared to Twitter. Joe Biden has a low range of
populist score, with low median for the both the platforms which indicates that Joe Biden
has a tendency to use less populist messages on both the platforms. For Nigel Farage we
28
Figure 4.3: User engagement by populist score in Twitter
Figure 4.4: Distribution of populist score by politicians and social media platform
can see that he has a wide range of populist messages for both the platforms. The median
is higher for Twitter which indicates that Nigel Farage has a tendency to use more populist
messages on Twitter.
Furthermore, I have plotted the average user engagement for the posts of opposition and
government actors over populist score. The different colors represent the populist scores
from 0 to 6. The figure 4.5 shows that with the increment of populist score, government
actors receive higher user engagement on average than the opposition actors.
With these visualizations for the exploratory data analysis, now we will look into the
regression analysis for the proposed hypothesis.
29
Figure 4.5: Mean user engagement for the posts of government and opposition
4.2 Hypothesis 1
To test H1 which stated Facebook will be the most commonly used platform for populist
messages for political actors. I have chosen social media platform as the independent variable,
and as the dependent variable, I have chosen the populist score. The type of political position
meaning opposition or the government in power, is considered as control variable.
To analyze the hypothesis I have done simple regression analysis 1 and included a table with
the beta values for the independent and control variables.
From the table 4.1 we can see that the social media platform Facebook (vs. Twitter) has
a positive coefficient of 0.151 which says that Facebook has 0.151 times higher populist
message than Twitter, with a p-value of 0.070 which is not significant.
The coefficient for Opposition actors is 0.158 which shows that the opposition actors use on
average 0.158 unit more populist messages than the government actors. This has a p-value
of 0.058 which is not significant at the 5% level.
The intercept is positive with a p < 0.001 which suggests that even without accounting
for social media platform and political position, there will be always a level of populist
messaging.
I have also found that the F-statistics has the value 3.51 with a p-value of 0.03 which is
statistically significant at the 5% level. The adjusted R2 value is 0.0054 which is very low.
It says that though choice of platform and political position has an effect on the populist
score but overall the model does not capture much variability.
1The regression model can be written as: Populist Score = β0 + β1 ∗ Social media platform + β2 ∗
Political position+ ϵ
30
(Model H1)
Predictors β (SE)1
Intercept 0.786∗∗∗
(0.070)
Social media platform: Facebook 0.151
(0.083)
Political position: Opposition 0.158
(0.083)
Observations 918
Adjusted R2 0.0054
(1) Standard errors (SE) in parentheses
∗ p < 0.05, ∗∗ p < 0.01, ∗∗∗ p < 0.001
Table 4.1: Regression of Populist score for social media choice
Overall, we can see that Facebook might not be the most used platform for populist messages.
With this analysis, we have to reject the hypothesis.
4.3 Hypothesis 2
To test H2 which stated, user engagement for populist messages is higher on Facebook than
on Twitter, we have chosen user engagement as the dependent variable; here we have con-
sidered like and share for the user engagement variable, social media platform (Facebook vs.
Twitter) as the independent variable, and populist Score and political position (Opposition
vs Government) as the control variable.
Since the dependent variable is a count variable we are using Negative Binomial regression
(NBR). We have considered both the Poisson regression and Negative Binomial regression,
since our data shows overdispersion we have used NBR. We can also see from the likelihood
ratio test that negative binomial regression is suitable for this hypothesis.
The regression table shows that the Wald chi-square test is statistically significant.
From the table 4.2 the social media platform: Facebook has a negative coefficient with a
significant p-value (p < 0.001). This shows that on average Facebook has 0.629 unit less
user engagement than Twitter. The political position: Opposition has a negative coefficient
with a p-value of 0.001 which is significant. It shows that opposition actors on average have
0.460 unit less user engagement in their social media posts. The control variable populist
score has a positive coefficient with p = 0.022, it says that the more populist a message is,
it gets more user engagement.
31
(Model H2)
Predictors β (SE)1
Intercept 11.501∗∗∗
(0.102)
Social media platform: Facebook -0.629∗∗∗
(0.133)
Political position: Opposition -0.460∗∗∗
(0.134)
Populist Score 0.114∗∗
(0.050)
Dispersion coefficient 0.718∗∗∗
(0.038)
Observations 918
Pseudo R2 0.0027
(1) Standard errors (SE) in parentheses
∗ p < 0.05, ∗∗ p < 0.01, ∗∗∗ p < 0.001
Table 4.2: Regression of User Engagement on social media platform
32
The intercept is 11.501 with a p < 0.001 which is statistically significant.
Overall we see that Facebook gets lower user engagement for populist messages than twitter
and we have to reject the hypothesis.
4.4 Hypothesis 3
To test H3 which stated, political actors in opposition will use populist messages more fre-
quently than those in governance, we set populist Score is the dependent variable, political
position (Opposition vs. Government) is the independent variable, and social media plat-
form(Facebook vs. Twitter) is the control variable.
(Model H3)
Predictors β (SE)1
Intercept 0.787∗
(0.070)
Political position: Opposition 0.158
(0.083)
Social media platform: Facebook 0.151
(0.083)
Observations 918
Adjusted R2 0.0054
(1) Standard errors (SE) in parentheses
∗ p < 0.05, ∗∗ p < 0.01, ∗∗∗ p < 0.001
Table 4.3: Regression of Populist score for Political Position
We have used a simple regression analysis here. The table 4.3 shows that the coefficient for
opposition is positive and has a p-value of 0.058 which is not statistically significant. It means
that the opposition actors might be slightly more populist messages than the government
actors but this result is not statistically significant. Also, the coefficient for Facebook is
positive and it has a p-value of 0.070 which is not significant. Though it says that we see
more populist messages in Facebook than Twitter but this result is not significant.
The intercept is positive and a p value of p < 0.001, which is statistically significant.
The analysis also shows that the F statistics is significant. But the R2 is very low which
suggests that the model does not capture the overall variability.
Overall, we can say that though the opposition actors use slightly more populist messages
than the government this result is not significant at the 5% level. We reject the hypothesis
based on this analysis.
33
4.5 Hypothesis 4
To test H4 which stated populist message will receive higher engagement than non-populist
message in social media, user engagement is the dependent variable, populist Score is the
independent variable, and political position(Opposition vs Government), Social media plat-
form (Facebook vs. Twitter) are the control variable.
Since the user engagement is a count variable we will use the negative binomial regression
for this hypothesis as we did for hypothesis H2.
