The President Is in: Public Opinion and the Presidential Use of Twitter

Abstract

Analyzing President Trump’s Tweets (N = 30,386) with the first tweet starting from 4 May 2009, this article looks at the nature of his conversations with the public and the building of public support for his candidacy, till he assumed office on January 2017. Drawing theoretically on deliberative democracy and technological populism as performance, this study, among the earliest to use interpretative qualitative analysis, reveals the different themes in his discourse, rather than only highlight specific attributes of his tweets. Our analysis shows that Trump tweets frequently and casts himself as a political outsider who can alone save America. His racist and sexist language with his confrontational style leaves no room for deliberative discourse. His messages may be populist in character, but they are aversive and uncivil and lack normative attributes of deliberation that one would expect in the leader of a powerful nation, such as the United States. These characteristics have been present in his tweets even as a private citizen. This research makes a new contribution to our understanding of how Trump uses Twitter, starting from before he emerged as a contender for the presidential office, and the discourses that emanate from his use of Twitter to make broader inferences about the messages the public is receiving from Trump.

Introduction

Donald Trump, the 45th president of the United States, uses Twitter to break news, feud with critics, and even conduct diplomacy. His Twitter use has drawn rampant media attention, as he often uses it to make controversial or even false claims, ignite his followers, or castigate his detractors. For example, in a pair of tweets in September 2018, Trump used Twitter to falsely cast doubt on the official death toll of nearly 3,000 people from Hurricane Maria, which ravaged Puerto Rico a year before (Qui, 2018). Earlier in 2018, Trump used Twitter to brag that his “nuclear button” is “much bigger” than North Korea’s (Baker & Tackett, 2018). Given the prominence of Trump’s tweets in public discourse, they have also drawn frequent scholarly attention. A quantitative analysis of 66,463 tweets from the primary season leading up to the 2016 election, for example, found that Trump made more “lying accusations” during that period than any other candidate (Kenski, Filer, & Conway-Silva, 2018). An analysis of candidates’ tweets during the 2016 general election campaign showed that Trump tended to tweet about people who endorsed or supported him or to criticize and attack others (Lee & Lim, 2016). The use of timestamps, the ways the tweets were typed (the capital letters, use of words), and tags were used deliberately to produce an “authentic form for Trump’s tweets to inhabit” (Shane, 2018).

This study builds on this research by qualitatively analyzing 303,086 of Trump’s tweets, retweets, and responses to his tweets over a longer period of time—from his pre-campaign days in 2009 to his inauguration in 2017—and using interpretative qualitative analysis to reveal the themes in his discourse, rather than only highlight specific attributes of his tweets, as other studies have done. While prior studies have analyzed Trump’s Twitter discourses as part of his campaign and afterwards as President, this study broadens our understanding of Trump’s public conversations and views, as a private citizen; much before stepping into political limelight and before any such influence may have had an impact. Trump’s tweets have been the subject of quantitative scholarly inquiry, but this is among the earliest studies that look at the corpus of his tweets through a qualitative analysis (Merriam, 2002). Thus, this research makes a new contribution to our understanding of how Trump uses Twitter and the discourses that emanate from his use of Twitter to make broader inferences about the messages the public is receiving from Trump. Theoretically, we draw on deliberative democracy (Fishkin, 1991; Guttmann & Thompson, 1996) and technological populism as performance (Baldwin-Philippi, 2018) to interpret how he communicates through Twitter. The theory of deliberative democracy argues that it is valuable for democracy for politicians and the public to be in open discourse, where multiple voices are heard, content is respectful and reasonable, and both sides are open to divergent viewpoints (Fishkin, 1991; Guttmann & Thompson, 1996; Jacobs, Cook, & Delli Carpini, 2009). Baldwin-Philippi (2018) argues that the Trump campaign’s use of digital platforms contributed to what she calls “technological performance of populism” (p. 2) by centering people in the campaign and speaking directly to them and highlighting their voices. As our findings show, Trump may portray himself as the lone outsider who can save the country, but he maintains no balance in populism and civility, using rhetorical devices like capital letters associated with incivility frequently in his tweets and retweeting only his supporters while being extremely insulting to detractors.

