World Leaders a-Twitter: Communication Platforms and Agenda-Building During the 2018 NATO Summit

Katja C. Walton International Affairs Program University of Colorado Boulder

Defended April 2, 2019

Honors Thesis Defense Committee

Dr. Janet Lynn Donavan, Primary Advisor Department of Political Science

Dr. Douglas Snyder, Honors Council Representative International Affairs Program

Dr. Levente Szentkirályi, Thesis Committee Member Program for Writing and Rhetoric

World Leaders a - T w i t t e r | 2

Abstract:

Twitter is a thriving microblogging service with growing prominence in the political sphere. In this study, I examine the differences between Twitter communications and verbal communications by three heads of state and government in relation to the most recent NATO Summit in July 2018. Through a three-step analysis, including descriptive statistics, content and tone analysis, and comparative analysis, the study investigates Twitter’s influence on content and tone and its agenda-building capacity for face-to-face summits. After hand-coding over 2,000 tweets and 15 verbal communications, I find that Twitter does not support more negative content and tone among world leaders than verbal communication. Rather, a leader’s tone remains consistent on both communication platforms, suggesting the salience of personality and political strategy as well as the importance of anonymity in online behavioral disinhibition. Findings also demonstrate that, in the case of Burden-Sharing negotiations during the 2018 NATO Summit, U.S. President Trump successfully implemented Twitter as an agenda-building tool to position Burden-Sharing as a prominent Summit topic. Ultimately, I conclude that the rejection of the platform’s legitimacy for diplomatic exchanges and the lack of direct discussion between politicians on Twitter demonstrates that Twitter is not a viable replacement for face-to-face summits. World Leaders a - T w i t t e r | 3

Table of Contents

Introduction...... 4

Background...... 8

Twitter………………………………………………………………………………………………………….………………..8

The North Atlantic Treaty Organization……………………………………………………..………………….10

Literature Review...... 12

Theory & Guiding Hypotheses...... 20

Data & Methodology...... 22

Descriptive Statistics...... 30

Twitter Summary Statistics……………………………………………………………………………..….…………30

Verbal Communication Summary Statistics……………………………………………………..…….……..33

Content and Tone Analysis...... 34

Content and Tone on Twitter...... 34

Content and Tone Through NATO Tweets...... 42

Content and Tone Through Verbal Communication...... 47

Comparative Analysis...... 52

Comparing Content: NATO Tweets vs. Verbal Communications...... 52

Comparing Tone: NATO Tweets vs. Verbal Communications...... 56

How Do Communications Relate to the Agenda? ...... 57

Conclusion.…………………………………………………………………………………………….…………………..……...……60

Appendix...... 67

Bibliography...... 73

World Leaders a - T w i t t e r | 4

Introduction

The increasing prominence of social media as an extension of social and political spheres poses a dual-use challenge to today’s population. Large bodies of literature have evaluated how social media have both benefitted societies and become weaponized. In 2011, social media perpetuated a democratic push across the Middle East and northern Africa that toppled authoritarian regimes in what became known as the Arab Spring. Social media have also given voices to the powerless, evidenced by hashtag activism, such as #BlackLivesMatter, #MeToo, and #Kony2012. Nevertheless, digital platforms have also served as the battlegrounds for disinformation and the undermining of democracy. Facebook, Twitter, and YouTube, for example, have all been used to disseminate malicious state and non-state propaganda.

In the wake of an active and politicized cyberspace, heads of state and government from all around the world have flocked to their keyboards and smartphones to disseminate their own messages. Twitter, in particular, has been a favored platform for politics, popularized by the

45th President of the United States, Donald J. Trump (Lüfkens 2018; Parmelee and Bichard

2012). With Twitter’s rising political popularity, heads of state and government have been using

Twitter to communicate with both their constituents and with one another. Twitter has therein provided the international system with a new platform to voice the successes and grievances of individual nations, moving conversations between world leaders out from behind closed doors and rolling cameras and into cyberspace.

Before social networking sites like Twitter, world leaders depended heavily on traditional mass media and face-to-face meetings to propagate their messages around the globe and to other leaders. Twitter, however, has permitted leaders to overcome these World Leaders a - T w i t t e r | 5 restrictions, providing them with 24-hour access to their domestic and international publics.

Although stark differences exist between verbal communications, such as at international summits and press conferences, and computer mediated communications in the Twittersphere, many world leaders still choose to actively engage in the platform. In fact, of all current United

Nations members, 97 percent have an official Twitter presence (Lüfkens 2018). Among these state representatives, President Donald Trump has become the most followed world leader on

Twitter with 58.4 million followers on his personal account, @realDonaldTrump, while his official account, @POTUS, ranks fifth with 25.2 million followers.1

Figure 1: Most Followed World Leaders in 2019 70,000,000

60,000,000

50,000,000

40,000,000

30,000,000 Followers 20,000,000

10,000,000

0

*Vatican data is a total of all nine different language accounts.

1 Data as of February 20, 2019. World Leaders a - T w i t t e r | 6

In the Age of the Twitter President, scholars have been investigating how and why politicians use Twitter. Despite the large body of academic work addressing these questions, few studies branch out beyond the United States to investigate international politicians’ implementation of Twitter, and none compare the platform to verbal political communications to investigate a difference in content and tone, or to understand Twitter’s agenda-building effects on face-to-face negotiations. This study seeks to fill this gap in the literature by evaluating Trump-era Twitter usage by three Western leaders in the context of the 2018 NATO

Summit.

The world leaders selected for this study include U.S. President Donald Trump, French

President , and Canadian Prime Minister – all of whom possess verified Twitter accounts. These world leaders have been selected due to their overwhelming popularity on Twitter with over 3.5 million followers each. These three are also interesting because of their positions as the leaders of powerful nations in the current international system that belong to the twelve original NATO members. Additionally, they were selected due to language restrictions since only English and French could be accurately understood and analyzed by the author and translator.

The 2018 NATO Summit sets the scene for this qualitative approach since all NATO member nations have official representation on Twitter, and the recent Summit on July 11-12,

2018 produced contemporary Twitter and verbal communication for data. Additionally, NATO summits follow a traditional diplomatic framework that is well-covered by the press, and as a stable 70-year-old alliance it will likely continue meeting for years to come, allowing future research to readdress this paper’s findings. Thus, the content and tone of three world leaders’ World Leaders a - T w i t t e r | 7 on- and offline communications will be compared, and communications will be evaluated with respect to the Summit agenda. Accordingly, I format my research questions:

(1) Do communication platforms influence the content and tone of statements made by heads of state and government?

(2) Do different communication strategies and content across platforms affect official, in- person summits through agenda-building?

This paper begins by describing in greater detail the mechanics and evolution of Twitter as well as an overview of NATO and its 2018 Summit in July. Then, I describe the existing literature concerning the differences between verbal and computer mediated communications both socially and politically, along with a synthesis of agenda-building studies. Much of this literature contributes to the theoretical basis for my arguments, so I subsequently discuss the relationship between these theories and my hypotheses. I first argue that Twitter contributes to greater negative content and tone than verbal communications because of online behavioral disinhibition, and next I argue that Twitter helps politicians build the agenda for face-to-face summits by producing a larger volume of tweets related to the topics they find most important than if communication was restricted to verbal communications alone. In order to test these hypotheses, I integrate descriptive statistics to summarize the data, conduct a content and tone analysis of both verbal and Twitter communications among the three world leaders, and offer a comparative analysis between the two communication platforms and their relationship to the

2018 NATO Summit agenda with a specific concentration on Burden-Sharing.

World Leaders a - T w i t t e r | 8

Background

Twitter

Twitter is an online microblogging site, created in the United States, that has rapidly globalized since its launch in 2006 (Aharony 2012). Twitter allows registered users to post short messages, called tweets, which may include links, images, or videos in addition to text. These messages allow users to share brief snapshots of their thoughts and activities with their network. Although tweets were originally restricted to 140 characters, the service has since doubled the limit for languages other than Chinese, Japanese, and Korean (Tsukayama 2017).2

The character limit increase is indicative of Twitter’s growing popularity, but it also establishes the company’s resolve to remain a microblogging service.

As a registered user, individuals can interact with one another by liking, commenting, and retweeting (sharing) each other’s tweets. Twitter users can also address and reply to one another through @mentions, or privately message one another through the company’s Short

Message Service (SMS). Another popular function, the hashtag, allows users to index tweets according to a common topic, producing a phenomenon known as hashtag activism in which individuals can show support for a movement by posting to certain hashtags (Yang 2016).

Hashtag activism has since raised awareness for salient topics, such as racism

(#BlackLivesMatter), sexual harassment (#MeToo), and international crimes against humanity

(#Kony2012).

2 These languages can incorporate more ideas in fewer characters because certain characters may denote entire words (Tsukayama 2017). World Leaders a - T w i t t e r | 9

In addition to Twitter’s contribution to social activism, Twitter has also become the most popular social networking system among government leaders (Lüfkens 2018; Parmelee and

Bichard 2012). According to Burson Cohn and Wolfe’s annual Twiplomacy study, which evaluates Twitter usage by world leaders, foreign ministers, and international organizations, a total of 951 personal and institutional Twitter accounts were identified as belonging to world leaders and foreign ministers (Lüfkens 2018).3 These Twitter accounts span across 187 different nations (Lüfkens 2018).

With respect to political Twitter accounts, only a few are personally run by the represented individual themselves. U.S. President Donald Trump is widely recognized for publishing his own tweets on his personal account, @realDonaldTrump, while his other institutional account, @POTUS, merely retweets from @realDonaldTrump and @WhiteHouse.

French President Emmanuel Macron, however, does not necessarily author all his own tweets on his personal account, @EmmanuelMacron (Lüfkens 2018). Likewise, Canadian Prime

Minister Justin Trudeau acknowledges in the profile of his personal Twitter account,

@JustinTrudeau, that the account is run by both himself and his staff. Although President

Macron and Prime Minister Trudeau do not author all their own tweets, it is imperative that their names and images are associated with these accounts, and that these accounts are denoted as official by Twitter.4 Due to these characteristics, the leaders are directly associated

3 Burson Cohn and Wolfe is a global, integrated communications firm with digital communications expertise that funds annual studies regarding heads of state and government, foreign ministers, and international organizations, and their activities on social networking platforms. These studies include “World Leaders on Instagram 2018,” “World Leaders on Facebook 2018,” and “Twiplomacy Study 2018.” 4 A blue checkmark beside an account name indicates the account officially represents the named individual. This is an important Twitter feature since many fake accounts exist for world leaders to troll, or offend and trick, users. World Leaders a - T w i t t e r | 10 with the tweets from their accounts and are nevertheless held responsible for whatever is published, despite any role a staff member may play. This accountability is comparable to politicians being held responsible for the speeches they deliver but do not write. Therefore, the authorship of tweets poses no concern for this study’s analysis since the tweets made from official accounts are still representative of the official. Additionally, exact authorship is also negligible because any discrepancies between a leader’s authentic tone and that of a staffer would remain insignificant due to the sheer volume of tweets and temporal dispersion in the study’s analysis.

In short, Twitter’s rapid growth and prominence in social and political spheres demonstrate the salience of this relatively new communication platform. As the primary platform for digital diplomacy, it is important to understand both the political content and tone associated with this format and its agenda-building effects for official face-to-face summits.

Therefore, this paper takes on both aspects through the in-depth analysis of communications from President Trump, President Macron, and Prime Minister Trudeau regarding the 2018

NATO Summit.

The North Atlantic Treaty Organization

The North Atlantic Treaty Organization, NATO, is a political and military alliance established April 9, 1949 as a response to the threat posed by the Soviet Union. Today, the alliance has expanded beyond the twelve original member nations from Western Europe and

North America to encompass twenty-nine member nations in all, including Eastern Europe and

Turkey. After the collapse of the Soviet Union, the alliance endured to protect democracy around the globe and cooperate on issues to prevent violent conflict. Although member nations World Leaders a - T w i t t e r | 11 are committed to peaceful resolutions, the alliance may still undertake collective crisis- management operations. Article 5 of the North Atlantic Treaty holds members to a principle of collective defense, meaning that member nations agree to consider any attack against one member as an attack on all members. Article 5, however, has only been invoked once – after the terrorist attacks on U.S. soil on September 11, 2001.

NATO’s most recent meeting took place during the 2018 NATO Summit from July 11 to

July 12, 2018 in Brussels, Belgium. While remaining committed to its three core tasks – collective defense, crisis management, and cooperative security – NATO announced six additional goals for the 2018 Summit. These goals included strengthening deterrence and defense, projecting stability and fighting terrorism, stronger cooperation with the European

Union, modernizing NATO, fairer burden-sharing, and supporting an alliance of shared values and transatlantic unity (“NATO Summit Guide” 2018).

The Brussels Summit Declaration (2018) publicized the outcomes of the Summit, denoting approximately eighty agreements and intentions. Among these agreements emphasized in NATO Secretary General ’s press conference on July 11, was a readiness initiative called the “Four Thirties.” This initiative refers to NATO’s attainment of 30 combat vessels, 30 air squadrons, and 30 mechanized battalions ready to use within 30 days by the year 2020. The Summit also led to an agreement to increase NATO Command Structure personnel and contributions to NATO deployments. Members agreed to support allies with energy security, cyber defense, and counter-propaganda strategies, while also launching a new training mission in Iraq and increasing support for Tunisia and Jordan to continue the fight against terrorism. Additionally, members agreed that the alliance did not possess fair burden- World Leaders a - T w i t t e r | 12 sharing, a chief concern of President Trump, so allies agreed to increase defense spending. For

2018, eight members agreed to spend at least 2% of their GDP on defense, with most members planning to accomplish the same by 2024 (“Brussels Summit Declaration” 2018).