(Model H4)
Predictors β (SE)1
Intercept 11.501∗∗∗
(0.092)
Populist Score 0.114∗∗
(0.038)
Social media platform: Facebook -0.629∗∗∗
(0.097)
Political position: Opposition -0.460∗∗∗
(0.097)
Dispersion coefficient 0.718∗∗∗
(0.038)
Observations 918
Pseudo R2 0.0027
(1) Standard errors (SE) in parentheses
∗ p < 0.05, ∗∗ p < 0.01, ∗∗∗ p < 0.001
Table 4.4: Regression of User Engagement for Populist Score
Here we can say that the Likelihood ratio chi square statistics is significant.
The table 4.4 shows that the coefficient for populist score is positive and has a p-value of
0.003 which is statistically significant. It tells us that the more populist a message is the
more user engagement it can get.
The coefficient for Facebook is negative with a p-value 0.001 which is significant. This means
that the user engagement on average is 0.629 unit less in Facebook than Twitter. Moreover,
we have a negative coefficient for the opposition actors with a p-value of p < 0.001 which
is statistically significant. This shows that the opposition actors receive 0.460 unit less user
engagement in social media holding all the other variables constant.
34
Overall, our regression analysis shows that populist messages receive higher engagement in
social media. We accept the hypothesis.
4.6 Hypothesis 5
To test H5 which stated the social media activity of government leaders will result in a
higher level of user engagement compared to the activity of opposition leaders we select user
engagement as the dependent variable. As independent variable we have considered political
Position (Opposition vs Government), and the control variable is social media (Facebook vs
Twitter)
We are using Negative Binomial regression for this hypothesis.
(Model H5)
Predictors β (SE)1
Intercept 11.607∗∗∗
(0.087)
Political position: Opposition -0.480∗∗∗
(0.097)
Social media platform: Facebook -0.585∗∗∗
(0.097)
Dispersion coefficient 0.725∗∗∗
(0.038)
Observations 918
Pseudo R2 0.0022
(1) Standard errors (SE) in parentheses
∗ p < 0.05, ∗∗ p < 0.01, ∗∗∗ p < 0.001
Table 4.5: Regression of User Engagement for Political Position
We can see from the table 4.5, coefficient for Opposition is negative, with a p-value of
p < 0.001 which is statistically significant. It shows that the opposition actors on average
get 0.480 unit less engagement in their social media posts than the government actors. The
coefficient for Facebook is also negative with a statistically significant p-value. This tells us
Facebook posts on average get 0.585 unit less engagement than Twitter.
From this analysis we can say that government actors receive higher engagement than the
opposition actors and we accept the hypothesis.
35
Chapter 5
Discussion and Conclusion
5.1 Concluding Discussion
This study addresses the rise of populist campaign communication on social media. Social
media allows populist leaders to communicate directly with citizens and bypass traditional
news channels. This shift raises important questions. With their open and fast paced
nature social media platforms can amplify populist messages which often frame politics as
struggle between "the people" and "the elite". The motivation for this research came from it.
Examples like the Brexit vote in the UK and the Capitol attack in the US show the power of
populist messages to create division and drive political action. These events highlight how
populist messages on social media can shape political behavior and public opinion. This
study tried to explore how populist messages attract citizen on social media and how they
participate in it.
This study focuses on the period before elections, when political messages are most in-
tense. The main investigation is based on which social media platform, Facebook or Twitter,
is most effective for populist messages in terms of engagement? How do the roles of political
actors, whether in government or opposition, affect their use of populist messaging and the
engagement it generates? To answer the research questions, this study uses theories that
explain why people respond to populist messages and why they engage with them. Two key
theories guided this study: Social Identity Theory and the Civic Voluntarism Model. Social
Identity Theory helps explain why people feel drawn to populist messages. It suggests that
people support messages that align with their social group or identity. The Civic Volun-
tarism Model explains why people participate in political actions, focusing on the resources,
engagement, and networks that encourage political involvement.
The expectation was that combining these theories would reveal the factors that drives
user-engagement with populist messages. Social Identity Theory would show the appeal
of populist ideas to people’s sense of belonging. Meanwhile, the Civic Voluntarism Model
would explain the practical reasons for people’s engagement. Together, these theories were
expected to explain both the attraction to and participation in populist communication on
social media.
To answer research questions, the researcher thoroughly reviewed the previous work. One
of the major finding by Bossetta (2018) and Cinelli et al. (2021) was Facebook’s algorithm
tends to prioritize emotionally charged content, making it especially effective for amplifying
36
populist messages and fostering solidarity among users. So the first hypothesis was formed as
Facebook will be the most commonly used platform for populist messages by political actors.
While looking for the pattern, this study found that the on average the user engagement
was higher in Twitter. After performing the regression analysis for the first hypothesis,
this study rejected it, so it is clearly understandable that Facebook is not most commonly
used platform for populist messages. The second hypothesis was about user engagement
for populist messages. In the second hypothesis, it was assumed that user engagement for
populist messages is higher on Facebook than on Twitter. The hypothesis was constructed
on the previous work by Engesser et al. (2017) and Kalsnes (2019) but, it was also rejected
due to lack of significance. According to the analysis this two hypothesis can answer the
first research question. Twitter is the most effective platform for populist messages in terms
of user engagement.
The second research question focused on the role of political actors, specifically how their
position—whether in government or opposition—affects their use of populist messaging.
According to research by Mudde (2004) and Pauwels (2011), political actors in opposition
are generally in a better position to articulate populist messages, as they often criticize the
establishment and appeal directly to "the people". In contrast, government leaders tend to
moderate their tone and focus on institutional stability and broad-based appeal (Moffitt and
Tormey (2014); de Vreese et al. (2018)). Based on these assumptions, the hypothesis was
formed that political actors in opposition would use populist messages more frequently than
those in government.
However, after testing this hypothesis, the study rejected it. Contrary to expectations,
opposition political leaders did not use populist messages more frequently. While specific
cases, such as Nigel Farage’s frequent use of populist rhetoric on Twitter, align with the
pattern, the overall findings did not support the hypothesis.
The fourth hypothesis posited that populist messages would receive higher engagement
than non-populist messages, supported by the research from Engesser et al. (2017), Krämer
(2014), and Stieglitz and Dang-Xuan (2013). These studies found that populist messages,
with their simple, polarizing language and clear "us vs. them" narratives, resonate strongly
with users, encouraging active participation through likes, comments, and shares. This hy-
pothesis was accepted, confirming that populist messages indeed outperformed non-populist
content in terms of user engagement.
The final hypothesis proposed that the social media activity of government leaders would
result in higher levels of user engagement compared to opposition leaders. This was based
on prior research by Larsson and Moe (2014), Lalancette and Raynauld (2018), and Krämer
(2014), which highlighted the authority and legitimacy associated with government lead-
ers’ posts. This hypothesis was also accepted, indicating that government leaders not only
generated higher engagement but were also more active in using populist messaging.