Literature Review

Presidential Use of Twitter

In the 2008 election, when the Obama campaign heavily invested in social media, few realized what a game changer it would be (Adams & McCorkindale, 2013; Towner & Dulio, 2011). Later, John McCain too used the Internet effectively as a fund-raising tool, but its success was nowhere near the Obama campaign (Adams & McCorkindale, 2013). The potentials of social media are most promising in the political context as they can be an enabler for more participation and democracy. Hong and Nadler (2011) have shown that the political use of Twitter by US politicians have no impact on changing or influencing public opinion either negatively or positively. Boulianne (2015) in her meta-analysis of research on social media use and participation has shown that there are positive relations between social media use and participation, but questions remain about whether these relationships are causal and transformative. But that does not stop politicians from using Twitter profusely. For example, Adams and McCorkindale (2013) showed that between, 1 February 2012 and 29 February 2012, when the campaigns were being actively run, Barack Obama “tweeted” the most with 273 tweets during February 2012, followed by Newt Gingrich and Rick Santorum, both with 113 tweets. Ron Paul had 59 tweets, and Mitt Romney “tweeted” the least, with 47 tweets (Adams & McCorkindale, 2013).

Twitter is increasingly being used within the sociopolitical domain as a channel through which to circulate information and opinions (Ross & Rivers, 2018). Adams and McCorkindale (2013) found that presidential candidates used Twitter mostly for information distribution; information exchange, interaction, and transparency did not feature significantly. For example, Obama’s Twitter account clearly stated that his campaign staff was responsible for the content of the tweets. The tweets by Obama were signed accordingly. Ron Paul’s tweets were either written in first person as “I” or third person as “Ron Paul.” His Twitter page did not state who was responsible for tweeting—whether his campaign staff managed the page or not. Romney, Santorum, and Gingrich all appeared to tweet themselves because each tweet was in first person. As Adams and McCorkindale (2013) stated, the electorate has no way of knowing whether the voice belongs to the candidate or not. None of the candidates in the study used retweeting as a means of engaging in meaningful political dialogue with their constituents. This clearly showed that candidates were only interested in disseminating information about themselves (Adams & McCorkindale, 2013). Few of the candidates worked to engage with their audiences. Only Obama and Gingrich asked questions through tweets; Obama retweeted questions his followers had for him, while Gingrich did not do so. There was little dialogue on Twitter even though each candidate had a significant number of followers.

A study (Golbeck, Grimes, & Rogers, 2010) that analyzed the contents of more than 6,000 tweets from Congress members found that members treated Twitter as a vehicle for self-promotion, using the platform to share information like news stories on them, daily activities, and their blogs rather than to interact with constituents. Participatory communication may earn legislators’ political capital more than proactive communicators who offer participatory opportunities that are more likely to build an online following (Lilleker & Koc-Michalska, 2013). Social media does have the potential to influence power relations in political parties as they allow individual candidates to campaign more independently of the central party (Karlsen & Enjolras, 2016).

In the 2016 presidential campaign, we saw a more nuanced use of Twitter. Twitter was used prolifically by both candidates, Hillary Clinton and Donald Trump, but the latter’s use was seen as especially unorthodox in the area of political campaigns due to the fact that his tweets came directly from him, unmediated by advisers and other campaign staff (Enli, 2017). While the Clinton campaign’s strategy only occasionally broke with expectations and using elements of “real talk” to underline her authenticity, the Trump campaign’s more amateurish and seemingly more authentic style in social media points toward deprofessionalization and amateurism as a counter-trend in political communication (Enli, 2017). Social media like Facebook and Twitter place the focus on the individual politician rather than the political party, thereby expanding the political arena for increased personalized campaigning (Enli & Skogerbø, 2013). Lee and Lim (2016) showed that Trump’s and Clinton’s campaign tweets, reflected the political candidates’ gendered communication strategies. The study found that Clinton emphasized her masculine personality traits and feminine issues and Trump mentioned masculine issues more with no particular attention to traits. Clinton focuses on expressing her opinions about public issues through tweets, while Trump uses Twitter to share citizens’ supportive quotes (Lee & Lim, 2016). Crockett (2016) noted that words like “good,” “bad,” and “sad” were very common in his Tweets besides the use of capital letters and exclamation marks. Among the earliest scholars to examine the rhetorical devices used by Trump, in his speeches, Jennifer Mercieca (2015) raises an important point in her essay titled, “The rhetorical brilliance of Trump the demagogue.” Analyzing, Trump’s December 7, Statement on Preventing Muslim Immigration, she says that while it drew widespread disdain, polls also found that 37% of voters agreed with a “temporary ban” on Muslims entering the United States. While his arrogance and capricious communication may be revolting to voters, he has extremely strong holds on certain parts of the voter base.