The NATO Summit concluded with a plethora of agreements and some undisclosed disagreements as well. This paper seeks to understand how the agenda for these negotiations unfolded by evaluating the different verbal and Twitter communications regarding the Summit.

Before analyzing these exchanges, however, I outline the existing explanations for the merits and shortcomings of in-person communication and computer mediated communication in both social and political contexts, along with a concise description of agenda-building studies in relation to Twitter.

Literature Review

Understanding Politics Through Verbal Communication

Prior to the creation and popularization of Twitter, communications between heads of state and government were traditionally restricted to face-to-face dialogue and other verbal platforms, including televised speeches and press conferences. In political communications, language and social interaction have remained salient because they produce senses of accountability (Robinson 2016). With verbal statements, a speaker is morally responsible for identifying, understanding, and adhering to “relevance rules,” or the behavioral norms of reasoning and conduct in a specific context (Robinson 2016). A failure to uphold relevance rules is punishable in the realm of reputation (Robinson 2016; Heritage 1984). Therefore, verbal World Leaders a - T w i t t e r | 13 communications influence political discussion by promoting accountability through contextual norms of behavior and threats of reputational consequence.

In addition to accountability, world leaders also use verbal communications, especially face-to-face exchanges, to determine credibility and sincerity based on personal impressions

(Groth 1964; Hall and Yarhi-Milo 2012). Rather than obtaining information through concrete evidence alone, personal contact plays an important role in establishing understanding and brokering political deals (Groth 1964; Hall and Yarhi-Milo 2012). Consequently, incorrect impressions of another leader’s intentions may result in costly repercussions.

Verbal communications also incorporate a complex set of mechanisms that signal both meaning and emotions to interlocutors (Dohen et al. 2010; Kendon 1988; Short et al. 1976). A large body of communication literature asserts that oral interactions are multimodal because they require speakers to hear and see the production of sound in conjunction with physical gestures (Dohen et al. 2010; Kendon 1988; Short et al. 1976). According to Marion Dohen, Jean-

Luc Schwarts, and Gérard Bailly (2010), prosody (intonation, rhythm, and phrasing) influences the understanding of spoken words. Sarcasm, for example, is often conveyed through a vocal flatness that negates the actual words that have been said. In this line, research also finds that physical gestures used while speaking, termed gesticulation, and communicative settings can similarly influence discourse cohesion by providing additional information and context for spoken words (Dohen et al. 2010; Kendon 1988). The literature concerned with the multimodality of verbal communication further demonstrates how verbal communications provide political leaders with more information to better facilitate negotiations. World Leaders a - T w i t t e r | 14

Thus, if verbal communications provide a greater wealth of information, it is likely that politicians will respond with greater attention to their context, suggesting their content and tone will be more appropriate. Additionally, agenda-building may benefit from supplementary information and fail to extend to social media platforms. This study evaluates content, tone, and Twitter’s agenda-building functions during the 2018 NATO Summit, arguing that Twitter’s content and tone will be largely negative in comparison to verbal communications and that

Twitter adds to the information provided by in-person communications to advance agenda- building.

Politics in 280 Characters or Less

Since Twitter’s launch in 2006 and its increasing incorporation into the political sphere, studies have examined the service’s communicative capacity (McFedries 2007; Honeycutt and

Herring 2009; Schandorf 2013). Courtenay Honeycutt and Susan C. Herring (2009) look at

Twitter as a platform for conversation and collaboration. They investigate how Twitter supports communicative exchanges between users and how people use Twitter. Their findings indicate that Twitter is not particularly conducive to conversational use, however, both long and short exchanges occur with relative coherence. Honeycutt and Herring suggest that Twitter conversations are reliant on the @ symbol to address users and track conversations. Still, they identify Twitter as a “noisy environment” in which a high volume of tweets disrupts discourse fluidity and impedes comprehension.

Conversely, Michael Schandorf (2013) argues that not only can Twitter facilitate conversation, but the same nonverbal mechanisms available in verbal communication are present in Twitter as well. He poses that the nonverbal mechanisms of face-to-face World Leaders a - T w i t t e r | 15 communication and the deliberate typography of computer mediated communication stem from the same cognitive origins and serve equivalent communicative purposes. Schandorf considers textual interjections (e.g., ‘grrrrr’) a potential replacement for vocal gestures that impart meaning and emotion. He also denotes how retweeting, following, and linking information can serve as digital gestures. Meanwhile, breaking up text can mimic verbal pauses to create rhythm, and imagistic gestures are present through emoticons and descriptions of gestures set off by asterisks (e.g. *high five*). Thus, Schandorf asserts that computer mediated communication includes various methods that mimic the nonverbal mechanisms of face-to-face communication, making Twitter communication possible and equally as coherent as verbal communication.

Whether or not Twitter is conducive to conversation is crucial for understanding the medium’s agenda-building functions. An inability for leaders to directly communicate can inhibit information-sharing functions and circumvent the relevant audience necessary to influence an upcoming agenda.

Jennifer Golbeck, Justin M. Grimes, and Anthony Rogers (2010) go beyond evaluating

Twitter’s communicative capacity to examine the primary functions of political communication on Twitter through an analysis of U.S. Congress members. Their results indicate that Congress members mainly use Twitter to disperse information to the public, especially through links to outside sources about themselves. Rather than establish transparency, the links provided by most Congress members serve as tools of self-promotion. Golbeck et al. also find that Twitter facilitates communication between Congress members and citizens, but this activity is less common. Golbeck et al. belong to an array of scholars studying Twitter’s political functionality World Leaders a - T w i t t e r | 16 in the United States (Parmelee and Bichard 2012; Shapiro and Hemphill 2017; Hemsley et al.

2018); therefore, their work recommends an international extension of their research to understand if other international politicians utilize Twitter in the same ways as U.S. Congress members. This paper answers the authors’ call by analyzing how politicians from other nations utilize Twitter and how these uses translate to diplomacy and international agenda-building.

Existing international Twitter studies primarily focus on Twitter’s domestic politics within a certain nation (Ahmed et al. 2017; Grant et al. 2010; Frame and Brachotte 2015). This study goes beyond a single international case to evaluate three different international politicians and considers bilateral and multilateral interactions instead of domestic exchanges.

Constance Duncombe (2017) also looks beyond Twitter’s domestic political implications to investigate Twitter’s use in international diplomacy. Through a study of U.S. and Iranian relations, Duncombe argues that Twitter plays an important role in negotiation strategy. She finds that states use Twitter to represent their identities and will produce tweets to reshape how they are recognized when they feel misrepresented. Therein, how a state represents itself through social media can signal political possibilities for change and rule out others. Similar to

Duncombe, I also consider Twitter’s role in diplomacy, but I compare its role to verbal diplomacy to better understand communicative discrepancies and to evaluate each platform’s conduciveness to agenda-building.

Opinion Change Through Different Communication Platforms

Numerous studies consider the difference in communicative persuasiveness based on communication platforms (Bower and Pulford 2013; Di Blasio and Milani 2008). Stephanie

Bower and Briony Pulford (2013) examine how individuals’ implementation of face-to-face World Leaders a - T w i t t e r | 17 advice differs from their implementation of computer mediated advice. By having participants play a prisoner’s dilemma game after receiving strategic advice through different communication platforms, Bower and Pulford find that individuals’ implementation of advice was high regardless of the advice format and advisor expertise. However, the study reveals that face-to-face (over video) advisors were perceived as friendlier and more pleasant than those who gave advice by text alone. Bower and Pulford acknowledge that verbal advice in the video condition provided participants with more information about their advisors, including gender, age, and ethnicity, but this additional information did not contribute to the persuasiveness of verbal advisors. Some research would attribute this phenomenon to confirmation bias in which individuals assume the identities of anonymous individuals online match their own (Schandorf

2013).

Paola Di Blasio and Luca Milani (2008) also investigate opinion change through verbal communication as compared to computer mediated communication. They argue that face-to- face communication is more persuasive because its physical components allow people to defend their positions energetically, resulting in opinion change. However, contrary to

Schandorf (2013), they argue that computer mediated communication does not allow for physical and non-verbal mechanisms, so individuals were more likely to uphold their original opinions.

In this study, I engage the subjects of advice implementation and opinion change between platforms by evaluating how online and verbal communications help set or change the agenda for the 2018 NATO Summit. I argue that Twitter supplements verbal interactions and, thusly, contributes to agenda-building. World Leaders a - T w i t t e r | 18

Behavioral Variance On- and Offline

Computer mediated communication research often investigates how communicative exchanges between digital users compare to verbal interactions. Shirley Ho and Douglas M.

McLeod (2008), for example, study the difference between an individual’s willingness to express opinions face-to-face and in an online chatroom. Their study finds that individuals are more likely to share opinions online than in person. They suggest that the anonymity and reduced observable social cues allotted by computer mediated communication make it easier for individuals to express their authentic opinions, whereas the physical presence of others during face-to-face communication contributes to an individual’s fear of criticism, which suppresses their opinions. This paper, however, aims to remove the veil of anonymity as

Twitter heads of state and government make their identities known with official, verified accounts.

Adam Joinson (1998) further investigates the differences between in person and computer mediated communication in his seminal contribution to Psychology and the Internet.

Joinson centers his scholarship on computer mediated communication behaviors and motivations, finding that individuals behave differently online than offline. This difference in behavior and lack of social constraint has been termed disinhibition. Joinson finds that individuals are more likely to share about themselves over a computer mediated communication system than in person or through video. Therefore, the study reveals that visual anonymity leads to higher levels of self-disclosure. Joinson also finds that individuals are more likely to flame online than in person. Flaming refers to impolite statements, profanity, exclamations, superlatives, or overt expressions of feelings towards someone else. Joinson World Leaders a - T w i t t e r | 19 suggests the tendency to flame is higher online than in person because there is a reduction in accountability and public self-awareness. In this paper, I consider how different online and offline behaviors contribute to the content and tone of political communications.

Agenda-Setting and Agenda-Building Through Twitter

Agenda-setting literature examines the connection between information sources, particularly the news media, and the creation of the public’s reality (McCombs and Shaw 1972).

As a logical extension of agenda-setting, agenda-building literature investigates how politicians set the agenda by influencing the media (Lee and Xu 2018; Kiousis et al. 2006; Kiousis and

Strömbäck 2010). With social media’s penetration into the political sphere, scholars have started evaluating how social media help politicians influence the political agenda (Yang et al.

2016; Lee and Xu 2018). Numerous studies demonstrate that Twitter has become a highly quoted source for opinion comments and information subsidies among journalists and has thereby accomplished agenda-building effects (Bane 2019; Metag and Rauchfleisch 2017;

Parmelee 2014). Through a quantitative content analysis, Kaitlin Bane (2019) finds that print news publications use Twitter for quotations, but web-only news publications quote from

Twitter more often. Moreover, agenda-building studies also evidence that popular tweets from popular accounts are more likely than other tweets to influence the coverage of reporters from traditional mass media (Parmelee and Bichard 2012; Parmelee 2014; Metag and Rauchfleisch

2017). Jayeon Lee and Weiai Xu (2018), for example, find that during the 2016 U.S. presidential election Donald Trump and his opposing candidate, Hillary Clinton, both drew attention to campaign issues by tweeting attacks about one another. World Leaders a - T w i t t e r | 20

Nevertheless, some scholars also suggest that journalists influence the Twitter agenda

(Metag and Rauchfleisch 2017; Conway-Silva et al. 2018). After assessing Swiss journalists, Julia

Metag and Adrian Rauchfleisch (2017) find that the journalists themselves believe they can influence politicians through Twitter and conclude that such a relationship is conceivable.

Bethany Conway-Silva et al. (2018) also find that a reciprocal agenda-setting relationship exists between newspapers and Twitter; however, the influence of the newspapers on Twitter was greater.

While these studies all address how social media influence either the public or other media sources, none consider how social media influence actual agenda topics discussed in bilateral or multilateral contexts. This study seeks to fill this gap in the literature by evaluating

Twitter’s capacity to influence the topics discussed at the 2018 NATO Summit.

Theory & Guiding Hypotheses

Through the investigation of two alternative hypotheses, this study intends to explain the differences between Twitter and verbal communications among world leaders, and their effects on content, tone, and agenda-building. The first guiding hypothesis addresses the difference in content and tone between tweets and verbal communications, stipulating that tweets are more likely to comment on grievances and express a negative tone than are verbal communications. This hypothesis draws from behavioral disinhibition theory (Joinson 1998) which argues that computer mediated communication distances individuals from the consequences of their online behaviors, increasing the likelihood of more extreme and inappropriate behaviors online than offline. Although world leaders do not enjoy the same World Leaders a - T w i t t e r | 21 anonymity associated with behavioral disinhibition due to their verified Twitter accounts, they still enjoy a similar psychological perception of physical safety and disconnection to potential consequences. Without the social accountability of spoken language, I propose that the psychological disassociation with consequences through computer mediated communication platforms, like Twitter, permits heads of state and government to speak more boldly about their demands of other leaders, or their grievances, and supports a negative tone that is less common in verbal communication.

The second hypothesis addresses Twitter’s agenda-building function. This hypothesis postures that heads of state and government are more likely to influence the agenda for in- person summits through Twitter use than verbal communications alone because they can control the volume of topics they consider important through their number of tweets. The second hypothesis thereby finds grounds in Maxwell McCombs and Donald Shaw’s (1972) agenda-setting theory of mass media. McCombs and Shaw (1972) argue that the media establish political reality, not by instructing individuals about what to think, but by emphasizing the importance of an issue through the amount of information provided. Agenda-building logically extrapolates from this theory by suggesting political actors influence media coverage and, thereby, indirectly set the public agenda (Lee and Xu 2018; Kiousis et al. 2006; Kiousis and

Strömbäck 2010). In accordance with agenda-building, the second hypothesis suggests that politicians will use Twitter as an agenda-building tool to push their grievances and interests to the forefront of negotiations by posting in higher volumes about topics, goals, or demands related to the upcoming summit. Therefore, I argue that that the volume with which leaders post about a certain topic signals the issue’s importance and provides other actors with World Leaders a - T w i t t e r | 22 information about which issues need to be prioritized during negotiations. Nevertheless, I expect that agenda-building is still viable through verbal communications, but in a lower volume and with less frequency. The study’s two hypotheses are outlined as follows:

Hypothesis 1 (a-b): If heads of state and government use Twitter for diplomacy, the posts are more likely to contain a greater volume of (a.) grievances (b.) negatively toned tweets due to online behavioral disinhibition and less social accountability than verbal communications.