This study answers the second research question by showing that government political
actors were more frequent users of populist messages, populist content generated higher user
engagement compared to non-populist content, and government leaders had more active and
engaging social media presence than opposition leaders.
In light of the theories applied, the findings of this study reveal key insights into how
populist messages function on social media. Social Identity Theory (SIT) helps to explain
why populist messages, which frame politics as a battle between “the people” and “the elite,”
37
resonate so strongly with audiences. Populist messages often appeal to people’s need for
belonging and identity. They create a clear in-group (“the people”) and out-group (“the
elite”), which makes it easy for users to identify with one side. This simplicity creates a
basic psychological drive to belong and to protect one’s group from perceived threats. The
high engagement with populist posts in this study which is measured by likes, shares, and
comments likely reflects users reinforcing their connection to the in-group and supporting
its values. In this way, SIT provides a strong foundation for understanding why users are
drawn to populist content and why they engage with it.
The Civic Voluntarism Model (CVM) also offers important insights into why people not
only connect with populist messages but also actively engage with them. According to
CVM, people participate more when they have the resources, psychological motivation, and
social networks that make engagement possible. Social media platforms reduce the barriers
to participation, making it easy for people to react and share without needing significant
time or effort. The emotional simplicity of populist messages on these platforms allows
people with various levels of digital skills to engage. This aligns with CVM’s focus on
accessibility and motivation, as the low-cost, high-reward nature of social media engagement
makes participation almost effortless. Additionally, social networks within these platforms
amplify engagement, as users are often influenced by seeing friends or followers interact with
similar content, which encourages them to join in.
An interesting finding in this study was that Twitter, rather than Facebook, generated
higher engagement for populist messages. This result challenges common assumptions that
Facebook, with its multimedia features and community-oriented design, would better support
engagement with populist content. Instead, Twitter’s fast-paced, real-time interaction model
may have made populist messages more visible and immediate. Twitter’s brief, direct format
aligns well with the straightforward, emotionally charged nature of populist rhetoric. This
suggests that Twitter’s design may amplify the appeal of populist messages by facilitating
quick and broad dissemination. While Facebook supports community-building, Twitter’s
rapid interactions seem better suited for the immediate, reaction-driven engagement that
populist messages generate. This finding highlights the role of platform design in shaping
how political content reaches and resonates with users and suggesting that the structure of
a social media platform can significantly influence political engagement.
Overall, these findings demonstrate the powerful bond between social identity, accessi-
ble participation, and platform design in fostering engagement with populist messages. SIT
and CVM together provide a comprehensive lens through which to understand why populist
communication is so effective on social media. The study’s insights into platform-specific
engagement also emphasize the importance of considering how different social media envi-
ronments can shape political dynamics in unique ways.
The unique political environment and campaign dynamics surrounding the U.S. and U.K.
elections could also explain why Twitter outperformed Facebook in this study. Twitter’s real-
time, concise format may have resonated more with audiences during the fast-paced election
periods, enabling quick dissemination of populist messages that needed immediate attention.
This aligns with findings from Kalsnes (2019), who observed that Twitter’s emphasis on speed
and brevity can be more effective for impactful, short-form political communication. Given
the urgent nature of election campaigns, Twitter might have become the preferred platform
for mobilizing immediate responses from supporters.
38
Individual politicians’ choices and communication styles may also account for these un-
expected results. Certain leaders, particularly those accustomed to engaging their base with
frequent updates, may prefer Twitter over Facebook. For instance, populist leaders often
utilize Twitter for its potential to spark real-time debates, escalate discussions quickly, and
directly address opponents or the public, which supports findings by Engesser et al. (2017)
on platform-specific message strategies among populist actors.
Another factor to consider is the type of engagement measured. While previous studies
found Facebook effective in fostering deep interactions (e.g., comments and long shares),
Twitter’s engagement metrics may better capture short, instant reactions like retweets. This
can amplify messages more widely during time-sensitive events.
Further more, it is also important to consider alternative explanations beyond the theories
of Social Identity Theory (SIT) and the Civic Voluntarism Model (CVM). One possible
factor is the role of social media algorithms. They are designed to prioritize content that
elicits strong emotional reactions, regardless of whether it is inherently populist (Bakshy
et al., 2015). Platforms like Facebook and Twitter use algorithms to promote posts that
generate high levels of engagement. That is often include emotionally charged or divisive
content. This means that populist messages may achieve greater visibility not solely because
of their appeal to social identity or group dynamics rather algorithms amplify content likely
to provoke responses. Thus, the high engagement observed may partly be an outcome of
platform design rather than a unique quality of populist messages themselves. Recognizing
the influence of algorithms helps to situate the findings within the broader digital media
landscape.
Another alternative explanation relates to general principles of content virality that ex-
tend beyond populism. Research on viral content suggests that messages that are simple,
emotionally resonant, and easy to understand are more likely to spread widely on social me-
dia (Berger, 2013). Populist messages often embody these qualities like using straightforward
language and framing issues in ways that elicit strong emotional responses. However, these
characteristics are not exclusive to populism. Any content that follows these principles such
as health campaigns, motivational posts, or trending challenges can also go viral. This sug-
gests that the engagement levels seen in this study could partly reflect broader social media
dynamics that favor emotionally engaging and easily shareable content, regardless of political
ideology. Considering these factors emphasizes the need to distinguish between engagement
driven by populist rhetoric and engagement driven by the general virality mechanisms that
govern social media platforms (Berger and Milkman, 2012).
5.2 Strength and Limitation
5.2.1 Strengths of the Study
This study makes a significant contribution by introducing a novel approach to blend three
different theories which are Populist Communication, Social Identity Theory and Civic Vol-
untarism Model. These theories can successfully explain citizen engagement on social media
towards populist messages. Another novel term coined by this research is triangle of populism
which are people, elite and political actors.
The paper employs a quantitative content analysis. To employ this process successfully,
39
the researcher needed to go through a rigorous process of data collection from social media
and to process it. The study checks the reliability and then validated it which enhances the
credibility and generalization of the research.
One strength of this research is its use of primary data. Collecting data directly from
social media posts offers firsthand insights into the communication strategies of political
actors. This approach allows for an accurate analysis of how populist messages are crafted
and received.
By focusing on the evolving role of social media, the paper addresses a timely topic to
ongoing debates about populism and democracy in the age of digitalization. Furthermore,
the clear and logical structure of the paper allows readers to follow the author’s arguments
and understand the implications of the findings with ease.
Last but not the least, this quantitative research present statistics and results in a way
that reader can easily visualize and understand the facts and results.