One notable feature of Trump’s Twitter use has been his repeated naming of mainstream media through pejorative identifications such as “fake news” and “fake media.” Ross and Rivers (2018) show that Trump uses these accusations to demonstrate allegiance and as a cover for his own spreading of misinformation, framed as truth. Semantic network analyses have shown that in using the campaign slogan, “Make America Great Again” (MAGA), Trump communicatively organized and controlled media systems by offering his followers an opportunity to connect with his campaign through the discursive hashtag, as well as exposed connections to overtly White supremacist groups within the United States and the United Kingdom throughout late November 2016 (Eddington, 2018). Shows that supporters on Twitter view Trump as truthful and frank. As a strategy of appeal, Trump’s aggressive, eccentric, and successful manner of attack may signify a possibly enduring political and party alignment (Appel, 2018).

Deliberative Democracy

The theory of deliberative democracy suggests that it is important or even essential for there to be a two-way discourse between politicians and their constituents because a government by the people requires a free exchange of ideas. Deliberation is intended to promote discussion, move toward solutions, and encourage truthful and balance discourse (Gastil, 2008; Guttmann & Thompson, 1996). It requires attributes such as logic, use of evidence, and rational arguments (Papacharissi, 2004; Stroud, Scacco, & Muddiman, 2015). The theory analyses political process as one in which people work with their fellow citizens to reach sharable conclusions that reflect the reasons offered in the deliberative process. Rawls (2005, pp. 134-72) argues that these considerations must be justifiable on grounds acceptable to members of a variety of distinct reasonable “comprehensive doctrines,” or worldviews, that form an “overlapping consensus” on citizens’ status as free and equal and on accompanying basic liberties (Layman, 2016). A central discussion on the theory of deliberative democracy in recent decades has focused on whether democratic deliberation, and consequently those participating in it, should aim, at least ideally, for political consensus (Martí, 2017). Disrespectful speech, which is defined as incivility (Coe, Kenski, & Rains, 2014), or more virulent forms of discourse, such as racist or sexist speech, violates the norms of deliberative discourse (Chen, 2017). Most theorists agree that deliberation includes some forms of “public talking” (Jacobs et al., 2009, p. 4) that contribute to public opinion, and increasingly, scholars have turned to social media as a form of this public talking (Camaj & Santana, 2015; Oz, Zheng, & Chen, 2018). Given that Trump used Twitter to talk directly to the American people before he became a candidate, during his campaign, and now since his election, this theory provides a robust way to understand the discourses inherent in his tweets.

Technological Performance of Populism

Theories of technological performance of populism also inform and help us interpret Trump’s tweets. The technological performance of populism involves a process by which “individuals create populist identities or qualities” (Baldwin-Philippi, 2018, p. 3) through their digital performativity. In this sense, digital performativity “implies not a simple expression of action, but a complex amalgam of a performance and production” (Leeker, 2017, p. 22) of an individual. Populism is an “appeal to ‘the people’ against both the established structures of power and the dominant values of a society” (Canovan, 1999, p. 3). Thus, populist movements, like social movements, challenge the power structure, but populism also targets elites, such as academics and the media (Canovan, 1999). Trump ran for election with a populist message of “draining the swamp” of political insiders, and his campaign engaged in what Baldwin-Philippi (2018) calls a “technological performance of populism” (p. 18) across digital platforms, such as Twitter and Instagram, in an effort to make the candidate appear authentic and unfiltered. We impose this same lens of technological populism performance on our analysis of Trump’s tweets from before he was a candidate to his inauguration day.

Given the increased relevance of political communication in social media, it is also important for politicians and parties to use social media more proactively to enter into dialogs and discussions with citizens (Dang-Xuan, Stieglitz, Wladarsch, & Neuberger, 2013). Thus, the deliberative quality of tweets becomes more important. Social media, such as Twitter, offers a fundamental change in traditional public communication, which has usually been exclusively initiated and managed by specific actors, for example, politicians, as well as journalists (Stieglitz & Dang-Xuan, 2013). Tweets themselves have become news and thus essentially free advertising for candidates (Gross & Johnson, 2016). For journalists, the use of Twitter has caused substantial changes to daily reporting practices and Twitter is considered more consequential for their job than any other form of social media, including Facebook (Parmelee, 2012). For citizens, the Internet offers easy access to political information, providing all kinds of opportunities to participate in political debates (Kruikemeier, Van Noort, Vliegenthart, & De Vreese, 2013). And there are few clearer examples than the 2016 elections.