Hypothesis 2: If heads of state and government use Twitter for diplomacy, they are more likely to help build the agenda for in-person summits by producing a greater volume of tweets related to the topics they find most important than if communication was restricted to verbal communications alone.

Data & Methodology

Research Design

To investigate these hypotheses, this study provides an analysis of tweets and verbal communications between June 1, 2018 and August 31, 2018. The study offers an international scope through the examination of three world leaders, including U.S. President Donald Trump,

French President Emmanuel Macron, and Canadian Prime Minster Justin Trudeau. The study uses three phases of analysis to thoroughly inspect the different communications:

(1) descriptive statistics; (2) content and tone analysis; and (3) comparative analysis

The first phase describes all the tweets and NATO-related verbal communications during the examination period. For the second and third phases, only NATO-related tweets and verbal communications are examined.

World Leaders a - T w i t t e r | 23

Data Collection

Twitter serves as the computer mediated communication platform for this study because it is the most popular social networking site among heads of state and government

(Lüfkens 2018; Parmelee and Bichard 2012). Twitter has also found relevance through U.S.

President Trump’s Twitter presence, based on his massive following and his high number of tweets. In this line, the study analyzes tweets and verbal communications from President

Trump, President Macron, and Prime Minister Trudeau. U.S. President Trump has been selected for this study because of his popularity on Twitter and the United States’ large contributions to

NATO. Additionally, the United States is a powerful nation that belongs to NATO’s founding member nations. President Macron and Prime Minister Trudeau have both been selected because of their Twitter popularity and their positions as leaders of two of the twelve original

NATO member nations.5 Moreover, President Macron and Prime Minister Trudeau both tweet and speak in French, for which the study has provided a native French translator.6 To cater to his domestic public, Prime Minster Trudeau also tweets and speaks in English. The comparative analysis of these three leaders provides an opportunity to understand how Trump-era Twitter

5 Compared with President Trump’s 58.4 million Twitter followers, President Macron and Prime Minister Trudeau’s popularity is dwarfed; however, President Macron and Prime Minister Trudeau enjoy 3.7 million and 4.46 million followers, respectively, which is still a large number relative to the 18,135 follower average from 2018 for world leaders (Lüfkens 2018). 6 Philomène Gonay is an undergraduate student from Chambéry, France who has studied English for over ten years and is currently studying philosophy at the University of Toronto, Canada. Gonay has a wealth of experience with French to English translations from her studies. To account for cultural differences and language barriers with President Macron and Prime Minister Trudeau, Gonay has volunteered to provide translations and note any cultural nuances, such as sarcasm, that may be imperceptible to outside readers. Gonay did not receive payment for her work. World Leaders a - T w i t t e r | 24 interactions truly impact content, tone, and agenda-building as compared to verbal interactions.7

Furthermore, the 2018 NATO Summit contextualizes this study because all NATO member nations have official representation on Twitter. The most recent NATO Summit took place during July 11-12, 2018, providing the study with contemporary and relevant information for analysis. The alliance’s high press coverage also provides more data, while its 70 years of stability makes it easier for future research to retest this study’s results. Twitter and verbal data have been collected from June 1, 2018 to August 31, 2018. This examination period captures one month of communications prior to the Summit, the month of the Summit, and one month afterward, including a total of 2,028 tweets and 15 verbal communications. Despite the temporal scope, NATO-related communications are not evenly distributed across this examination period and cluster around the Summit in July.

Twitter data for U.S. President Trump were downloaded from the Trump Twitter

Archive, a public archive updated hourly to include all tweets from @realDonaldTrump, including those later deleted.8 The archive provides information about the tweet itself, any tweet attachments (links, videos, images, articles), the date of publication, and the number of favorites, retweets, and comments that the tweet received. Brendan Brown, a programmer

7 At this point, a reader may wonder why Chancellor Merkel was not included as a subject in the study. Despite the Chancellor’s position as leader of a powerful NATO member nation whose country was specifically addressed by tweets from President Trump, Chancellor Merkel refuses to participate in Twitter. Her government created an official account for her use, but Chancellor Merkel purposefully does not engage in the platform. Nevertheless, many other politicians represent Germany on Twitter. Future research should conduct an experimental study examining the difference between heads of state and government active on Twitter and those who do not use the platform. 8 Retrieved from: http://www.trumptwitterarchive.com/. World Leaders a - T w i t t e r | 25 from Boston, Massachusetts, created the archive as an independent project and continues managing the archive and its code. For President Macron and Prime Minister Trudeau, however, similar archives are not available. Therefore, Twitter data for @EmmanuelMacron and @JustinTrudeau were collected manually. This process included going back to June 1, 2018 on each leader’s timeline and copying and pasting all tweets and retweets into Excel spreadsheets through August 31, 2018. This included the manual input of the text itself, any attachments, the date, and the number of favorites, retweets, and comments. Moreover, the tweets collected only encompass public tweets, so private direct messages are not included in the sample.

Verbal communications data were derived from NATO searches on the official government websites of each leader. Media coverage detailing specific NATO-related comments or speeches was also used to track down any verbal communications not included on the government websites. C-Span, YouTube, and Twitter links provided the necessary visual footage of all transcribed and analyzed verbal communications. For President Trump, most transcriptions were collected directly from the official White House website. Other transcription processes included a transcription software, transcription and translation, and manual transcription. The transcription software, Happy Scribe, automatically transcribes videos regardless of the language.9 Therefore, this service was used to transcribe verbal communications given in English and French. Once the verbal communications were transcribed through the software, the French transcripts were translated into English by a native French

9 Happy Scribe is a paid automatic transcription software made in Europe. Retrieved from: https://www.happyscribe.co/. World Leaders a - T w i t t e r | 26 speaker.10 In some cases, the audio was unclear and could not be transcribed through the software. In these instances, videos were transcribed manually by the author.

Data Analysis

The study begins with descriptive statistics regarding all President Trump, President

Macron, and Prime Minister Trudeau’s tweets from June 1, 2018 to August 31, 2018. The descriptive statistics provide a summary of the collected data to quantify results. This information includes basic data, such as the average number of tweets by month and the average amount of interactions with tweets from each account. The study also integrates similar descriptive statistics for verbal communications, describing the number of press conferences, joint press conferences, bilateral meetings, rallies, brief media statements, or

NATO panels given by all three leaders within the examination period regarding NATO.

After describing the basic composition of the data, the second phase of analysis evaluates the content and tone of all collected tweets. Through an inductive, clustering approach, every tenth tweet within the three chronological datasets was categorized, and new categories were developed whenever a tweet did not match a preexisting category. After this random sample, the categories devised nine macro-topics across all three leaders. The remaining tweets were then coded according to these topics. These coding categories are defined as follows:

• Domestic Issues: Information and/or opinions on domestic (policy) topics or issues; stating a position on a domestic issue; • Domestic Events: Information on upcoming, occurring, or past domestic events to raise awareness about the event; does not include sports;

10 See Footnote 6, Philomène Gonay. World Leaders a - T w i t t e r | 27

• International Issues: Information and/or opinions on international (policy) topics or issues; stating a position on an international issue; • International Events: Information on upcoming, occurring, or past international events to raise awareness about the event; does not include sports; • Sports: Information, support, or congratulations regarding a national or international sporting event or team; • Gratitude/Congratulations: Sincerely thanking or congratulating an actor or actors; does not include sports; • Condolences: Sincerely expressing sympathy for a domestic or international tragedy or death; • Self-Promotion/Job Performance: Promoting accomplishments for the purpose of improving approval ratings; differs from ‘Domestic Issues’ by including little to no facts; and • Opponent: Attacking, blaming, or countering a critic or group of critics; insulting or uncovering unfavorable opponent actions

Once all tweets are analyzed for content, a tone analysis of all tweets categorizes them as either positive, negative, or neutral. These tonal categories were originally coded by numerous emotions and later sorted as follows:

• Positive: Language signaling optimism, kindness, good relations, or cheer; language evoking good humor or comfortability; • Neutral: Language that cannot be convincingly sorted into either positive or negative tone categories; language that avoids making direct statements or revealing attitudes toward a certain topic; language without attitude or emotion; and • Negative: Language signaling pessimism, unkindness, bad relations, or animosity; language that evokes anger or defensiveness

A more detailed content and tone analysis of only the NATO-related tweets follows. This phase uses the same inductive, clustering approach to develop content categories specific to

NATO topics. After coding every other NATO-related tweet for each leader, ten NATO content categories emerged. The remaining NATO tweets were then coded accordingly. Thus, the NATO coding categories are:

• Importance/Success: Remarks on the importance of NATO, the success of the NATO Summit or any smaller NATO meetings, and the successful evolution of NATO; World Leaders a - T w i t t e r | 28

• Call for Change: Remarks asking for specific changes to be made to NATO or intimating that changes are necessary; • Unity: Remarks acknowledging NATO’s solidarity and ability or necessity to work together; • Commitment: Remarks about each leader’s respective country’s commitments to NATO outside of burden-sharing; • Burden-Sharing: Remarks related to NATO member nations’ defense expenditures or equipment and military personnel expenditures; • Positive Relations: Remarks relating to good relations with another country or leader; • Security: Remarks on the defensive goals of NATO or the protection of NATO member nations; • Russia: Remarks about the Russian pipeline in Germany or Russia as a threat to NATO; • Trade: Remarks relating to trade between NATO allies; and • Event: Information on NATO as an upcoming event or updates about NATO events that are occurring or have already taken place

Following the content analysis of the NATO-related tweets, a more in-depth tone analysis considers the semantics and syntax of each tweet. The secondary tone analysis evaluates the use of punctuation, capitalization, rhetorical questions, idioms, and direct reference to another individual or state to determine positive, negative, or neutral tone as previously defined. Tone will refer to attitudes towards NATO, rather than the overall tone of the tweet.

The same content and tone analyses are applied to the transcripts and videos of the verbal communications from all three leaders. Since all verbal communications address NATO, the same categories developed for the NATO-related tweets were applied to the verbal communications. Every time the speaker referred to one of the NATO coding categories, the reference was assigned to a single word. Any references made to the categories by other speakers or audience members are not included. To account for the difference in length of all the verbal communications and allow for consistent comparison across leaders, these references, or words, were then totaled for each respective leader and divided by the leader’s total word count for all his verbal communications. This process was completed for each World Leaders a - T w i t t e r | 29 category and for each leader to determine the percentage of total verbal communications dedicated to a specific category. Since references were only associated with a single word, the word counts were taken after filtering out stop words and only counted the words spoken by the leader to increase the size of the percentages. Stop words stem from computer science and refer to common and insignificant words that are normally ignored by search engines. Some stop words, for example, include: a, an, at, the, and which. All stop words were removed through the same free app, “Cleanse Stop Words.”11

A similar process was also applied to each individual leader for their verbal communications before, during, and after the Summit. In this case, references were still attributed to a specific word and the communications were filtered for stop words. Then, the number of references was divided by the total number of words for all of a leader’s verbal communications either before, during, or after the Summit. This provided a percentage for the amount of verbal communications dedicated to a specific category in each time period. To evaluate tone for verbal communications, videos of these communications were coded for gesticulation, body language, verbal phrasing, and verbal volume or emphasis. The tone categories still included positive, negative, or neutral.

The final phase of analysis for this study is a comparative analysis of Twitter and verbal communications, in addition to a comparison of the communication content and the 2018

NATO Summit agenda. This section includes a comparison of NATO-related tweets and verbal communications regarding the types of content and tone. By evaluating the predominant content and tone between communication platforms, the study tests Hypotheses 1a-b. Next,

11 Retrieved from: https://demos.datasciencedojo.com/demo/stopwords/. World Leaders a - T w i t t e r | 30 the analysis compares Twitter and verbal statements to identify whether or not any demands or criticisms made about NATO were addressed during the Summit as a result of tweets. This analysis concentrates on Burden-Sharing, which is the only clear demand made within the examination period, and tests Hypothesis 2 by evaluating Twitter’s agenda-building function.

Descriptive Statistics

Twitter Summary Statistics

After collecting all the tweets published by President Trump (@realDonaldTrump),

President Macron (@EmmanuelMacron), and Prime Minister Trudeau (@JustinTrudeau) between June 1, 2018 and August 31, 2018, the study’s Twitter data captured a total of 2,028 tweets. President Trump was responsible for tweeting 1,049 times during the examination period, composing 51.7% of the total Twitter data. President Macron tweeted the second most often with 498 tweets, or 24.6% of the total, and Prime Minister Trudeau tweeted 481 times, or

23.7% of the total. Prime Minister Trudeau, however, often publishes the same tweet twice, once in French and once in English. Nevertheless, observations for Prime Minister Trudeau are not divided in half because there are still some tweets that are not translated into the other language and some tweets in which the translation occurs within the same post.