5.2.2 Limitation of the Study
First of all, this study focuses on two platforms so it does not capture what is going on with
other platforms like Instagram or TikTok. Each social media platform has its own unique
style and audience. As a result, a broader view could not be be achieved. Additionally, the
data was collected six weeks before the election but researcher makes an artificial week for
convenience. The research can be more impactful if the amount of data is more. Not only
that, studying engagement year-round might reveal different patterns when politics is less
intense.
In this study, only like, comment and share has been taken as engagement metric. But,
as some comments are missing from Donald Trump’s Twitter account, the researcher choose
to stick to like and share. More engagement metric can portray the results more accurately.
This study categorizes each variable as either 1 or 0, which can be limiting. Some mes-
sages don’t fit neatly into one category or the other but still carry populist undertones. This
approach may overlook subtleties in political messaging that are important for understanding
how political actors communicate.
In this research, author chose only politician from US and UK who are prominent. It
is a limitation. The study may not fully represent how populist messages resonate with
audiences in other regions or from different types of political figures. Other countries or
political contexts, especially from global south, could reveal different trends in how populist
messages engage people.
5.3 Future Research
First of all this study mainly focus on Facebook and Twitter. These two platform were com-
pared on the basis of which platform draws the most attention for populist communication.
It helps to understand where these messages get the most attention and why one platform
might be better suited for populist communication. But, there are other social media plat-
forms which are popular, specially among young people. According to Pew Research Center
(2021), TikTok and Instagram are used mostly by young generation compared to other social
40
media platforms. The same research could be conducted for those platforms and understand
the difference.
Secondly, the study examines how leaders use specific tactics in their messaging, such as
emphasizing “the people” versus “the elite” and calling for a return of power to ordinary citi-
zens. This thesis identifies three core dimensions of populist communication strategies: anti-
elitism, people-centrism, and restoring sovereignty. However, other communication strategies
and styles also exist that could deepen our understanding of populist messaging. The author
hopes these additional avenues will be explored in future research.
Thirdly, the thesis also explores the differences in engagement levels for populist messages
based on whether the political actors is in government or opposition. In future, a study
could track how engagement with populist messages changes overtime. Then it is possible to
investigate the shifts of populist messages when opposition leaders transition into government
roles or vice versa. It would reveal how populist messaging strategies and their effectiveness
evolve on the basis of political power.
This thesis focuses on the United States and United Kingdom and more specifically
the period before election of these two countries. In future, it can be possible to do sim-
ilar research for other countries context. Furthermore, comparing populist communication
strategies in countries with different political systems (e.g., multiparty systems vs. two-party
systems) could provide insights. Another possibility is to Investigating whether engagement
with populist content fluctuates during different political cycles, such as election versus
non-election periods. This could add depth to the understanding of populist engagement
patterns.
In this thesis, only three engagement metric (e.g., like, comment and share) are under
consideration. But, there are other metric available for different platforms. Examining those
emotional responses of audiences to populist messages reveal new avenue in future research.
As here is used quantitative content analysis as research method, but in future possibly using
sentiment analysis or survey methods may uncover how the tone of populist rhetoric (e.g.,
anger, solidarity) influences engagement levels.
Additionally, future research could be possible by blending methods. That means, com-
bining quantitative analysis with qualitative approaches like interviews or discourse analysis.
That will capture the subtleties of how populist messages resonate with audiences of gov-
ernment and opposition political actor. It can be also possible to do experimental studies
which could explore how people react to simulated populist messages from political actors. It
will be possible to pinpoint what drives engagement through this. Lastly, machine learning
could add a predictive angle by identifying what aspects of a message are most likely to spark
engagement for government versus opposition leaders. This is a more advanced technique
for coder reliability or software-assisted coding to improve accuracy, particularly for larger
datasets so that the findings might portray big picture.
The author hope these indications will encourage researchers in different fields to work
on the triangle of populism and social media.
41
Bibliography
Akkerman, T., de Lange, S., & Rooduijn, M. (Eds.). (2016). Radical right-wing populist
parties in western europe: Into the mainstream? (1st). Routledge. https://doi.org/
10.4324/9781315687988
Albertazzi, D., & McDonnell, D. (2008). Introduction: The sceptre and the spectre. In D.
Albertazzi & D. McDonnell (Eds.), Twenty-first century populism: The spectre of
western european democracy (pp. 1–11). Palgrave.
Aylott, N., & Bolin, N. (2023). A new right: The swedish parliamentary election of september
2022. West European Politics, 46 (5), 1049–1062. https://doi.org/10.1080/01402382.
2022.2156199
Bakshy, E., Messing, S., & Adamic, L. A. (2015). Exposure to ideologically diverse news
and opinion on facebook. Science, 348 (6239), 1130–1132. https://doi.org/10.1126/
science.aaa1160
Bandura, A. (1997). Self-efficacy: The exercise of control. W.H. Freeman; Company.
Bartels, J.-E., & Remke, M. L. (2023). ‘parties in the shadows’ –do small and marginal
parties cater to a niche? German Politics, 32 (2), 267–298. https://doi.org/10.1080/
09644008.2021.1913725
Behr, R. (2024). Britain and the us face vital elections. the decline of democracy is at stake
[The Guardian]. Retrieved November 3, 2024, from https://www.theguardian.com/
commentisfree/2024/jan/03/britain-us-elections-democracy-decline-donald-trump
Berger, J. (2013). Contagious: Why things catch on. Simon & Schuster.
Berger, J., & Milkman, K. L. (2012). What makes online content viral? Journal of Marketing
Research, 49 (2), 192–205. https://doi.org/10.1509/jmr.10.0353
Bertot, J. C., Jaeger, P. T., & Grimes, J. M. (2012). Promoting transparency and ac-
countability through icts, social media, and collaborative e-government. Transform-
ing Government: People, Process and Policy, 6 (1), 78–91. https://doi.org/10.1108/
17506161211214831
Blassnig, S., Ernst, N., Engesser, S., & Esser, F. (2019). Populism and social media popular-
ity: How populist communication benefits political leaders on facebook and twitter.
In Power shift? political leadership and social media (1st, p. 15). Routledge. https:
//doi.org/10.4324/9780429466007
Block, E., & Negrine, R. (2017). The populist communication style: Toward a critical frame-
work. International Journal of Communication, 11, 20.
Bobba, G. (2019). Social media populism: Features and ’likeability’ of lega nord communica-
tion on facebook. European Political Science, 18 (1), 11–23. https://doi.org/10.1057/
s41304-017-0141-8
42
Bobba, G., & McDonnell, D. (2016). Different types of right-wing populist discourse in
government and opposition: The case of italy. South European Society and Politics,
21 (3), 281–299. https://doi.org/10.1080/13608746.2016.1211239
Bonikowski, B., & Gidron, N. (2016). The populist style in american politics: Presidential
campaign discourse, 1952–1996. Social Forces, 94 (4), 1593–1621. https://doi.org/10.