Twitter in particular can be accused of promoting frequently denigrating and dehumanizing discourse (Ott, 2017). Lee and Lim (2016) have shown that while Trump actively retweets citizen supporters’ tweets, 10.5% of his tweets were uncivil. Ott (2017) argues that Twitter’s requirement of short messages demands simplicity that “ultimately trains us to devalue others, thereby, cultivating mean and malicious discourse” (p. 60). Ott (2017) says Twitter is defined by three key features: simplicity, impulsivity, and incivility. Messages on Twitter cannot be sophisticated; there is little space to explain, analyze, or asses. As Kapko (2016) has reiterated, Twitter cripples our capacity to deliberate about issues and events in complex ways. Stieglitz and Dang-Xuan (2013) have shown that heavy Twitter users appear to have a desperate, even compulsive, need for attention, and to ensure that they get that attention, “they tend to post more emotionally charged tweets” (p. 214). Considering President Trump, Ott (2017) believes that “Trump’s natural style of speaking and Twitter’s underlying logic are wholly homologous” (p. 63). These studies have contributed greatly to our understanding of Trump’s style of political communication on Twitter, but they do not examine the common themes that arise from the organic body of tweets and the significance of those subjects, through a qualitative lens, nor do they examine their value to enhance deliberation and democratic conversation.

Trump and Twitter

Trump has had a long history on Twitter. As the BBC (http://www.bbc.com/news/world-us-canada-38245530) has reported, in May 2009, Trump, then a private businessman, sent his first tweet promoting a forthcoming appearance on the Late Show with David Letterman. During the 2016 campaign, Trump was able to generate considerable unpaid or free media for himself, often directly through Twitter (Francia, 2017). Scholars have asserted that he successfully blurred lines between entertainment and politics, fusing “mediation, visibility and attention” and using his brand name and notoriety to become a sort of populist hero (Wells et al., 2016). In a 2015 article, “Pithy, Mean and Powerful: How Trump Mastered Twitter for the 2016 Race,” the New York Times (NYT) wrote, Trump had made Twitter “a centerpiece of his campaign” using it for “political promotion, distraction, score-settling and attack.” Previous studies have examined the impact of attacking opponents based on their personal traits, issue positions, or the political party to which they belong (Skaperdas & Grofman, 1995) and focused on the effects of voter turnout (Peterson & Djupe, 2005).

Few have examined why candidates will choose to be negative in the first place. As Gross and Johnson (2016) show candidates grow increasingly negative and tweet more frequently as the field narrows and as voting progresses and opportunities dwindle. Generally, negativity is nearly always directed from lower to higher status candidates. But in the 2016 Grand Old Party (GOP) nomination contest, Trump flouted this norm with brutal remarks aimed at even low-polling candidates (Gross & Johnson, 2016). As the NYT said, “To an unprecedented degree in American history, Mr. Trump has made personal insults and attacks part of his campaign” (Lee, & Quealy, 2016). He also made more lying accusations—55—before the 2016 primary campaign started, and this constituted half of all lying accusations made by major party candidates during this period (Kenski et al., 2018). Therefore, in this study, we focus on the discourses that surface in Trump’s tweets, rather than focus on particular attributes of the tweets. We aimed to interpret and understand how deliberative and populist discourse emanate from his tweets to provide a more holistic view of the meaning inherent in his tweets. Thus, we proposed the following research questions:

• RQ1. What are the major themes in Trump’s tweets during the period from before he became a presidential candidate to his inauguration?

• RQ2. What (a) deliberative discourses and (b) populist discourses emanate from these themes?

Method

This study examined a total corpus of 30,386 tweets from 4 May 2009 to 27 January 2017. This comprised the period from when Trump was a private citizen to his inauguration as the 45th US President on 20 January 2017. Trump announced his candidacy for the US presidency at Trump Tower in New York City on Tuesday, 16 June 2015, and by a year after his election, by 2017, he was the most unpopular president in the history of modern opinion polls (Enten, 2018). In 2016, he averaged 375 tweets a month through to the end of November, according to TrumpTwitterArchive.com, a searchable database dedicated to cataloging all of Trump’s tweets. The current data set was downloaded from the archive. This set includes all of Trumps tweets as well as his retweets and responses to his tweets. Our intention was to use qualitative textual analysis and examine Trump’s tweets to understand his intentions as articulated on Twitter, leading up to the presidency.