Throughout the examination period, the distribution of tweets over time was not consistent among all three Twitter leaders. Instead, all three leaders tweeted the most in June, as depicted in Figure 2. Additionally, President Trump tweeted the least in July, whereas

President Macron and Prime Minister Trudeau tweeted the least in August. President Trump tweeted 35% of his total tweets in June, 30.5% of his total tweets in July, and 34.5% of his total World Leaders a - T w i t t e r | 31

Figure 2: Tweets by Month 400 350 300 250 200 Tweets 150 100 50 0 June July August

Trump Macron Trudeau

tweets in August; President Macron tweeted 51.4% of his total tweets in June, 32.1% in July, and 16.5% in August; and, Prime Minister Trudeau tweeted 45.3% in June, 36.2% in July, and

18.5% in August.

While President Trump demonstrates consistent tweet volume over time, President

Macron and Prime Minister Trudeau both demonstrate a higher volume of tweets in June and

July. After evaluating tweet content, the differing volumes of both the French and Canadian leaders can be explained by a large volume of tweets addressing the G7 Summit that took place in Canada on June 8-9, 2018. President Macron’s exceptionally high volume of tweets in June is not only explained by the G7 Summit, but also by a European Council meeting in late June.

Table 1 provides a basic description of the study’s Twitter data. Table 1 indicates that

President Trump retweets far more often than President Macron, while Prime Minister Trudeau never retweets and only publishes original content. Despite President Trump’s higher rate of retweeting, it is noteworthy that President Trump often retweets his own posts from the past. World Leaders a - T w i t t e r | 32

Table 1: Twitter Interactions

% Posts That % Tweets with Average # of Average # of Average # of Are Retweets an Attachment Comments per Favorites per Times a Tweet Tweet Tweet Is Retweeted President 13.7% 26.3% 18,423 81,874 19,962 Trump President 2.4% 63.5% 243 2,606 743 Macron Prime Minister 0% 88.4% 229 3,347 685 Trudeau

Excluding retweets from @realDonaldTrump, 9.7% of his posts are retweets and most retweets are originally published by Assistant to the President and White House Director of Social Media,

Dan Scavino (@Scavino45 and @DanScavino). President Macron does not retweet his own posts and most of his retweets are originally published by @Élysée, the institutional account for the French presidency.

Table 1 also demonstrates that Prime Minister Trudeau utilizes attachments more often than the other two leaders. Attachments include links, images, videos, articles, or links to other tweets. These attachments generally supplement the text, but they do not need to be associated with the text at all. Attachments offer leaders the opportunity to provide additional information to their followers via a mixed-media approach. Of Prime Minister Trudeau’s total attachments from within the examination period, 46.6% of attachments are images, 22.7% are articles, 16.5% are links to other tweets, 11.8% are videos, and 2.4% are links. Of President

Macron’s attachments 29.4% are images, 0.3% are articles, 23.4% are links to other tweets, 46% are videos, and 0.9% are links. For President Trump’s attachments, 27.2% are images, 16.3% are articles, 15.2% are links to other tweets, 40.6% are videos, and 0.7% are links. Hence, Prime

Minster Trudeau prefers to attach images while President Macron and President Trump both World Leaders a - T w i t t e r | 33 prefer videos. Either way, all three politicians enrich their tweets with visual media more often than additional text, such as with links to other tweets, links to websites, or articles.

With respect to interactions from Twitter users, Table 1 indicates that President Trump receives the most interactions. Unsurprisingly, President Trump enjoys a higher average number of comments, favorites, and retweets per tweet. Trump’s larger number of interactions are easily attributed to his 58.4 million Twitter followers. President Macron enjoys the second highest average number of comments and retweets per tweet, but Prime Minister Trudeau enjoys the second highest average number of favorites per tweet.

Verbal Communication Summary Statistics

Figure 3: NATO Communication Before, During, and After the Summit 6

5

4

3

2

# of Verbal Verbal of Communications # 1

0 Before During (July 11-12) After

Trump Macron Trudeau

*Bilateral meetings are counted as one communication for each actor.

After evaluating official government websites and news media for NATO-related communications from each of the three world leaders between June 1, 2018 and August 31, World Leaders a - T w i t t e r | 34

2018, the study’s verbal communication data captures a total of 15 verbal communications. The verbal commentary ranges from a minimum length of forty seconds to full-length press conferences over half an hour long, but data has been collected with respect to word counts following the removal of stop words to account for these differences. Figure 3 depicts the number of public NATO-related communications given by each leader throughout the examination period. President Donald Trump engaged in the most public NATO-related communications with a total of 4 bilateral meetings, 2 press conferences, 1 joint press conference, 1 rally, and 1 brief media statement. President Macron engaged in 1 bilateral meeting, 1 press conference, and 1 brief media statement. Meanwhile, Prime Minister Trudeau provided 1 press conference, 1 brief media statement, and 1 NATO panel.

President Trump is the only leader who provided public NATO-related communications in June, but he is not the only one to discuss NATO prior to the Summit. Although Prime

Minister Trudeau did not discuss NATO in June, he did discuss NATO earlier in July.

Unsurprisingly, a majority of public NATO-related communications occurred during the Summit in July. August, on the other hand, presented no public NATO-related communications by any of the study’s leaders. In fact, President Trump is the only leader that publicly discussed NATO following the Summit later in July.

Content and Tone Analysis

Content and Tone on Twitter

All 2,028 tweets were read and assigned a category describing content and tone. The content categories included: Domestic Issues, Domestic Events, International Issues, World Leaders a - T w i t t e r | 35

International Events, Sports, Gratitude/Congratulations, Condolences, Self-Promotion/Job

Performance, and Opponent.12 The tone categories included positive, negative, or neutral.

Table 2: Twitter Content President Trump President Macron Prime Minister Trudeau Domestic Issues 38.6% 23.3% 30.2% Domestic Events 8.2% 9.2% 15.2% International Issues 13.3% 44.1% 13.3% International Events 4.4% 10.1% 13.5% Sports 1.4% 4.3% 2.5% Gratitude/ 8.2% 2.8% 18.9% Congratulations Condolences 1.1% 5.4% 6.4% Self-Promotion/ 6.3% 0.8% 0% Job Performance Opponent 18.5% 0% 0%

*All tweets were given a value of 1. Tweets assigned two categories were thusly valued at .5 for the first category and .5 for the second category. No tweets were assigned more than two categories.

Table 2 illustrates the results of the content analysis for the complete Twitter dataset.

Accordingly, the study finds that President Trump and Prime Minister Trudeau both tweeted about Domestic Issues more often than other content. President Macron, on the other hand, tweeted most about International Issues. The informative functions of these two categories support a more generalizable conclusion for Golbeck et al. (2010) who find that U.S. Congress members primarily use Twitter to disperse information to the public. These findings, however, also indicate that different leaders may define their publics differently. In order to remain relevant to their Twitter audiences, heads of state and government are likely to tweet about issues connected to their publics. Therefore, based on the large volume of Domestic Issues tweets, President Trump and Prime Minister Trudeau are likely to tweet with their domestic

12 See p. 26-27 for specific category definitions. World Leaders a - T w i t t e r | 36 populations in mind as their audience. Following the same logic, President Macron is likely to consider a more international audience when tweeting due to his concentration on

International Issues.

These results speak to the inter-connectedness of Europe and, therefore, the more international focus of European leaders. As a member of the European Union, and as a personal proponent of European unity over nationalism, French President Macron demonstrates a stronger consideration of his international public. Although President Macron’s higher volume of International Issues tweets may simply correlate with a greater number of international issues due to France’s membership in the European Union or the popularity of his international platform, President Macron also demonstrates a greater concern for his international audience by tweeting in foreign languages. Throughout the examination period, President Macron primarily tweeted in French, but other tweets were published in English, Japanese, and Greek without any supplementary French text within the same post. Therefore, certain tweets were exclusively intended for the consumption of an international audience. Neither President

Trump nor Prime Minister Trudeau tweeted in any languages other than their national languages, reinforcing their focus on a domestic audience. Hence, President Trump and Prime

Minister Trudeau’s focus on Domestic Issues and lack of foreign language tweets may indicate that North American geographic isolation, or a small number of shared borders, contribute to a stronger focus on a domestic public when tweeting. More research is necessary to substantiate these observations. World Leaders a - T w i t t e r | 37

Table 2 further establishes that

President Trump is the only Twitter leader in the sample to tweet about

Opponents. In fact, Opponents is the second most common topic covered by

President Trump’s tweets between

June and August, composing 18.5% of his total tweets. Conversely, President Macron and Prime Minister Trudeau do not dedicate a single tweet to Opponents. These findings diverge from the anticipated results grounded in behavioral disinhibition theory, which would predict Twitter to serve as an outlet for negativity, especially directed at political opponents.

Multiple arguments exist to help describe these unexpected results. First, the difference in Opponent tweeting may be attributed to President Trump personally authoring most of his tweets without staff as a filter for tweets attacking opponents. President Macron and Prime

Minister Trudeau both have accounts that are in part written or filtered by staff which may eliminate controversial Opponent tweets. However, this argument is unsatisfactory as it attributes too much power to social media staffers and does not align with the normative behaviors that both President Macron and Prime Minister Trudeau have embodied throughout their terms. Second, President Trump may be more susceptible to online behavioral disinhibition than either President Macron or Prime Minister Trudeau, but such an assertion would necessitate a deeper psychoanalysis of the three Twitter leaders that is beyond the scope of this study. A far more likely argument regarding behavioral disinhibition is that online World Leaders a - T w i t t e r | 38 behavioral disinhibition is more reliant on anonymity than originally anticipated. Instead of a psychological disassociation with social accountability, the association between heads of state and government and their official Twitter accounts is strong enough to overcome the lack of social-constraints and avoid online disinhibition. The increasing coverage of tweets from politicians in the media further strengthens the accountability linkage between a world leader and their Twitter account.

In addition to anonymity serving as a crucial determinant of online behavioral disinhibition, I also argue that the discrepancies between President Trump and the other two leaders concerns personality and political strategy. President Trump demonstrates a personality conducive to addressing opponents, while President Macron and Prime Minister Trudeau are more inclined to avoid public attacks on their opponents. Based on observations throughout all three politicians’ terms, it is evident that President Trump exemplifies a more aggressive and contentious personality, making him a particularly controversial figurehead both within and outside of the United States. However, many politicians differ in their personalities, but do not publicly desecrate their opponents. Therefore, I subsequently hold that personality operates in conjunction with political strategy to influence a politician’s content. President Trump’s unbridled personality is evident of a political strategy that uses populist rhetoric to appeal to ordinary citizens who have become fed up with the “political correctness” of Washington.

President Macron, on the other hand, defeated his populist rival, Marine Le Pen, and Prime

Minister Trudeau, as the son of one of Canada’s most favored heads of government, is inherently counter to populism. Moreover, by openly attacking his opponents, President Trump implements a defensive political strategy to influence the public narrative, whereas President World Leaders a - T w i t t e r | 39

Macron and Prime Minister Trudeau avoid attacking their opponents as a political strategy to maintain a friendly image.

In a similar line, 6.8% of President Trump’s total tweets within the examination period are coded as Self-Promotion/Job Performance. President Macron tweets for Self-

Promotion/Job Performance 0.8% of the time, and Prime Minister Trudeau omits Self-

Promotion/Job Performance entirely.

Again, multiple explanations address these alternative implementations of

Twitter. I argue that personality plays an important role in engaging in public self- promotion or general references to job performance. Different characteristics, including confidence or narcissism, contribute to a greater volume of Self-Promotion/Job Performance tweets. I also argue that the difference between using and not using Self-Promotion/Job Performance tweets relates to the domestic norms of political behavior. Golbeck et al. (2010) find that U.S. Congress members primarily use links provided via Twitter for self-promotion. Therefore, a domestic expectation of self- promotion already exists within the United States that may be considered taboo in French or

Canadian political spheres. Furthermore, a leader’s current approval ratings may also contribute to a greater or lesser volume of Self-Promotion/Job Performance. If a leader has a high approval rating, this may correlate positively with a greater volume of Self-Promotion/Job

Performance tweets to publicize a leader’s success. On the other hand, a low approval rating may also correlate with a greater volume of Self-Promotion/Job Performance tweets to World Leaders a - T w i t t e r | 40 increase approval among a Twitter audience; however, more research is necessary to validate this relationship.

In sum, the three leaders all use Twitter to address different macro-topics. While all three leaders use Twitter to disseminate information or opinions about Domestic or

International Issues, President Trump is the only politician to use Twitter to comment on

Opponents. President Trump also uses Twitter for Self-Promotion/Job Performance more than the other two leaders. President Macron and Prime Minister Trudeau both use Twitter to discuss Sports, express Condolences, and publicize Domestic and International Events more than President Trump; however, President Macron’s large volume of Sports tweets are tied to

France’s World Cup victory from June 14 to July 15, 2018.

Table 3: Twitter Tone President Trump President Macron Prime Minister Trudeau Positive 42.4% 48.8% 65.6% Neutral 13.3% 43.8% 33.9% Negative 44.3% 7.4% 0.5%

Along with coding for content, all tweets were also coded for a positive, negative, or neutral tone. Throughout the examination period, President Trump’s tweets were mostly associated with a negative tone. For President Macron and Prime Minister Trudeau, most tweets were associated with a positive tone. Nevertheless, all three leaders produced a high volume of positive tweets. As depicted in Table 3, President Macron and Prime Minister

Trudeau were more neutral than President Trump. For President Trump, 86.7% of tweets were coded as either negative or positive. President Macron had 56.2% of tweets coded as either negative or positive and Prime Minister Trudeau had 66.1% of tweets coded as either negative World Leaders a - T w i t t e r | 41 or positive. I find that President Macron and Prime Minister Trudeau’s neutrality is indicative of a standard diplomatic approach to controversial issues.