1093/sf/sov120
Bos, L., & Brants, K. (2014). Populist rhetoric in politics and media: A longitudinal study
of the netherlands. European Journal of Communication, 29 (6), 703–719. https://
doi.org/10.1177/0267323114545709
Bossetta, M. (2018). The digital architectures of social media: Comparing political cam-
paigning on facebook, twitter, instagram, and snapchat in the 2016 u.s. election.
Journalism & Mass Communication Quarterly, 95 (2), 471–496. https://doi.org/10.
1177/1077699018763307
Brady, H. E., Schlozman, K. L., & Verba, S. (1999). Beyond ses: A resource model of political
participation. American Journal of Political Science, 43 (2), 271–294. https://doi.org/
10.2307/2082425
Bruns, A. (2020). Real-time applications (twitter). In Handbuch soziale praktiken und digitale
alltagswelten (pp. 87–95). Springer. https://doi.org/10.1007/978-3-658-08357-1_8
Bryman, A. (2016). Social research methods. Oxford university press.
Bugaric, B. (2019). The two faces of populism: Between authoritarian and democratic pop-
ulism. German Law Journal, 20 (3), 390–400. https://doi.org/10.1017/glj.2019.20
Canovan, M. (1999). Trust the people! populism and the two faces of democracy. Political
Studies, 47 (1), 2–16. https://doi.org/10.1111/1467-9248.00184
Cassell, K. J. (2021). The populist communication strategy in comparative perspective. The
International Journal of Press/Politics, 28, 725–746. https : / / doi . org / 10 . 1177 /
19401612211055695
Chadwick, A. (2013). The hybrid media system: Politics and power. Oxford University Press.
Cinelli, M., De Francisci Morales, G., Galeazzi, A., Quattrociocchi, W., & Starnini, M. (2021).
The echo chamber effect on social media. Proceedings of the National Academy of
Sciences, 118 (9), e2023301118. https://doi.org/10.1073/pnas.2023301118
CNN. (2021, January). Twitter permanently suspends president donald trump’s account [Ac-
cessed: 2024-11-10]. https ://edition . cnn .com/2021/01/08/tech/trump- twitter -
ban/index.html
Collins, M. (2023). The real lesson to be learned from twitter’s rebrand. Retrieved July 30,
2023, from https ://www.forbes .com/sites/marcuscollins/2023/07/30/the- real -
lesson-to-be-learned-from-twitters-rebrand/
Creswell, J. W. (2014). Research design: Qualitative, quantitative, and mixed methods ap-
proaches. Sage publications.
Crothers, L., & Burgener, G. (2021). Insurrectionary populism? assessing the january 6 attack
on the u.s. capitol. Populism, 4 (2), 129–145. https://doi.org/10.1163/25888072-
bja10025
de Vreese, C. H., Esser, F., Aalberg, T., Reinemann, C., & Stanyer, J. (2018). Populism as
an expression of political communication content and style: A new perspective. The
International Journal of Press/Politics, 23 (4), 423–438. https://doi.org/10.1177/
1940161218790035
43
De Bruycker, I., & Rooduijn, M. (2021). The people’s champions? populist communication
as a contextually dependent political strategy. Journalism Mass Communication
Quarterly, 98, 896–922. https://doi.org/10.1177/1077699021998648
Delli Carpini, M. X. (2004). Mediating democratic engagement: The impact of communica-
tions on citizens’ involvement in political and civic life. In L. L. Kaid (Ed.), Handbook
of political communication research (pp. 395–434). Lawrence Erlbaum Associates.
D’heer, E., & Verdegem, P. (2014). Conversations about the elections on twitter: Towards a
structural understanding of twitter’s relation with the political and the media field.
European Journal of Communication, 29 (6), 720–734. https : //doi . org/10 .1177/
0267323114544866
Engesser, S., Ernst, N., Esser, F., & Büchel, F. (2017). Populism and social media: How politi-
cians spread a fragmented ideology. Information, Communication Society, 20 (8),
1109–1126.
Enli, G., & Moe, H. (2013). Introduction to special issue on social media and election cam-
paigns—key tendencies and ways forward. Information, Communication & Society,
16 (5), 637–645.
Enli, G. S., & Skogerbø, E. (2013). Personalized campaigns in party-centred politics: Twitter
and facebook as arenas for political communication. Information, Communication &
Society, 16 (5), 757–774. https://doi.org/10.1080/1369118X.2013.782330
Ernst, N., Engesser, S., Büchel, F., Blassnig, S., & Esser, F. (2017). Extreme parties and
populism: An analysis of facebook and twitter across six countries. Information, Com-
munication & Society, 20 (9), 1347–1364.
Etikan, I., Musa, S. A., & Alkassim, R. S. (2016). Comparison of convenience sampling and
purposive sampling. American Journal of Theoretical and Applied Statistics, 5 (1),
1–4. https://doi.org/10.11648/j.ajtas.20160501.11
Fernández-García, B., & Luengo, Ó. G. (2020). Populist contagion or anti-elitism in oppo-
sition? psoe’s response to the emergence of podemos. Revista Española De Ciencia
Política, (54), 13–37. https://doi.org/10.21308/recp.54.01
Freeden, M. (1998). Is nationalism a distinct ideology? Political Studies, 46 (4), 748–765.
https://doi.org/10.1111/1467-9248.00165
Gandini, A., Ceron, A., & Lodetti, P. (2022). Populists or influencers? the use of facebook
videos by populist leaders. International Journal of Communication, 16, 1–20.
Geert wilders abandons bid to become dutch prime minister. (2023, May). https://www.bbc.
com/news/world-europe-68557036
Goodwyn, L. (1976). Democratic promise: The populist moment in america. Oxford Univer-
sity Press.
Hameleers, M., Bos, L., & de Vreese, C. H. (2017). "they did it": The effects of emotionalized
blame attribution in populist communication. Communication Research, 44 (6), 870–
900. https://doi.org/10.1177/0093650216644026
Hameleers, M., van der Meer, T. G. L. A., & Boumans, J. W. (2023). Communications,
48 (4), 588–607. https://doi.org/doi:10.1515/commun-2021-0105
Harwood, J. (2020). Social identity theory. In The international encyclopedia of media psy-
chology (pp. 1–7). John Wiley Sons, Ltd. https://doi.org/https://doi.org/10.1002/
9781119011071.iemp0153
44
Hawkins, K. A. (2009). Is chavez populist?: Measuring populist discourse in comparative
perspective. Comparative Political Studies, 42 (8), 1040–1067. https://doi .org/10.