Research using qualitative textual analysis focuses on the characteristics of language and visuals as communication with attention to the content or contextual meaning of the text and visual (McTavish & Pirro, 1990). Our goal was to reveal the discourses in which these tweets operate, illuminating implied meaning to make more sweeping statements and inferences (McTavish & Pirro, 1990) about the meaning in Trump’s tweets leading up to his election and ultimately his inauguration. Qualitative analysis often views culture as a narrative or story-telling process in which particular “texts” or “cultural artifacts” (i.e., a pop song or a TV program or visual) consciously or unconsciously link themselves to larger stories at play in the society. While quantitative approaches are very useful to count the number of tweets and attach scores of ‘sentiment’ to individual tweets, can such methods identify issues such as the presence or absence of empathy or incivility (Karamshuk, Shaw, Brownlie, & Sastry, 2017)? Thus, content analysis, with its variable analytic style, is more suitable for answering the “why” questions and those that deal with causes and effects while the focus of qualitative analysis is on the “how” questions—focusing on the processes through which things come to be the way they are (D’Angelo, 2002; Ettema, 2005; Reese & Lewis, 2009). Our focus is the how of Trumps tweets and that is why qualitative textual analysis is a suitable method. The product of such qualitative inquiry is rich in description of the phenomenon observed and the results are more effectively conveyed in words and pictures rather than numbers (Thomas, 2006). Such a methodology focuses on the description of the context and data are usually in the form of quotes from documents or platforms rather than percentages or numbers.

Our analysis was guided by prior literature, principle characteristics of qualitative interpretive analysis, and the theories of deliberative democracy and technological populism. Thus, we examined the words and language used to analyze how much deliberation did Trump encourage on issues, how much discussion did he promote, and how did he create an identity online through his tweets. Studies by Ott (2017), Lee and Lim (2016), and others have shown that Trump actively retweets citizen supporters’ tweets, while about 10.5% of his tweets were uncivil. Such tweets were designed to devalue “others” and promote “mean and malicious discourse” (Ott, 2017, p. 60). Thus, we were aware of certain elements (incivility, demeaning nature of tweets) that may have been present, and we were also aware that this corpus of tweets was from a time Trump was a private citizen and had not steeped into the political arena. Thus, to overcome bias and ensure a thorough examination of the data, the whole corpus of tweets was read four times (in total) by both researchers involved in the study before themes organically emerged. This approach involves focusing on the “underlying ideological and cultural assumptions of the text” (Fürsich, 2009, p. 240), rather than only emphasizing the literal meaning of words and sentences. Thus, we both focused on specific words, such as “media,” “immigration,” “president,” “great,” and “business,” but also on context, syntax, metaphor, and meaning in those words. This analysis strategy allowed us the “opportunity to step into the mind of another person to see and experience the world as they do” (McCracken, 1988, p. 9) through their public words. Of course, no one can actually know what another person thinks or feels, but this approach uses the public words to make inferences about those thoughts and feelings. The first reading (done by both researchers) initiated a multi-stage process of categorizing and coding the data where the corpus was coded roughly along major categories (Flick, 2007). Content was then categorized into broad groupings, which were narrowed, using a constant comparative method (Johnson, 2000). This involved merging similar categories or finding new categories based on a closer reading of data (Cresswell, 1994). This iterative process was repeated until consistent themes emerged. In the next phase, the main topical groups were further developed and characterized based on the number of tweets that fell into each category. The entire data set was then read again, based on the categories, and the grouping of the tweets was refined. The final reading involved a rechecking of the data coded according to the final set of categories determined. There was an agreement on all the themes between the two readers.

Results

In this section, we discuss some overall findings about Trump’s Twitter style before detailing and providing examples of the three major themes that surfaced in our data and answer our specific research questions. Our analysis of Trump’s tweets showed that for the President, Twitter is a medium of attack and defense. Overall, the content and language of his tweets, especially the extensive use of words in all capital letters—a grammatical device linked to incivility (Chen, 2017)—were striking. Very often, his tweets either angrily refuted what others were saying or reacted against perceived slights. For example, in a 2016 Tweet, Trump wrote, “Happy New Year to all, including to my many enemies and those who have fought me and lost so badly they just don’t know what to do. Love!”