President Trump differs significantly from the other two leaders in his larger volume of negative tweets. While 44.3% of President Trump’s tweets are negative, only 7.4% of President

Macron’s tweets are considered negative, and only 0.5% of Prime Minister Trudeau’s tweets are negative. The first explanation for President Trump’s observed affinity for publishing negative tweets is a correlation between tone and the type of content. Based on President

Trump’s tweets within the examination period, 98.1% of tweets coded as Opponent were negative and 97.8% of tweets coded as Gratitude/Congratulations were positive. The Opponent category is highly conducive to a negative tone because it is defined as “attacking, blaming, or countering a critic or group of critics” just as the Gratitude/Congratulations category lends itself to a positive tone. Therefore, a greater volume of positive or negative tweets is at least partially dependent on content. Since both President Macron and Prime Minister Trudeau avoid

Opponent content, their lower volume of negative tweets naturally ensues.

Other explanations follow the same logic presented to address the difference between leaders and their Twitter content. Therefore, I argue that Twitter tone is highly correlated with a culmination of a politician’s personality and political strategy. Personality influences the tone of political rhetoric by affecting what a politician deems appropriate in a given context.

Pessimistic leaders are more likely to post negative tweets, while those with optimistic characteristics are more likely to post positively. Similarly, political strategy affects tone by dictating what tone is necessary to protect or convey a certain image to constituents. Negativity exudes firmness and power, but it can also express dissent about a policy a leader needs World Leaders a - T w i t t e r | 42 support to repeal and replace. Meanwhile, positivity can establish a friendly image and soothe concerned voters. Therefore, tone is dependent on a leader’s personality, what statements they are comfortable making, and a leader’s political strategy, what they intend to accomplish through certain statements.

These findings further strengthen the aforementioned conclusion that anonymity is significant in facilitating online behavioral disinhibition. With a plurality of President Macron and Prime Minister Trudeau’s tweets promoting a positive tone, I find that behavioral disinhibition is an unlikely explanation for the tone of political leaders’ tweets due to the direct association between the individual and their official account and a lack of anonymity to hide behind when making negative statements.

Content and Tone Through NATO Tweets

After all the tweets were coded for content and tone, only the subset of tweets pertaining to NATO were recoded to specifically address Summit communications and provide comparison to the NATO verbal communications. The number of NATO-related tweets totaled

63 tweets in all, or 3.1% of the original tweet dataset. 2.9% of President Trump’s tweets regarded NATO, 1.4% of President Macron’s tweets discussed NATO, and 5.4% of Prime

Minister Trudeau’s tweets also discussed the Summit. The content categories developed for the

NATO tweets include: Importance/Success; Call for Change; Unity; Commitment; Burden-

Sharing; Positive Relations; Security; Russia; Trade; and Event.13 The in-depth tone categories remain the same: positive, negative, and neutral.

13 See p. 27-28 for specific category definitions. World Leaders a - T w i t t e r | 43

Table 4: NATO Tweet Content President Trump President Macron Prime Minister Trudeau Importance/Success 14.1% 18.2% 12.5% Call for Change 9.4% 0% 0% Unity 0% 54.5% 12.5% Commitment 0% 0% 29.1% Burden-Sharing 37.7% 0% 0% Positive Relations 1.2% 0% 16.7% Security 8.2% 0% 12.5% Russia 8.2% 0% 0% Trade 11.8% 0% 0% Event 9.4% 27.3% 16.7%

*Every reference to a NATO category was valued at 1. Each tweet could contain more than one reference. Percentages were taken by dividing the total number of references to one category by the total number of references made to all NATO categories.

Table 4 illustrates the percentage of references made to each content category among all NATO-related tweets for each leader. In accordance with agenda-setting and agenda- building theory, the categories referred to most often indicate the topics most important to each actor (McCombs and Shaw 1972; Kiousis et al. 2006; Kiousis and Strömbäck 2010).

Therefore, a higher volume of references to a specific issue may serve as a strategy to influence the Summit agenda.

The data indicates that President Trump tweeted about Burden-Sharing far more than any other topic. Additionally, 80% of President Trump’s calls for change directly related to

Burden-Sharing. Thus, it is safe to conclude that President Trump considered Burden-Sharing the most predominant issue facing NATO. When the data is divided according to tweets preceding the Summit, tweets during the Summit, and tweets afterwards, the data also shows that President Trump tweeted the most about Burden-Sharing prior to the Summit as well.14

14 For specific data on NATO-related content before, during, and after the Summit for each leader, see Tables 4a-c in the Appendix. World Leaders a - T w i t t e r | 44

Hence, it is likely that President Trump not only regarded Burden-Sharing as the most important topic, but also wanted to ensure this topic was addressed during the Summit by bringing it up multiple times beforehand.

In accordance with the data, agenda-setting and agenda-building theory also suggest that President Macron considered NATO’s unity the most important issue facing the alliance, whereas Prime Minister Trudeau was most concerned with expressing Canada’s independent contributions and commitment to NATO.

Table 5: NATO Tweet Tone President Trump President Macron Prime Minister Trudeau Positive 26.7% 85.7% 76.9% Neutral 23.3% 14.3% 23.1% Negative 50% 0% 0%

After evaluating the NATO-related tweets for semantics and syntax, the NATO-related tweets were re-coded for tone according to an attitude toward NATO, rather than the overall tone of the tweet. This re-coding resulted in a different tone category as compared to the first general analysis for 4.8% of all NATO-related tweets. Despite the NATO-related tweets only encompassing a small subset of the larger Twitter dataset, the tone results produce a similar pattern. Again, President Trump’s tweets are largely negative, while President Macron and

Prime Minister Trudeau remain significantly positive. Table 5 indicates that half of President

Trump’s NATO-related tweets were negative, while none of President Macron’s or Prime

Minister Trudeau’s tweets were negative. As with the tones from the general dataset, I find that certain NATO content relates to a specific tone toward NATO. For example, 100% of Success tweets are positive toward NATO and 86% of Russia tweets are negative. Therefore, the leaders World Leaders a - T w i t t e r | 45 who post more about certain content will also have more of a certain tone. The findings for the

NATO-related tweets therein mimic the results from the general dataset tones, further supporting that a correlation between content and tone and a culmination of personality and political strategy influence tone, while behavioral disinhibition fails to be significant due to a lack of anonymity.

While all three nations express commitment to NATO based on membership, these differences in attitude toward NATO are best explained by each leader’s election platform.

Election platforms are crucial to the statements and actions of democratic leaders because they serve as promises to the constituency and subsequently hold politicians accountable. President

Trump was elected under his “America First” platform, which explains his negativity towards

NATO. President Trump’s tweets argue that the United States’ NATO contributions are disproportionately high and unfair. In fact, President Trump references the unfairness of NATO burden-sharing in 16.7% of his tweets along with references to the unfairness to U.S. taxpayers.

Therefore, President Trump’s negative attitude towards NATO stems from his election promise to place the interests of the United States before any international interests, such as by reducing U.S. payments to NATO and saving taxpayers money.15

15 It is important to note that President Trump’s dissatisfaction with NATO was also expressed by former U.S. president Barack Obama, but President Trump has been bolder and more overt in his negativity than President Obama was during his presidency. World Leaders a - T w i t t e r | 46

President Macron, on the other hand, was elected under what was, in part, a platform of internationalism and European unity. Now and during his election campaign, President

Macron expressed a desire to create a European Union defense fund to protect European

Union member nations from threats. Throughout his presidency, President Macron has also been a strong critic of nationalism. Therefore, President Macron’s positive attitude toward

NATO is likely due to his election platform’s interest in internationalism and unity.

As a prime minister, Justin Trudeau did not run under his own personal platform but was instead represented by his party’s platform. However, as the leader of his party, the Liberal

Party of Canada, Prime Minister Trudeau was a main candidate for prime minister and a face for the party’s policies. During the year of Prime Minister Trudeau’s indirect election, the Liberal

Party of Canada presented a platform that promised to restore

Canada’s leadership in the world and renew a commitment to international peacekeeping operations (The Liberal Party of

Canada 2015). Hence, Prime

Minister Trudeau’s positive World Leaders a - T w i t t e r | 47 response to NATO is strongly related to his party’s platform encouraging international engagement.

Furthermore, each leader’s tone toward NATO was evaluated before, during, and after the Summit.16 The data indicates that 90% of President Trump’s NATO-related tweets prior to the Summit were negative, while only 14.3% were negative after the Summit. Additionally,

President Trump did not post a single NATO-related tweet with a positive tone prior to the

Summit, but after the Summit 85.7% of his NATO-related tweets were positive. This data demonstrates that the 2018 NATO Summit had a strong effect on President Trump’s attitude toward NATO. Based on Trump’s verbal and Twitter communications, it is evident that the

President’s tonal shift stems from his favorable perception of the Summit’s outcomes.

Content and Tone Through Verbal Communication

The study’s verbal communications, all of which relate to NATO, were coded according to the same coding categorization scheme as the NATO-related tweets. These content and tone categories were applied to both communication platforms to allow for a more direct comparison. Again, the verbal communications were coded for tone according to an attitude toward NATO, rather than the overall tone. The content of the verbal communications and

NATO-related tweets were relatively consistent with one another with the exception of a few categories specific to certain leaders. President Macron, for example, often spoke about soldiers and the “seriousness” owed to the men and women who fight for NATO member

16 For specific data on NATO-related tweet tone before, during, and after the Summit for each leader, see Tables 5a-c in the Appendix. World Leaders a - T w i t t e r | 48 nations; whereas, Prime Minister Trudeau spoke largely about “diversity” and the significance of diversity in the modern world order.

Table 6: Verbal Communication Content President Trump President Macron Prime Minister Trudeau Importance/Success 0.2789% 0.4283% 0.3087% Call for Change 0.1395% 0% 0% Unity 0.0349% 0.4711% 0.3344% Commitment 0.0174% 0.0857% 0.5660% Burden-Sharing 0.5056% 0.5996% 1.0291% Positive Relations 0.2789% 0.4283% 0% Security 0.2615% 0.5139% 0.1029% Russia 0.4184% 0.2998% 0.0772% Trade 0.1046% 0.1713% 0.1801% Event 0% 0% 0%

*Percentages appear small because each reference to a topic has been assigned to a single word and the total number of references is divided by the word count to account for the difference in length of the verbal communications. Stop words were filtered out to increase the percentages. This method allows for comparison across all three leaders.

When discussing NATO between June 1, 2018 and August 31, 2018, Burden-Sharing served as the most commonly discussed topic for all three world leaders. Meanwhile, no leaders spoke to publicize NATO as an event. President Macron and Prime Minister Trudeau never made calls for change to NATO, leaving Trump as the only leader with strong demands.

Prime Minister Trudeau also never verbally addressed Canada’s positive relations with NATO member nations or leaders.

Table 6 is interesting because it depicts that a smaller percentage of President Trump’s total verbal communications addressed Burden-Sharing than President Macron and Prime

Minister. However, these findings do not necessarily indicate that President Macron and Prime

Minster Trudeau regarded Burden-Sharing as more important than President Trump. Afterall, World Leaders a - T w i t t e r | 49

President Trump may have the smallest percentage of Burden-Sharing discussion, but he also provided the most verbal communications and covered the most NATO-related topics.

President Trump’s wider variety of verbal communications and settings thereby depresses the number of references to Burden-Sharing. On the other hand, Prime Minister Trudeau spoke the most about Burden-Sharing percentage-wise, but he also addressed the fewest NATO topics.

The predominance of Burden-Sharing for all three heads of state and government supports several conclusions. First, some may argue that Burden-Sharing was simply the most important NATO topic for all NATO members. This argument, however, falls short because neither President Macron nor Prime Minister Trudeau tweeted about Burden-Sharing. Second, while it may be possible that all three leaders addressed Burden-Sharing the most by chance this conclusion in highly unlikely, especially when reviewing the transcripts of the verbal communications which evidence that leading questions by journalists guided discussions toward Burden-Sharing. These leading questions thusly present two additional arguments for the verbal plurality of Burden-Sharing. The first argument pertaining to the media suggests that

President Trump may have successfully engaged in agenda-building to bring Burden-Sharing to the forefront of the negotiations, while the second poses that the press may have guided the conversation towards Burden-Sharing for their own coverage, regardless of President Trump’s actions. Both media-based arguments are later tested in the comparative analysis.

When verbal communications are divided temporally in relation to the Summit, Burden-

Sharing remains the most popular topic among all world leaders apart from President Trump.

President Trump speaks the most about Positive Relations before the Summit, composing

1.92% of President Trump’s verbal communications preceding the Summit. In contrast to the World Leaders a - T w i t t e r | 50 negative commentary on Twitter, President Trump’s verbal profession of Positive Relations represents a strategic attempt to explain why he will accomplish his goals at the Summit.

During the Summit, President Trump spoke about Russian issues with Germany and Russian threats to NATO the most (0.43%) with Burden-Sharing as a strong secondary topic (0.40%). I find that the prominence of Russia in President Trump’s verbal communications during the

Summit was driven by his upcoming summit with Russian President Vladimir Putin only four days after the NATO Summit, on July 16, 2018, making Russia a hot topic throughout discussions. Finally, after the Summit, President Trump referenced the Importance/Success of

NATO the most, totaling 0.79% of total verbal communications following the Summit, which was intended to signal his diplomatic success at the Summit.17

Figure 4: Verbal Communication Tone 6

5

4

3

2

1 # of Verbal Verbal of Communications #

0 Positive Neutral Negative

Trump Macron Trudeau

17 See Tables 6a-c in the Appendix for more information. World Leaders a - T w i t t e r | 51

Tone for verbal communications were based on language, gesticulation, body language, phrasing, and volume or emphasis. Gesticulation and body language generally remained constant among all leaders. Most verbal communications were given in front of a podium or while seated and hand gestures were often used too frequently to indicate particular emphasis.