1177/0010414009331721
Hermida, A. (2010). Twittering the news: The emergence of ambient journalism. Journalism
Practice, 4 (3), 297–308. https://doi.org/10.1080/17512781003640703
Hester, J. B., & Dougall, E. (2007). The efficiency of constructed week sampling for content
analysis of online news. Journalism & Mass Communication Quarterly, 84 (4), 811–
824. https://doi.org/10.1177/107769900708400410
Hsieh, H.-F., & Shannon, S. E. (2005). Three approaches to qualitative content analysis.
Qualitative health research, 15 (9), 1277–1288.
Iyengar, S., & Simon, A. F. (2000). New perspectives and evidence on political communi-
cation and campaign effects. Annual Review of Psychology, 51 (1), 149–169. https:
//doi.org/10.1146/annurev.psych.51.1.149
Jagers, J., & Walgrave, S. (2007). Populism as political communication style: An empir-
ical study of political parties’ discourse in belgium. European Journal of Political
Research, 46 (3), 319–345. https://doi.org/10.1111/j.1475-6765.2006.00690.x
Kalsnes, B. (2016). The social media paradox explained: Comparing political communication
on social media and television. International Journal of Digital Television, 7 (3), 277–
293.
Kalsnes, B. (2019). Examining the populist communication logic: Strategic use of social
media in populist political parties in norway and sweden. Central European Journal
of Communication, 12 (2), 187–205. https://doi.org/10.19195/1899-5101.12.2(23).5
Kaltwasser, C. R. (2012). The ambivalence of populism: Threat and corrective for democracy.
Democratization, 19 (2), 184–208. https://doi.org/10.1080/13510347.2011.572619
Katz, J. E., Barris, M., & Jain, A. (2013). The social media president: Barack obama and
the politics of digital engagement. Palgrave Macmillan.
Krämer, B. (2014). Media populism: A conceptual clarification and some theses on its effects.
Communication Theory, 24 (1), 42–60. https://doi.org/10.1111/comt.12029
Krämer, B. (2017). Populist online practices: The function of the internet in right-wing
populism. Information, Communication & Society, 20 (9), 1293–1309. https://doi.
org/10.1080/1369118X.2017.1328520
Kreis, R. (2017). The “tweet politics” of president trump. Journal of Language and Politics,
16 (4), 607–618. https://doi.org/10.1075/jlp.17032.kre
Kriesi, H. (2013). Conceptualizing the populist challenge. Johns Hopkins University.
Kriesi, H. (2014). The populist challenge. West European Politics, 37 (2), 361–378. https:
//doi.org/10.1080/01402382.2014.887879
Krippendorff, K. (2018). Content analysis: An introduction to its methodology. Sage publi-
cations.
Laclau, E. (2005). On populist reason. Verso.
Lalancette, M., & Raynauld, V. (2018). The power of political image: Justin trudeau, insta-
gram, and celebrity politics. American Behavioral Scientist, 62 (2), 243–259. https:
//doi.org/10.1177/0002764217744838
Larsson, A. O., & Ihlen, Ø. (2015). Birds of a feather flock together? party leaders on twitter
during the 2013 norwegian elections. European Journal of Communication, 30 (6),
666–681. https://doi.org/10.1177/0267323115595525
45
Larsson, A. O., & Moe, H. (2014). Triumph of the underdogs? comparing twitter use by
political actors during two norwegian election campaigns. SAGE Open, 4 (4), 1–13.
https://doi.org/10.1177/2158244014559015
Lee, N. M., & VanDyke, M. S. (2015). Set it and forget it: The one-way use of social media
by government agencies communicating science. Science Communication, 37 (4), 533–
541. https://doi.org/10.1177/1075547015588600
McGregor, S. C., Lawrence, R. G., & Cardona, A. (2017). Personalization, gender, and social
media: Gubernatorial candidates’ social media strategies. Information, Communica-
tion & Society, 20 (2), 264–283. https://doi.org/10.1080/1369118X.2016.1167228
Mény, Y., & Surel, Y. (2002). The constitutive ambiguity of populism. In Y. Mény & Y.
Surel (Eds.), Democracies and the populist challenge (pp. 1–24). Palgrave.
Milroy, L. (2000). Britain and the united states: Two nations divided by the same language
(and different language ideologies). Journal of Linguistic Anthropology, 10 (1), 56–89.
Retrieved November 3, 2024, from http://www.jstor.org/stable/43103225
Moffitt, B. (2016). The global rise of populism: Performance, political style, and representa-
tion. Stanford University Press.
Moffitt, B., & Tormey, S. (2014). Rethinking populism: Politics, mediatisation and political
style. Political Studies, 62 (2), 381–397. https://doi.org/10.1111/1467-9248.12032
Mudde, C. (2004). The populist zeitgeist. Government and Opposition, 39 (4), 541–563. Re-
trieved March 10, 2024, from http://www.jstor.org/stable/44483088
Neuendorf, K. A. (2017). The content analysis guidebook. Sage publications.
Norris, P., Curtice, J., Sanders, D., Scammell, M., & Semetko, H. A. (1999). On message:
Communicating the campaign. Sage Publications.
Obama warns against irresponsible social media use. (2017, December). https://www.bbc.
com/news/world-us-canada-42491638
Ott, B. L. (2017). The age of twitter: Donald j. trump and the politics of debasement. Critical
Studies in Media Communication, 34 (1), 59–68. https://doi.org/10.1080/15295036.
2016.1266686
Parmelee, J. H., & Bichard, S. L. (2011). Politics and the twitter revolution: How tweets
influence the relationship between political leaders and the public. Lexington Books.
Paterson, M., Wilshire, S., & Tobin, P. (2023). The rise of anti-net zero populism in the uk:
Comparing rhetorical strategies for climate policy dismantling. Journal of Compara-
tive Policy Analysis: Research and Practice, 26 (3–4), 332–350. https://doi.org/10.
1080/13876988.2023.2242799
Pauwels, T. (2011). Measuring populism: A quantitative text analysis of party literature
in belgium. Journal of Elections, Public Opinion and Parties, 21 (1), 97–119. https:
//doi.org/10.1080/17457289.2011.539483
Pew Research Center. (2021). Social media use in 2021 [Accessed: 2021-04-07]. https://www.
pewresearch.org/internet/2021/04/07/social-media-use-in-2021/
Piazza, J. A. (2024). Populism and support for political violence in the united states: As-
sessing the role of grievances, distrust of political institutions, social change threat,
and political illiberalism. Political Research Quarterly, 77 (1), 152–166. https://doi.
org/10.1177/10659129231198248
Priester, K. (2011). Definitionen und typologien des populismus [definitions and typologies
of populism]. Soziale Welt, 62 (2), 185–198. http://www.jstor.org/stable/23060014
46
Project, T. A. P. (2024). Advanced search results for speeches and remarks by president
donald trump. Retrieved November 4, 2024, from https://www.presidency.ucsb.edu/
advanced- search?field - keywords=&field - keywords2=&field - keywords3=&from%
5Bdate%5D=09 - 15 - 2020& to%5Bdate%5D=11 - 05 - 2020&person2=200301&
category2%5B%5D=423&items_per_page=100
Putnam, R. D. (2000). Bowling alone: The collapse and revival of american community.