Trump is a prolific user of Twitter, tweeting four to eight times a day. The tone of his tweets is conversational, and the language is colloquial. He responds to praise and brickbats almost as if he is having a face-to-face conversation with an adoring audience at large. For example, writing about the NYT and Clinton, he says, “The failing @nytimes has gone nuts that Crooked Hillary is doing so badly. They are willing to say anything, has become a laughingstock rag!.” His tweet about Clinton’s campaigning: “Why isn’t President Obama working instead of campaigning for Hillary Clinton?” He refers to Edward Snowden, a former Central Intelligence Agency (CIA) operative who leaked classified US government information, as “that piece of human garbage” in a 2014 tweet.

Twitter also seems to be a way for him to gauge public reaction. For example, in 2014, he tweeted, “So many people think I will not run for President. Wow, I wonder what the response will be if I do. Even the haters and losers will be happy!” But these conversations are often one-sided. Trump does not encourage two-way exchanges in his tweets unless the other person is praising or supporting him. He speaks directly to the public to offset what is being said in news stories. For example, during a trial for a fraud case arising out of Trump’s for-profit Trump University, Trump retweeted “@seankesser: @realDonaldTrump Trump University had 98% approval ratings, beats Harvard, NYU & other top universities” (2014). Or he retweets those who compliment him, such as this tweet from 2016: “@alphainparis: @seanhannity DonaldTrump is the 21st century Ronald Reagan. Intelligent. Bold. Surrounded by good people. Gets the job done.”

He seldom uses news sources, facts, or figures. Many of his tweets are opinions. For example, in 2014, he tweeted, “The U.S. accidentally air dropped a large shipment of military weapons and supplies right into the middle of ISIS as enemy laughs! Very sad! (2014).” No credible news organization reported on this story.

“Disgusting” is a very commonly used word and sentiment in regard to politics and the media. For example, in 2013, he tweeted “Will be on Fox & Friends tomorrow morning at 7.00. Will be discussing the disgusting and wasteful $635 million website rollout and more!” Another example, from 2016, “.@CNN is so disgusting in their bias, but they are having a hard time promoting Crooked Hillary in light of the new e-mail scandals.” He also favors similar words, such as “disrespectful” and “horrible.” For example, in 2016, he tweeted about Barney Frank, a former member of the US House of Representatives and board member of the New York–based Signature Bank, writing “Barney Frank looked disgusting—nipples protruding—in his blue shirt before Congress. Very very disrespectful.”

Trump also used Twitter extensively to promote his public appearances or imagined public service. For example, he tweeted in 2015, “I have a proven track record supporting our Veterans. Veterans deserve universal access to care. VA scandal proves politicians are inept.” The tweets do not contain links to support his contention that he has helped veterans. His television appearances are routinely publicized. For example, he tweeted in 2016, “My @greta int. on @FoxNews on how to defeat ISIS, Obama losing ground to ISIS, & Making America Great Again! http://t.co/1d4mPmJFaa.”

His family is also given space, especially his son Eric and daughter Ivanka. This tweet from 2016 illustrates this point:

My wonderful son, Eric, will no longer be allowed to raise money for children with cancer because of a possible conflict of interest with . . . my presidency. Isn’t this a ridiculous shame? He loves these kids, has raised millions of dollars for them, and now must stop. Wrong answer!

Trump’s emphasis on his children’s accomplishments makes him sound almost like a boss than a parent. Clearly, the insistence is on achievement and it is promoted as an approach that gets things done.

Major Themes

Besides the idiosyncrasies of Trump’s tweeting style explained above, three major themes surfaced in the data. These were the outsider who will make America great, which focuses on Trump’s populist promises to restore the nation to an imagined former glory and give the country back to the people; racism, misogyny, and hate, which exposes the acerbic underbelly of his discourse; fake news, which includes his attacks on the news media and foreign policy by Twitter, which comprised his beliefs about Russia and North Korea. Each theme is described in detail in the following. Overall, the three themes answer our first research question that asked, what major themes surface in his tweets. The answers to our second research question, which examined both the deliberative and populist discourses in his tweets, are explained in the context of each theme.

The Outsider Who Will Make America Great

This theme suggested that Trump views himself as a political outsider—or populists—because of his lack of previous political experience, but he sees this lack of experience as an asset to the country. This theme is prevalent in his tweets. An example is this tweet from 2016: “Politicians are all talk and no action. Washington can only be fixed by an outsider. Let’s make America great again!” which links to his campaign page. In another tweet from 2016, he promised that, as someone who is not a Washington elite, “. . . our administration will ALWAYS have your back. We will ALWAYS be with you!”