Nevertheless, certain upward and downward hand motions were particularly relevant for

President Trump when emphasizing Burden-Sharing. Phrasing and volume also remained relatively consistent among leaders with respect to the predominantly formal settings in which verbal communications were given. These settings included press conferences at the Summit, bilateral meetings, and brief media statements. The only instance in which meaning was significantly affected by verbal phrasing, volume, and emphasis “And I said, ‘You know Angela, I can’t guarantee it, but we’re protecting you and it was during President Trump’s means a lot more to you than protecting us because I don’t know how much protection rally in Montana, during which a we get by protecting you.’” mocking tone was used to – President Trump, Montana Rally, July 5, 2018 express discontent with NATO.

After utilizing audio-visual analysis to code transcripts for tone, President Trump was the only leader whose verbal communications were coded with a negative tone towards NATO.

President Macron and Prime Minister Trudeau both remained positive throughout all verbal communications. Of President Trump’s negative communications, 50% occurred before the

Summit and 50% occurred during the Summit, with no negative comments following the

Summit. However, 55.6% of President Trump’s total verbal communications were positive. 60% of President Trump’s positive communications took place during the Summit and 40% occurred World Leaders a - T w i t t e r | 52 afterward, with no positive comments preceding the Summit. This indicates a verbal shift from a negative attitude towards NATO to a positive attitude due to President Trump’s satisfaction with the Summit’s outcomes.

Comparative Analysis

Comparing Content: NATO Tweets vs. Verbal Communications

Table 7: Most-Referenced Content NATO Tweets Verbal Communications President Trump Burden-Sharing Burden-Sharing President Macron Unity Burden-Sharing Prime Minister Trudeau Commitment Burden-Sharing

When comparing overall content between NATO-related tweets and verbal communications, the findings suggest that President Trump is the only leader who consistently focused on Burden-Sharing across both communication platforms, as shown in Table 7.

President Macron and Prime Minister Trudeau also addressed Burden-Sharing the most through verbal communications, but President Macron primarily tweeted about NATO’s Unity while Prime Minister Trudeau’s tweets addressed Canada’s Commitments.

When comparing content over time, a plurality of President Trump’s tweets before the

Summit addressed Burden-Sharing, while his verbal communications focused on Positive

Relations. During the Summit, he continued to tweet about Burden-Sharing, but spoke mostly about Russia. Following the Summit, President Trump demonstrated consistency by tweeting and speaking about the Importance/Success of NATO more than any other topic.18 Prime

18 See Table 7a in the Appendix. World Leaders a - T w i t t e r | 53

Minister Trudeau, however, primarily tweeted about NATO as an event and spoke about

Burden-Sharing before the Summit. During the Summit, the Prime Minister tweeted a lot about

Commitment and Positive Relations, but again he spoke with a concentration on Burden-

Sharing.19 President Macron only tweeted and spoke about NATO during the Summit, so accounting for temporal distribution with respect to the Summit does not change the findings from Table 7.20

Interestingly, President Macron and Prime Minister Trudeau largely addressed Burden-

Sharing verbally, but they did not tweet about Burden-Sharing. While some may argue that avoiding Burden-Sharing on Twitter is intentional and strategic, I maintain that this is not the case. President Macron only tweeted about three of the ten NATO-related topics, focusing on far more topics verbally than through Twitter. Additionally, his tweets were limited to topics conducive to positivity and optimism including Importance/Success, Unity, and Event.

Therefore, President Macron’s disregard for Burden-Sharing, along with six other topics, is less of a strategic avoidance and more of a reflection on his implementation of Twitter as a tool to positively disseminate information. Moreover, Prime Minister Trudeau does not directly discuss

Burden-Sharing on Twitter, but he does not ignore the subject either. Instead, Prime Minister

Trudeau tweets about Canada’s commitments, which compose a plurality of his NATO-related tweets, to indirectly respond to calls for fairer burden-sharing. During a NATO panel discussion, the Prime Minister made this linkage clear when he described expenditures as a less significant

19 See Table 7c in the Appendix. 20 See Table 7b in the Appendix. World Leaders a - T w i t t e r | 54 measure of NATO contributions than actual positive outputs. Hence, his tweets about

Commitment address Burden-

Sharing but try to move “Announcing money put in, announcing inputs, isn't nearly as important as conversation about demonstrating outputs.” – Prime Minister Trudeau, NATO Engages, July 11, 2018 contributions away from expenditures and toward action.

These results indicate that Twitter content and verbal content often differ. The difference in content, therefore, indicates a difference in how leaders implement Twitter communications as opposed to their verbal communications. For President Trump, Twitter initially served as a platform to air grievances about NATO Burden-Sharing, whereas verbal communications were used to establish a positive image and a perception of Positive Relations.

Once discussions began, both platforms were used to address issues, and after the Summit both platforms were used to express positive accomplishments. Twitter also allowed President

Macron to express his values of unity, based in his election platform, while verbal communications provided him with a platform to detail policy-related information. For the

Canadian Prime Minister, Twitter served as a platform to publicize the Summit beforehand, whereas his stances on NATO Summit issues were addressed verbally. In this case, Twitter functioned as a platform for basic information, while verbal communications provided more detailed, policy-related information. During the Summit, Prime Minister Trudeau’s verbal communications continued to provide detailed policy information, while his tweets regarded his nation’s contributions to Twitter, an indirect reference to Burden-Sharing, and expressed

Positive Relations. World Leaders a - T w i t t e r | 55

Furthermore, President Trump is the only leader to utilize Twitter to express grievances, such as the “unfairness” of NATO burden-sharing, and the only leader to actively call for changes via Twitter to address these grievances. President Trump, however, also utilizes verbal communications to express the same grievances. Dependent on temporal distribution in relation to the Summit, the U.S. President sometimes posted a higher percentage of grievances on Twitter and sometimes expressed a higher percentage of grievances verbally. Therefore,

President Trump posts more tweets containing grievances than other leaders, but he also voices his grievances more often, and does not consistently post more grievances on Twitter. These findings, in conjunction with the omittance of grievances from President Macron and

Prime Minister Trudeau entirely, rejects

Hypothesis 1a and confirms the null hypothesis.

In short, the comparative content analysis demonstrates that leaders implement Twitter differently while verbal communications were predominantly used to express detailed policy issues by all three leaders. The comparative content analysis also establishes that President

Trump may have produced more grievances than other leaders on both communication platforms, but neither he nor any other leaders support Hypothesis 1a, which anticipated a greater volume of grievances on Twitter than in person.

World Leaders a - T w i t t e r | 56

Comparing Tone: NATO Tweets vs. Verbal Communications

When comparing tone, President Trump is the only leader who produces a large amount of negatively toned tweets and verbal communications.21 Overall, President Trump was more negative over Twitter than he was verbally; however, when considering communications in relation to the Summit, sometimes a higher percentage of verbal communications were negative than the percentage of negative NATO-related tweets. Meanwhile, President Macron and Prime Minister Trudeau were both equally as positive toward NATO online as they were in person. Without any NATO-related tweets or verbal communications coded as negative for both President Macron and Prime Minister Trudeau, a temporal analysis does not provide an alternative conclusion. Therefore, all three leaders reject Hypothesis 1b, which anticipated a greater volume of negatively toned tweets than verbal communications, confirming the null hypothesis.

As mentioned earlier, these findings reinforce the salience of personality and political strategy in determining tone, helping to explain the observed differences in tone among the three leaders. The different election platforms of all three leaders also clarify their differing attitudes toward NATO. Additionally, the overwhelmingly positive tones of President Macron and Prime Minister Trudeau demonstrate a linkage of social accountability between leaders and their Twitter accounts, suggesting that anonymity is an important determinant for behavioral disinhibition and that disinhibition cannot explain President Trump’s negative tweets which rejects the logic behind Hypothesis 1b.

21 See Table 5 for NATO tweets and Figure 4 for verbal communications. World Leaders a - T w i t t e r | 57

How Do Communications Relate to the Agenda?

To identify Twitter’s agenda-building function for in-person summits, Twitter and verbal communications must be compared to Summit discussions. Going into the Summit, President

Trump was the only leader in the study who asked for specific changes to be made to NATO. On multiple occasions, President Trump demanded that NATO member nations increase their defense spending to establish fairer burden-sharing both on Twitter and verbally. Additionally,

President Macron only tweeted during the Summit and never tweeted before or after; thus, his references to Unity cannot necessarily be considered an attempt to influence the agenda.

Before the Summit, Prime Minister Trudeau referenced NATO as an event more than Canada’s personal commitments, indicating he was more interested in publicizing the event beforehand than setting an agenda. Therefore, President Trump and his calls for fairer burden-sharing will serve as the case for the analysis of Hypothesis 2, which seeks to understand Twitter’s agenda- building capabilities as compared to the capabilities of verbal communications.

Table 8 compares the demands President Trump made for increased defense spending on Twitter and through verbal communications before the Summit. This comparison indicates that President Trump commented on the issue three times more often on Twitter than verbally.

Additionally, 42.8% of President Trump’s NATO-related tweets before the Summit referenced

Burden-Sharing, while his verbal communications were primarily concerned with addressing

Positive Relations. According to agenda-setting and agenda-building theory, the amount of information provided on a certain topic indicates importance (McCombs and Shaw 1972;

Kiousis et al. 2006; Kiousis and Strömbäck 2010); therefore, President Trump demonstrated a

World Leaders a - T w i t t e r | 58

Table 8: Trump Demands for Fair Burden-Sharing on Twitter and In Person Twitter Verbal

“....And add to that the fact that the U.S. pays close to “I have a very good relationship with , the entire cost of NATO-protecting many of these but on NATO we’re paying 4.2% on a much larger same countries that rip us off on Trade (they pay only [GDP]…So we just can’t have it that way.” June 12, a fraction of the cost-and laugh!). The European Union 2018 had a $151 Billion Surplus-should pay much more for Military!” June 11, 2018

“....Germany pays 1% (slowly) of GDP towards NATO “We have 150 billion dollars in trade deficits with the while we pay 4% of a MUCH larger GDP. Does anybody EU. And on top of that, they kill us with NATO. They kill believe that makes sense? We protect Europe (which us. So we pay 4% of a huge GDP, which got a lot bigger is good) at great financial loss and then get unfairly since I became your president. And Germany, which is clobbered on Trade. Change is coming!” June 11, 2018 the biggest country of the EU, European Union, Germany pays 1%.” July 5, 2018

“The United States is spending far more on NATO than any other Country. This is not fair nor is it acceptable. While these countries have been increasing their contributions since I took office they must do much more. Germany is at 1% the U.S. is at 4% and NATO benefits...... ” July 9, 2018

“...Europe far more than it does the U.S. By some accounts the U.S. is paying for 90% of NATO with many countries nowhere close to their 2% commitment. On top of this the European Union has a Trade Surplus of $151 Million with the U.S. with big Trade Barriers on U.S. goods. NO!” July 9, 2018

“Getting ready to leave for Europe. First meeting - NATO. The U.S. is spending many times more than any other country in order to protect them. Not fair to the U.S. taxpayer. On top of that we lose $151 Billion on Trade with the European Union. Charge us big Tariffs (& Barriers)!” July 10, 2018

“NATO countries must pay MORE the United States must pay LESS. Very Unfair!” July 10, 2018

“Many countries in NATO which we are expected to defend are not only short of their current commitment of 2% (which is low) but are also delinquent for many years in payments that have not been made. Will they reimburse the U.S.?” July 10, 2018

“The European Union makes it impossible for our farmers and workers and companies to do business in Europe (U.S. has a $151 Billion trade deficit) and then they want us to happily defend them through NATO and nicely pay for it. Just doesn’t work!” July 10, 2018 World Leaders a - T w i t t e r | 59 higher level of importance for Burden-Sharing online than in person. President Trump’s call for an emergency meeting during the NATO Summit to discuss increased spending further evidences his perception of Burden-Sharing as exceptionally important. But how can we know if

Twitter really helped build the agenda?

Table 9: Twitter-Verbal Crossover Speaker Twitter-Verbal Crossover

Tomas LeCrass, Croatia Daily Newspaper, “We understand your message [about burden- President Trump’s NATO Press Conference, July sharing], but some people ask themselves, will 12, 2018 you be tweeting differently once you board the Air Force One? Thank you.”

Philip Rucker, Washington Post, President “You tweeted yesterday, “What good is NATO?” Trump’s NATO Press Conference, July 12, 2018 And you’ve talked about NATO as an alliance that benefits Europe, that defends and protects Europe. Do you see any value of NATO to the United States vis-à-vis Russia? Does it help protect the United States from Russia, in your view?”

President Emmanuel Macron, President Macron’s “I read 140-character messages, but I also NATO Press Conference, July 12, 2018 attended the debates. The debates were not of this nature. The debates were debates of truth, but there was no stigmatization or disrespect. They were respectful, active and demanding debates.”

President Emmanuel Macron, President Macron’s “I’m telling you again to conclude this press NATO Press Conference, July 12, 2018 conference, as in all summits, sometimes the corridors, the comments, the tweets, become more important than what is negotiated or said or endorsed, including sometimes by the Heads of State and Government themselves in the final declarations.”

First, President Trump dedicated a majority of his tweets before the Summit to Burden-

Sharing. In order to influence the Summit agenda, the President needed to address the issues he wanted covered during the Summit before the Summit occurred. Second, although private World Leaders a - T w i t t e r | 60 meetings certainly occurred to plan the Summit agenda, Burden-Sharing became a topic pushed by the press. Agenda-building theory suggests that political actors are capable of influencing the media to indirectly set the public, or in this case Summit, agenda; therefore, the prominence of Burden-Sharing questions among the press indicate that they may have been influenced to ask these questions by President Trump’s agenda-building strategy. Of questions asked by the press regarding Burden-Sharing, two reporters directly mentioned tweets, as shown in Table 9, indicating that agenda-building through Twitter was successful. Additionally,

President Macron referenced tweets in his own NATO press conference, also shown in Table 9, thereby evidencing that the actors were also aware of the comments President Trump had made on Twitter about Burden-Sharing and were subsequently responding to them. Therefore,

President Macron’s comments also indicate that President Trump’s tweets successfully helped set the agenda. Ultimately, this data provides support for Hypothesis 2, suggesting that

Twiplomacy contributes to agenda-building for in-person summits based on a greater volume of tweets related to topics leaders deem most important.