Simon & Schuster.
Reicher, S. D., Spears, R., & Postmes, T. (1995). A social identity model of deindividuation
phenomena. European Review of Social Psychology, 6 (1), 161–198. https://doi.org/
10.1080/14792779443000049
Riffe, D., Lacy, S., & Fico, F. (2019). Analyzing media messages: Using quantitative content
analysis in research. Routledge.
Rooduijn, M. (2013). The mesmerising message: The diffusion of populism in public debates
in western european media. Political Studies, 62 (4), 726–744.
Rosenstone, S. J., & Hansen, J. M. (1993). Mobilization, participation, and democracy in
america. Macmillan.
Ruth-Lovell, S. P., & Grahn, S. (2022). Threat or corrective to democracy? the relationship
between populism and different models of democracy. European Journal of Political
Research, 62 (3), 677–698. https://doi.org/10.1111/1475-6765.12564
Silver, L. (2022). Populists in europe, especially those on the right, have increased their vote
shares in recent elections [Accessed: 2024-05-29].
Sintes-Olivella, M., Casero-Ripollés, A., & Yeste-Piquer, E. (2020). The inclusionary populist
communication style on facebook: The case of ada colau in barcelona. Communication
and Society, 33, 193–208. https://doi.org/10.15581/003.33.2.193-208
Smith, D., Deacon, D., & Downey, J. (2021). Inside out: The uk press, brexit and strategic
populist ventriloquism. European Journal of Communication, 36 (1), 21–37. https :
//doi.org/10.1177/0267323120940917
Smith, J. (2020). Social media strategies in political campaigns: Timing and impact. Journal
of Political Communication, 35 (4), 457–478.
Stanley, B. (2008). The thin ideology of populism. Journal of Political Ideologies, 13 (1),
95–110. https://doi.org/10.1080/13569310701822289
Stieglitz, S., & Dang-Xuan, L. (2013). Emotions and information diffusion in social me-
dia—sentiment of microblogs and sharing behavior. Journal of Management Infor-
mation Systems, 29 (4), 217–248. https://doi.org/10.2753/MIS0742-1222290408
Stier, S., Bleier, A., Lietz, H., & Strohmaier, M. (2018). Election campaigning on social
media: Politicians, audiences, and the mediation of political communication on face-
book and twitter. Political Communication, 35 (1), 50–74. https://doi.org/10.1080/
10584609.2017.1334728
Taggart, P. (2000). Populism. Open University Press.
Tajfel, H. (1981). Human groups and social categories: Studies in social psychology. Cam-
bridge University Press.
Tajfel, H., & Turner, J. C. (1979). An integrative theory of intergroup conflict. In W. G.
Austin & S. Worchel (Eds.), The social psychology of intergroup relations (pp. 33–47).
Brooks/Cole.
47
Tajfel, H., & Turner, J. C. (1986). The social identity theory of intergroup behavior. In
S. Worchel & W. G. Austin (Eds.), Psychology of intergroup relations (pp. 7–24).
Nelson-Hall.
The Guardian. (2020, September 22). Joe biden accuses trump of ’not doing his job’ as us
covid-19 death toll nears 200,000 [Published: Tue 22 Sep 2020 00:18 CEST]. Retrieved
December 9, 2023, from https://amp.theguardian.com/us-news/video/2020/sep/21/
joe-biden-trump-coronavirus-deaths
Trochim, W. M. K. (2020). Research methods: The essential knowledge base. Cengage Learn-
ing.
Tucker, J. A., Guess, A., Barbera, P., Vaccari, C., Siegel, A., Sanovich, S., Stukal, D., &
Nyhan, B. (2018). Social media, political polarization, and political disinformation:
A review of the scientific literature. SSRN Electronic Journal. https://doi.org/10.
2139/ssrn.3144139
Turner, J. C. (1982). Towards a cognitive redefinition of the social group. In H. Tajfel (Ed.),
Social identity and intergroup relations (pp. 15–40). Cambridge University Press.
van Dijck, J. (2013). The culture of connectivity: A critical history of social media. Oxford
University Press. https://doi.org/10.1093/acprof:oso/9780199970773.001.0001
Velasquez, A., & LaRose, R. (2015). Social media for social change: Social media political
efficacy and activism in student activist groups. Journal of Broadcasting & Electronic
Media, 59 (3), 456–474. https://doi.org/10.1080/08838151.2015.1054998
Verba, S., Schlozman, K. L., & Brady, H. E. (1995). Voice and equality: Civic voluntarism
in american politics. Harvard University Press.
Vogel, E. A., Rose, J. P., Roberts, L. R., & Eckles, K. (2014). Social comparison, social
media, and self-esteem. Psychology of Popular Media Culture, 3 (4), 206–222. https:
//doi.org/10.1037/ppm0000047
Vossen, K. (2011). Classifying wilders: The ideological development of geert wilders and
his party for freedom. Politics, 31 (3), 179–189. https://doi .org/10.1111/j .1467-
9256.2011.01417.x
Waisbord, S., & Amado, A. (2017). Populist communication by digital means: Presidential
twitter in latin america. Information, Communication & Society, 20 (9), 1330–1346.
https://doi.org/10.1080/1369118X.2017.1328521
Weber, R. P. (1990). Basic content analysis. Sage.
Wettstein, M., Esser, F., Engesser, S., Wirth, W., Wirz, D., Schulz, A., Büchel, F., Ernst,
N., Caramani, D., Manucci, L., Bernhard, L., Steenbergen, M., Hänggli, R., Dalmus,
C., & Weber, E. (2016, May). The appeal of populist ideas, strategies and styles: A
theoretical model and research design for analyzing populist political communication.
https://doi.org/10.13140/RG.2.2.20358.37440
Wirth, W., Esser, F., Wettstein, M., Engesser, S., Wirz, D., Schulz, A., Ernst, N., Büchel, F.,
Caramani, D., Manucci, L., Steenbergen, M., Bernhard, L., Weber, E. C., Hänggli, R.,
Dalmus, C., Schemer, C., & Müller, P. (2016). The appeal of populist ideas, strategies
and styles: A theoretical model and research design for analyzing populist political
communication. NCCR Democracy at the University of Zurich. https://doi.org/10.