Conclusion

Efficacy of Twiplomacy

Twitter is likely to remain a thriving and relevant communication platform in the foreseeable future, and Twiplomacy will therein continue shaping political communications in cyberspace. The next NATO gathering for heads of state and government is expected to occur in

April 2019 to mark the alliance’s 70th anniversary. Details of the meeting have yet to be confirmed, but Twitter may soon provide insight on the event and its agenda. World Leaders a - T w i t t e r | 61

In November 2018, two months after this study’s examination period, President Trump posted several controversial tweets criticizing President Macron surrounding the U.S.

President’s visit to Paris. The tweets attacked President Macron’s call for a European Union army, deeming it a strategy “to protect Europe against the U.S.” and suggesting “Europe should first pay its fair share of NATO” (Trump 2018). In a CNN interview with Fareed Zakaria (2018),

President Macron was asked about the Twitter tirade. The French President stated that he felt the Paris discussions had gone well and that President Trump had confirmed to the press that he was “okay” (Zakaria 2018). President Macron followed these comments saying, “I always prefer having direct discussion or answering questions than making “I always prefer having direct discussion or answering questions than making my my diplomacy through tweets” diplomacy through tweets.” – President Macron, CNN Interview with Fareed Zakaria, (Zakaria 2018). President Macron November 11, 2018 also reiterated his dedication to fairer burden-sharing in NATO, but he noted that increased defense expenditures in Europe should also contribute to European autonomy (Zakaria 2018).

These communications indicate the distinct limitations of Twiplomacy. Although leaders, like U.S. President Trump, may embrace Twitter as a platform for political communication, the rejection of this platform as a legitimate space for diplomatic exchanges by other leaders suggests that it is no replacement for in-person summits. President Macron is not the only prominent European leader to reject Twitter as a diplomatic discussion platform, but German

Chancellor Angela Merkel also refuses to engage with the platform entirely. Moreover, in the

November tweets, President Trump referenced President Macron by name, but did not utilize World Leaders a - T w i t t e r | 62

Twitter’s @mention function to directly engage President Macron. During the examination period, leaders were also seldom addressed through this function. In the rare cases of

@mentions, only one reply was documented by Prime Minister Trudeau within the study’s scope. Hence, the political communicative exchanges between leaders on Twitter are predominantly one-sided, indicating that Twitter may best serve informative purposes and agenda-building functions while failing to facilitate the discourse fluidity necessary for diplomatic dialogue.

Findings

The current study has presented a robust analysis of Twitter and verbal communications by three heads of state and government between June 1, 2018 and August 31, 2018. The 2018

NATO Summit serves to contextualize this study, providing a salient topic of discussion for all leaders. The study implemented three levels of analysis, including descriptive statistics, content and tone analysis, and comparative analysis to investigate whether communication platforms influence the content and tone of politicians’ statements and to identify whether Twitter functions as an agenda-building tool for international summits. The study’s findings demonstrate that Twitter does not support more negative content and tone among world leaders than verbal communications, and rather a leader’s tone remains consistent on both communication platforms. I also find that, in the case of Burden-Sharing discussions during the

2018 NATO Summit, President Trump successfully engaged in agenda-building though Twitter to influence the press and other NATO leaders and position Burden-Sharing as a prominent

Summit topic. World Leaders a - T w i t t e r | 63

Additional findings presented in this study indicate that the North American leaders tweeted about Domestic Issues more than any other subject, while the European leader,

President Macron, tweeted about International Issues more than any other subject. When focusing on the context of NATO, the study finds President Trump tweeted the most about

Burden-Sharing, President Macron tweeted mainly about Unity, and Prime Minister Trudeau tweeted primarily about Canada’s Commitments. Verbally, President Trump, President Macron, and Prime Minister Trudeau all spoke about Burden-Sharing the most. With regard to tone,

President Trump was found to be the only leader to express negativity about NATO both on

Twitter and in person. Overall, President Macron and Prime Minister Trudeau tweeted negatively far less frequently than President Trump.

I argue that the difference in content and tone are most likely tied to a coalescence of a leader’s personality and political strategy. While personality influences how a leader perceives situations and how they are comfortable expressing their opinions about different topics, political strategy affects the goals of certain statements and helps frame information. Together, both factors determine what is said and how it is said.

Contrary to the original logic behind Hypotheses 1a-b, I also find that the veil of social anonymity provided by computer mediated communication is an important determinant of online behavioral disinhibition. Due to the official verification of all three leaders’ identities on

Twitter, they possess a stronger link to their accounts which produces social accountability. The direct association of leaders with their posts on Twitter thereby depresses the effects of behavioral disinhibition and provides a greater level of consistency with grievances and tone across different communication platforms than initially anticipated. World Leaders a - T w i t t e r | 64

Finally, the rejection of Twitter as a legitimate platform for diplomatic negotiations by different world leaders and the lack of direct dialogue between leaders reveals that Twiplomacy will not become a viable replacement for in-person summits in the near future. Still,

Twiplomacy will likely remain significant in disseminating political information and as an agenda-building tool.

Limitations and Future Research

As with all studies, this study faces certain limitations. The study is limited in data by the three-month time period which restricts the sample size for NATO-related tweets and verbal communications, but it is also significantly restricted by the verbal negotiations privately conducted behind closed doors. Greater information of private negotiations would help identify the leaders pushing different agendas and would clarify the different goals of each politician.

However, the confidentiality of these meetings is an important factor for open and honest communications between nations to reach agreements without the additional pressures of public scrutiny and will, thusly, remain a barrier to political communications literature. Another limitation includes a small inter-rater subset. With qualitative studies such as this, inter-rater reliability is necessary to account for bias, and an ideal study would implement multiple coders and capture a larger percentage of the total data.

Nevertheless, these limitations make room for future research. Future studies should address these limitations by increasing the examination period and increasing the analysis to more than three leaders. A larger examination period and more leaders will provide more data to test the findings of this study. By extending analysis to other leaders, future research will also produce a more comprehensive understanding of political Twitter use around the globe. The World Leaders a - T w i t t e r | 65 expansion of international leaders should also accompany an extension of this study’s findings to future NATO Summits and other in-person negotiations, such as the G7.22 By encompassing different organizations and summits, the literature will draw a fuller description of Twitter’s influence on a variety of face-to-face diplomatic discussions. Future studies should also consider comparing Twitter use between all politicians in one country to those of other countries to better understand the difference in online political norms.

Implications

The findings of this paper offer significant implications. As the first study to compare the content and tone of Twitter and verbal communications between world leaders and apply it to the context of an international summit, this study opens a new and contemporary topic for both political science and communications scholars in assessing the communicative differences and political strategies associated with alternative platforms. In addition to broadening the scholarship, this study also provides important evidence that Twitter possesses discernable consequences for meaningful diplomatic discussions. I have demonstrated that behaviors in the

Twittersphere carry into reality through agenda-building strategies and are no longer limited to cyberspace. Twitter’s growing presence in the news media and salient international negotiations further reveal the value of studying this thriving social networking system as it continues to manifest in the offline world.

22 All three of the study’s leaders produced tweets about the 2018 G7 meeting. In these tweets, leaders referred directly to other leaders to request change. Additionally, leaders indirectly responded to these tweets with their own public tweets. Despite the available dialogue and tweets for the 2018 G7 meeting, it was not selected as this paper’s case-study because the Summit’s events were abnormal and highly controversial, evidenced by and President Trump’s premeditated early exit. World Leaders a - T w i t t e r | 66

This study’s significance is also indicative through its demonstration that the political behaviors and Twitter tirades arising in American politics are not yet prevalent in other countries, such as Canada and France. The difference in Twitter behaviors between President

Trump, President Macron, and Prime Minister Trudeau indicates that President Trump’s

Twiplomacy does not represent a new normal for political discourse, and similarly suggests that

President Trump is embracing an alternative, untested form of direct and aggressive diplomacy through an online medium. This form of diplomacy will require greater research in the future both in respect to its efficacy throughout President Trump’s term and the extension of his

Twiplomacy strategies to leaders of other nations over time.

Finally, this study is imperative in that it shows the agenda-building capabilities of

Twitter. Twitter’s ability to shape the agenda of international summits and other diplomatic discussions suggests that social media has effectively extended the strategic political sphere for world leaders, no longer serving as only an informative message dissemination platform.

Twitter’s agenda-building also demonstrates that leaders can effectively communicate the negotiation topics most important to them to other leaders across the globe before summit discussions even begin. This Twitter function also exposes the relationship between the media, politicians, and the public’s reality which is becoming ever more complex through the increasing prominence of social networking services. As studies continue evaluating the predominant and reciprocal relationships between journalists, politicians, and the interplay of social media, it is crucial for scholars and the public to remember that politicians push their political aims both verbally and on Twitter. The public must ultimately remain critical to information, even when it is presented in a user-friendly format with only 280 characters. World Leaders a - T w i t t e r | 67

Appendix

Inter-rater Reliability

For qualitative studies of this nature, an assessment of inter-rater reliability (IRR) is necessary to account for potential observational bias and demonstrate consistency. Therefore, another undergraduate student coded a subset of the study’s tweets and verbal communications.23 The subset totals 10% of total Twitter data, about 50% of NATO Twitter data, and approximately 20% of verbal communications data. In order to receive a chronologically dispersed subset, the secondary coder was instructed to code every tenth tweet in each of the three complete datasets and every other tweet for the in-depth NATO-related tweets. Furthermore, President Trump’s first transcript, President Macron’s second transcript, and Prime Minister Trudeau’s third transcript were selected for IRR evaluation in this randomly devised sequence. Although content categories were originally coded through an inductive, clustering approach, the secondary coder received the macro-topics ahead of time. The secondary coder was also trained to understand the definitions of all categories and the coding process in a practice trial with tweets from January 1-10, 2019 derived from @realDonaldTrump and a NATO-related press conference given by President Trump from the examination period. Oftentimes, studies incorrectly implement statistical methods to describe coding variance, such as by only calculating the percentages of agreement (Hallgren 2012). Studies that only calculate percentages of agreement fail to account for any agreement that occurs by chance. In order to account for chance, this study uses Cohen’s (1960) kappa for nominal variables, which produces a standardized IRR index. Kappa statistics are measured according to the equation:

푃(푎) − 푃(푒) 퐾 = 1 − 푃(푒) wherein P(a) represents agreement between coders, and P(e) represents the probability of agreement due to chance (Hallgren 2012). P(a) is calculated by dividing the number of agreements by the total number of observations. P(e) is calculated by multiplying the row total of observations by the column total of observations in a contingency table for each individual variable, in this case each categorization, and dividing the sum by the total number of observations. Possible kappa statistic values range from -1 to 1, wherein 1 represents complete agreement, 0 represents entirely random agreement, and -1 represents complete disagreement (Hallgren 2012). Krippendorff (1980) suggests a conservative measure of interpretation for Cohen’s kappa statistic in which values below 0.67 are considered unsatisfactory, values

23 Kyoseo (David) Koo is an undergraduate student studying computer science at Colorado School of Mines. Koo did not receive payment for his work. World Leaders a - T w i t t e r | 68 between 0.67 and 0.80 indicate tentative agreement, and values above 0.80 indicate conclusive agreement. Cohen’s kappa statistic was applied to all three datasets: total Twitter data, NATO- related Twitter data, and verbal communications data. Following this process, total Twitter data content agreement was valued at 0.84, total Twitter data tone was valued at 0.86, NATO- related Twitter data content was valued at 0.81, and NATO-related Twitter data tone was valued at 0.89, verbal communications content was valued at 0.73, and verbal communications tone was valued at 0.86. Therefore, the IRR is conclusive, with the exception of verbal communications content which remains tentative.