13140/RG.2.2.20358.37440
Wojcik, S., & Hughes, A. (2019). Sizing up twitter users [Pew Research Center]. https :
//www.pewresearch.org/internet/2019/04/24/sizing-up-twitter-users/
48
Yabanci, B. (2018). The akp after 15 years: Emergence of erdoganism in turkey. Third World
Quarterly, 39 (9), 1812–1830. https://doi.org/10.1080/01436597.2018.1447371
49
Appendix A
Literature Search Process
The selection process for this literature review on populist communication on social media
by political actors before an election involves several systematic steps. The review focuses
on two main research questions. The first question examines the extent to which political
actors engage in populist campaigns on social media before an election and identifies the most
commonly used social media platform for these campaigns. The second question explores
the level of engagement between political actors and citizens on social media , comparing the
engagement levels of populist versus non-populist messages, and determining which platform,
Facebook or Twitter, garners more user engagement. The search for relevant literature uses
a comprehensive list of keywords, including ” Populism,” ”Populist Communication,” ”Social
Media,” ”Populist Campaign,” ”Social Identity theory,” ”Civic VoluntarismModel,” ”Populism
in Power,” and ”Populism in Opposition.” Different combinations of these keywords guide
the search across Google Scholar, Scopus, and Gothenburg University Library. From these
sources, a total of approximately 16,084, 15,466, and 18,985 papers respectively are identified.
The criteria for including papers are strict; they must relate to the main themes and be
peer-reviewed. Papers not meeting these criteria are excluded to ensure the review remains
focused and of high quality. The screening process begins with reading titles and abstracts to
identify relevant studies. Full-text reviews follow, focusing on the methodology and results
sections to assess the quality and relevance of each paper. Of the vast pool of identified
papers, 21 research papers are selected and used in this review. Thematic analysis organizes
these selected papers. This approach ensures a comprehensive and current overview of the
field.
50
Appendix B
Codebook
Table B.1: Codebook for Variables
Dimension Variable Response Alter- Coding Instruction
native
Format
V1. Politician 1 = Nigel Farage What is the political leader’s
Identification name?
2 = Boris Johnson
3 = Donald
Trump
4 = Joe Biden
51
Dimension Variable Response Alter- Coding Instruction
native
V2. Identification Serial number Each tweet or Facebook post
Number is given a unique ID as fol-
lows:
• 6-digit ID
• 1st digit starts with
either one of number
between 1 to 4
1 = Nigel
2 = Boris
3 = Trump
4 = Biden
• 2nd digit will be either
1 or 2.
1 = Tweet
2 = Facebook
• Last 4 digits: unique
post identifier
• Example: 210001 is
Boris Johnson’s first
Twitter post.
V3. Date Format: Date post was published.
MMDDYYYY
V4. Share Numerical Value Number of shares (Facebook)
or retweets (Twitter).
Engagement
V5. Comment Numerical Value Number of comments (Face-
book) or replies (Twitter).
52
Dimension Variable Response Alter- Coding Instruction
native
V6. Like Numerical Value Number of likes (Facebook)
or favorites (Twitter).
Anti-Elitism V7a. Discrediting 0 = No Are elites accused of being
the elite 1 = Yes malevolent, criminal, lazy,
stupid, extremist, racist,
undemocratic, etc. The elite
are called names and denied
morality, charisma, credibil-
ity, intelligence, competence,
consistency?
If no then insert 0 otherwise
1.
V7b. Blaming the 0 = No Are elites described as a
elite 1 = Yes threat/burden, responsible
for negative developments/si-
tuations, or as having
committed a mistake or
crime?
Are elites described as not
being a source of enrichment
or responsible for a positive
development/situation?
If no then insert 0 otherwise
1.
53
Dimension Variable Response Alter- Coding Instruction
native
V7c. Detaching the 0 = No Are elites described as not
elite from the people 1 = Yes belonging to the people, not
being close to the people,
not knowing the people,
not speaking for the people,
not caring for them, or not
performing everyday actions?
If no then insert 0 otherwise
1.
V8a. Stressing the 0 = No Are the people bestowed with
People-Centrism people’s virtues 1 = Yes morality, charisma, credibil-
ity, intelligence, competence,
consistency, etc.
Are the people exempt from
being malevolent, criminal,
lazy, stupid, extremist, racist,
undemocratic, etc.?
If no then insert 0 otherwise
1.
V8b. Praising the 0 = No Are the people described
people’s achieve- 1 = Yes as being an enrichment or
ments responsible for a positive
development/situation?
Are the people described
as not being a threat/bur-
den, not being responsible
for a negative developmen-
t/situation, nor as having
committed a mistake or
crime?
If no then insert 0 otherwise
1.
54
Dimension Variable Response Alter- Coding Instruction
native
V8c. Stating a 0 = No Are people described as
monolithic people 1 = Yes sharing common feelings,
desires, or opinions?
If no then insert 0 otherwise
1.
V8d. Demonstrat- 0 = No Does the speaker describe
ing closeness to the 1 = Yes himself as belonging to the
people people, being close to the
people, knowing the people,
speaking for the people, car-
ing for the people, agreeing
with the people, or perform-
ing everyday actions?
Does the speaker claim to
represent or embody the
people?
If no then insert 0 otherwise
1.
Restoring Sovereignty
V9a. Demanding 0 = No Does the speaker argue
popular sovereignty 1 = Yes for general institutional
reforms to grant the people
more power (by introducing
direct-democratic elements
or increasing political partic-
ipation)?
Does the speaker argue in
favor of granting more power
to the people within the
context of a specific issue
(e.g., election, immigration,
security)?
If no then insert 0 otherwise
1.
55
Dimension Variable Response Alter- Coding Instruction
native
V9b. Denying elite 0 = No Does the speaker argue in
sovereignty 1 = Yes favor of granting less power
to elites within the context of
a specific issue (e.g., election,
immigration, security)?
If no then insert 0 otherwise
1.
Note. For missing values or technical errors, insert 999.
56
Appendix C
Inter-coder Reliablity Test and Cohen’s
Kappa
Table C.1: Intercoder Reliability Test Results
Variable Percentage of Agreement Cohen’s Kappa
Discrediting the elite 92.39% 0.74
Blaming the elite 95.65% 0.87
Detaching the elite from the people 97.83% 0.94
Stressing the people’s virtues 100.00% 1.00
Praising the people’s achievements 95.65% 0.79
Stating a monolithic people 94.57% 0.79
Demonstrating closeness to the people 94.57% 0.75
Demanding popular sovereignty 96.74% 0.55
Denying elite sovereignty 96.74% 0.55
Note: The table shows inter-coder reliability results, including percentage of agreement
and Cohen’s Kappa for each variable.
57