World Leaders a - T w i t t e r | 69

Table 4a: Trump NATO Tweet Content Over Time Before During After Importance/Success 0% 5.6% 47.6% Call for Change 14.3% 8.3% 4.8% Unity 0% 0% 0% Commitment 0% 0% 0% Burden-Sharing 42.8% 36.1% 33.3% Positive Relations 0% 0% 4.8% Security 14.3% 8.3% 0% Russia 0% 13.9% 9.5% Trade 25% 8.3% 0% Event 3.6% 19.4% 0%

Table 4b: Macron NATO Tweet Content Over Time Before During After Importance/Success 18.2% Call for Change 0% Unity 54.5% Commitment 0% Burden-Sharing 0% Positive Relations 0% Security 0% Russia 0% Trade 0% Event 27.3%

Table 4c: Trudeau NATO Tweet Content Over Time Before During After Importance/Success 10% 14.3% Call for Change 0% 0% Unity 10% 14.3% Commitment 30% 28.6% Burden-Sharing 0% 0% Positive Relations 0% 28.6% Security 10% 14.3% Russia 0% 0% Trade 0% 0% Event 40% 0%

World Leaders a - T w i t t e r | 70

Table 5a: Trump NATO Tweet Tone Over Time Before During After Positive 0% 15.4% 85.7% Neutral 10% 46.1% 0% Negative 90% 38.5% 14.3%

Table 5b: Macron NATO Tweet Tone Over Time Before During After Positive 85.7% Neutral 14.3% Negative 0%

Table 5c: Trudeau NATO Tweet Tone Over Time Before During After Positive 83.3% 71.4% Neutral 16.7% 28.6% Negative 0% 0%

World Leaders a - T w i t t e r | 71

Table 6a: Trump Verbal Communication Content Over Time Before During After Importance/Success 0% 0.0805% 0.7898% Call for Change 0.2740% 0.1878% 0% Unity 0% 0% 0.1215% Commitment 0% 0.0268% 0% Burden-Sharing 1.3699% 0.4025% 0.5468% Positive Relations 1.9178% 0.1073% 0.3034% Security 0.5479% 0.3220% 0.0607% Russia 1.3699% 0.4293% 0.1823% Trade 0.5479% 0.1073% 0% Event 0% 0% 0%

Table 6b: Macron Verbal Communication Content Over Time Before During After Importance/Success 0.4283% Call for Change 0% Unity 0.4711% Commitment 0.0857% Burden-Sharing 0.5996% Positive Relations 0.4283% Security 0.5139% Russia 0.2998% Trade 0.1713% Event 0%

Table 6c: Trudeau Verbal Communication Content Over Time Before During After Importance/Success 0.8547% 0.2918% Call for Change 0% 0% Unity 0% 0.3448% Commitment 0.8547% 0.5570% Burden-Sharing 3.4188% 0.9549% Positive Relations 0% 0% Security 0% 0.1061% Russia 0% 0.0796% Trade 0% 0.1857% Event 0% 0%

World Leaders a - T w i t t e r | 72

Table 7a: Trump Most-Referenced Content Over Time NATO Tweets Verbal Communications Before Burden-Sharing Positive Relations During Burden-Sharing Russia After Importance/Success Importance/Success

Table 7b: Macron Most-Referenced Content Over Time NATO Tweets Verbal Communications Before During Unity Burden-Sharing After

Table 7c: Trudeau Most-Referenced Content Over Time NATO Tweets Verbal Communications Before Event Burden-Sharing During Commitment & Positive Burden-Sharing Relations After

World Leaders a - T w i t t e r | 73

Bibliography

Aharony, Noa. 2012. “Twitter Use by Three Political Leaders: An Exploratory Analysis.” Online Information Review, 36 (4): 587-603.

Ahmadian, Sara, Sara Azarshahi, and Delroy L. Paulhus. 2017. “Explaining Donald Trump via Communication Style: Grandiosity, Informality, and Dynamism.” Personality and Individual Differences, 107: 49-53.

Ahmed, Saifuddin, Jaeho Cho, and Kokil Jaidka. 2017. “Leveling the Playing Field: The Use of Twitter by Politicians During the 2014 Indian General Election Campaign.” Telematics and Informatics, 34 (7): 1377-1386.

Antoci, Angelo, Alexi Delfino, Fabio Paglieri, Fabrizio Panebianco, and Fabio Sabatini. 2016. “Civility vs. Incivility in Online Social Interactions: An Evolutionary Approach.” PLOS One, 11 (11): 1-17.

Bane, Kaitlin C. 2019. “Tweeting the Agenda: How Print and Alternative Web-Only News Organizations Use Twitter as a Source.” Journalism Practice, 13 (2): 191-205.

Bower, Stephanie D., and Briony D. Pulford. 2013. “Utilization of Advice from Face-To-Face and Internet-Mediated Advisors.” Journal of Technology in Human Services, 31 (4): 304- 320.

Coesemans, Roel, and Barbara De Cock. 2017. “Self-Reference by Politicians on Twitter: Strategies to Adapt to 140 Characters.” Journal of Pragmatics, 116 (2017): 37-50.

Cohen, Jacob. 1960. “A Coefficient of Agreement for Nominal Scales.” Educational and Psychological Measurement, 20 (1): 37-46.

Conway-Silva, Bethany A., Christine R. Filer, Kate Kenski and Eric Tsetsi. 2018. “Reassessing Twitter’s Agenda-Building Power.” Social Science Computer Review, 36 (4): 469-483.

Di Blasio, Paola, and Luca Milani. 2008. “Computer-mediated Communication and Persuasion: Peripheral vs. Central Route to Opinion Shift.” Computers in Human Behavior, 24 (3): 798-815.

Dohen, Marion, Jean-Luc Schwartz, and Gérard Bailly. 2010. “Speech and Face-To-Face Communication – an Introduction.” Speech Communication, 52 (6): 477-480.

Duncombe, Constance. 2017. “Twitter and Transformative Diplomacy: Social Media and Iran– US Relations.” International Affairs, 93 (3): 545-562.

World Leaders a - T w i t t e r | 74

Ewa, Kuzniar, and Filimoniuk Nataliia. 2017. “E-diplomacy on Twitter: International Comparison of Strategies and Effectivity.” Social Communication, 3 (2): 34-41.

Frame, Alex, and Gilles Brachotte. 2015. “Le Tweet Stratégique: Use of Twitter as a PR Tool by French Politicians.” Public Relations Review, 41 (2): 278-287.

Golbeck, Jennifer, Justin M. Grimes, and Anthony Rogers. 2010. “Twitter Use by the U.S. Congress.” Journal of the American Society for Information Science and Technology, 61 (8): 1612-1621.

Grant, Will J., Brenda Moon, and Janie Busby Grant. 2010. “Digital Dialogue? Australian Politicians' Use of the Social Network Tool Twitter.” Australian Journal of Political Science, 45 (4): 579-604.

Groth, Alexander J. 1964. “On the Intelligence Aspects of Personal Diplomacy.” Orbis 7: 833- 884.

Hall, Todd, and Keren Yarhi-Milo. 2012. “The Personal Touch: Leaders' Impressions, Costly Signaling, and Assessments of Sincerity in International Affairs.” International Studies Quarterly, 56 (3): 560-573.

Hallgren, Kevin A. 2012. “Computing Inter-Rater Reliability for Observational Data: An Overview and Tutorial.” Tutorials in Quantitative Methods for Psychology, 8 (1): 23-34.

Hemsley, Jeff, Jennifer Stromer-Galley, Bryan Semaan, and Sikana Tanupabrungsun. 2018. “Tweeting to the Target: Candidates' Use of Strategic Messages and @Mentions on Twitter.” Journal of Information Technology & Politics, 15 (1): 3-18.

Heritage, John. 1984. Garfinkel and Ethnomethodology. Cambridge, England: Polity Press.

Ho, Shirley S., and Douglas M. McLeod. 2008. “Social-Psychological Influences on Opinion Expression in Face-To-Face and Computer-Mediated Communication.” Communication Research, 35 (2): 190-207.

Holmes, Marcus. 2013. “The Force of Face-To-Face Diplomacy: Mirror Neurons and the Problem of Intentions.” International Organization, 67 (4): 829-861.

Honeycutt, Courtenay, and Susan C. Herring. 2009. “Beyond Microblogging: Conversation and Collaboration via Twitter.” Paper presented at the 42nd Hawaii International Conference on System Sciences.

Huberman, Bernardo, Daniel M. Romero, and Fang Wu. 2008. “Social Networks That Matter: Twitter Under the Microscope.” First Monday 14 (1): n.p.

World Leaders a - T w i t t e r | 75

Jiang, Jing, Bohan Dai, Danling Peng, Chaozhe Zhu, Li Liu, and Chunming Lu. 2012. “Neural Synchronization During Face-To-Face Communication.” The Journal of Neuroscience: The Official Journal of the Society for Neuroscience, 32 (45): 16064-16069.

Joinson, Adam. 1998. “Disinhibition and the Internet.” In Psychology and the Internet: Intrapersonal, Interpersonal, and Transpersonal Implications, ed. Jayne Gackenbach. San Diego, CA: Academic Press.

Kendon, Adam. 1988. “How Gestures Can Become like Words.” In Fernando Poyatos (ed.), Crosscultural Perspectives in Nonverbal Communication. Lewiston, NY: Hogrefe, 131- 141.

Kiousis, Spiro and Jesper Strömbäck. 2010. “The White House and Public Relations: Examining the Linkages Between Presidential Communications and Public Opinion.” Public Relations Review, 36 (1): 7–14.

Kiousis, Spiro, Michael Mitrook, Xu Wu, and Trent Seltzer. 2006. “First- and Second-Level Agenda-Building and Agenda-Setting Effects: Exploring the Linkages Among Candidate News Releases, Media Coverage, and Public Opinion During the 2002 Florida Gubernatorial Election.” Journal of Public Relations Research, 18 (3): 265-285.

Krippendorff, Klaus. 1980. Content Analysis: An Introduction to Its Methodology. Beverly Hills, CA: Sage Publications.

Lee, Jayeon, and Weiai Xu. 2018. “The More Attacks, the More Retweets: Trump’s and Clinton’s Agenda Setting on Twitter.” Public Relations Review, 44 (2): 201-213.

Li, Xiguang and Jing Wang. 2010. “Web-Based Public Diplomacy: The Role of Social Media in the Iranian and Xinjian Riots.” Journal of International Communication, 16 (1): 7-22.

Lüfkens, Matthias. 2018. “Twiplomacy Study 2018.” Twiplomacy. July 10. https://twiplomacy.com/blog/twiplomacy-study-2018/. (September 20, 2018).

McCombs, Maxwell E., and Donald L. Shaw. 1972. “The Agenda-Setting Function of Mass Media.” The Public Opinion Quarterly, 36 (2): 176-187.

McFedries, Paul. 2007. “Technically Speaking: All A-Twitter.” IEEE Spectrum, 44 (10): 84–84.

Metag, Julia, and Adrian Rauchfleisch. 2017. “Journalists' Use of Political Tweets: Functions for Journalistic Work and the Role of Perceived Influences.” Digital Journalism, 5 (9): 1155- 1172.

World Leaders a - T w i t t e r | 76

Nagel, Friederike, Marcus Maurer, and Carsten Reinemann. 2012. “Is There a Visual Dominance in Political Communication? How Verbal, Visual, and Vocal Communication Shape Viewers' Impressions of Political Candidates.” Journal of Communication, 62 (5): 833- 850.

“NATO Summit Guide.” 2018. North Atlantic Treaty Organization. https://www.nato.int/nato_static_fl2014/assets/pdf/pdf_2018_07/20180718_180711- summit-guide-brussels.pdf (December 14, 2018).

North Atlantic Treaty Organization. 2018. Brussels Summit Declaration [Press release]. July 11. https://www.nato.int/cps/en/natohq/official_texts_156624.htm (September 23, 2018).

North Atlantic Treaty Organization. 2018. Press Conference by NATO Secretary General Jens Stoltenberg Following the Meeting of the at the Level of Heads of State and Government (Nato Summit Brussels) [Press conference]. July 11. https://www.nato.int/cps/ua/natohq/opinions_156733.htm?selectedLocale=en (September 23, 2018).

Parmelee, John H. 2014. “The Agenda-Building Function of Political Tweets.” New Media & Society, 16 (3): 434-450.

Parmelee, John H., and Shannon L. Bichard. 2012. Politics and the Twitter Revolution: How Tweets Influence the Relationship Between Political Leaders and the Public. Lanham, MD: Lexington Books.

Patterson, Miles L., Mary E. Churchill, Gary K. Burger, and Jack L. Powell. 1992. “Verbal and Nonverbal Modality Effects on Impressions of Political Candidates: Analysis from the 1984 Presidential Debates.” Communication Monographs, 59 (3): 231-242.

Robinson, Jeffrey D. 2016. Accountability in Social Interaction. New York, NY: Oxford University Press.

Schandorf, Michael. 2013. “Mediated Gesture: Paralinguistic Communication and Phatic Text.” Convergence, 19 (3): 319-344.

Shapiro, Matthew A., and Libby Hemphill. 2017. “Politicians and the Policy Agenda: Does Use of Twitter by the U.S. Congress Direct New York Times Content?” Policy & Internet, 9 (1): 109-132.

Short, John, Ederyn Williams, and Bruce Christie. 1976. The Social Psychology of Telecommunication. London: John Wiley & Sons.

World Leaders a - T w i t t e r | 77

The Liberal Party of Canada. 2015. Real Change: A New Plan for A Strong Middle Class. https://www.liberal.ca/wp-content/uploads/2015/10/New-plan-for-a-strong-middle- class.pdf (February 27, 2019).

Trump, Donald (realDonaldTrump). 9 November, 2018, 1:10pm. “President Macron of France has just suggested that Europe build its own military in order to protect itself from the U.S., China and Russia. Very insulting, but perhaps Europe should first pay its fair share of NATO, which the U.S. subsidizes greatly!” Tweet.

Tsukayama, Hayley. 2017. “Twitter Is Officially Doubling the Character Limit to 280.” The Washington Post. November 7. https://www.washingtonpost.com/news/the- switch/wp/2017/11/07/twitter-is-officially-doubling-the-character-limit-to- 280/?noredirect=on&utm_term=.87a58059c61d (December 14, 2018).

Wallace, Patricia. 1999. The Psychology of the Internet. Cambridge, UK: Cambridge University Press.

“Cleanse Stop Words.” n.d. Data Science Dojo. https://demos.datasciencedojo.com/demo/stopwords/ (February 28, 2019).

“What is NATO?” n.d. North Atlantic Treaty Organization. https://www.nato.int/nato-welcome/index.html (December 14, 2018).

Yang, Guobin. 2016. “Narrative Agency in Hashtag Activism: The Case of #BlackLivesMatter.” Media and Communication, 4 (4): 13-17.

Yang, Xinxin, Bo-Chiuan Chen, Mrinmoy Maity, and Emilio Ferrara. 2016. “Social Politics: Agenda Setting and Political Communication on Social Media.” Social Informatics, 2016: 330-344.

Zakaria, Fareed. [CNN]. 2018. “French President Macron: I Always Prefer Having Direct Discussion.” [Video File]. November 11. https://www.youtube.com/watch?v=Epb5RwLXA8U&t=2s. (March 21, 2019).

Zhao, Dejin, and Mary Beth Rosson. 2009. “How and Why People Twitter: The Role That Micro-Blogging Plays in Informal Communication at Work.” Paper presented at the ACM 2009 International Conference on Supporting Group Work.