He eschews diplomatic language and focuses on hot-button issues that he knows will resonate with his followers, casting himself as the savior for America. For example, he tweeted in 2013, “We should be concerned about the American worker & invest here. Not grant amnesty to illegals or waste $7B in Africa.” He draws on his business experience in his tweets, to suggest that his business acumen will build up the US economy. His tweets consistently blame Mexico for US trade deficits, with little appreciation for the nuances of how trade affects the economy, as exemplified by this tweet from 2017: “Mexico has taken advantage of the U.S. for long enough. Massive trade deficits & little help on the very weak border must change, NOW!” In his tweets, he links his campaign slogan—Make America Great Again (or MAGA)—to stimulating jobs and expanding US nuclear capabilities, offers some insight into what he sees as a great America. These tweets from 2017 illustrated this point: “Great meeting with automobile industry leaders at the @WhiteHouse this morning. Together, we will #MAGA! and “The United States must greatly strengthen and expand its nuclear capability until such time as the world comes to its senses regarding nukes.”

In tweets leading up to the election, Trump cast himself as the decisive leader who alone could protect the people. For example, his disparagement of President Barak Obama in 2015 set a tone for this campaign both on and off Twitter: “All he does is go on television is talk, talk, talk, but incapable of doing anything.” His tweets highlight what he saw as the indecisiveness of his predecessors whom he claims sold the country out. For example, Trump tweeted in 2011: “@BarackObama sold guns to the Mexican drug cartels. They were used in the murders of Americans. Where is the outrage?” He credits himself as to the leader who can fix problems. Two tweets from 2017 are a good example: “If Chicago doesn’t fix the horrible ‘carnage’ going on, 228 shootings in 2017 with 42 killings (up 24% from 2016), I will send in the Feds!” and “As your President, I have no higher duty than to protect the lives of the American people. https://t.co/o7YNUNwb8f.”

A subtheme surfaced within this theme that emphasized Trump’s populist message of restoring America to its citizens when he is elected. In early 2017, he tweeted that his presidency was not just “. . . transferring power from one Administration to another, or from one party to another—but we are transferring . . . power from Washington, D.C. and giving it back to you, the American People. #InaugurationDay.” Other tweets emphasize his populist message from 2017: “. . . Even if I don’t always agree, I recognize the rights of people to express their views” and “What truly matters is not which party controls our government, but whether our government is controlled by the people.” Trump, who has registered as both a Democrat and Republican over the years, seems to be saying that he would not be beholden to any party but rather will focus his attention on the American people—a pledge that resonated strongly with many in the electorate. He repeatedly promised to “drain the swamp” of Washington insiders, although as his presidency continued he actually appointed many insiders to key posts. Many tweets exemplify this message, including these tweets from 2016: “I will Make Our Government Honest Again—believe me. But first, I’m going to have to #DrainTheSwamp” and “In order to #DrainTheSwamp & create a new GOVERNMENT of, by, & for the PEOPLE, I need your VOTE! . . . LET’S #MAGA!”

Overall, this theme offers evidence of Trump’s deployment of a technological performance of populism that mirrors the tact his campaign employed across e-mail, Twitter, Instagram, Facebook, and campaign-created apps (Baldwin-Philippi, 2018). Thus, this theme answers RQ2b, which asked about the populist discourse that surface in Trump’s tweets. A populist theme was quite predominant in Trump’s tweets, suggesting a strong resonance with the identity that Trump is trying to perform through social media.

Racism, Misogyny, and Hate Speech

A second predominant theme that surfaced in the data relates to the language, tone, and content of Trump’s tweets and how they castigate people of color, immigrants, women, and other marginalized groups. All these types of speeches fit the general definition of uncivil speech, which is content with a disrespectful tone (Coe et al., 2014) that ranges from mere impoliteness to hate speech (Chen, 2017). However, some of these tweets also may fit the definition of hate speech, which is content that perpetuates negative stereotypes or biases or is intended to humiliate, insults, or denigrate marginalized groups (Davidson, Warmsley, Macy & Weber, 2017).

For example, Trump uses racially loaded terms, such as thugs, or denigrates immigrants by calling them “disgusting.” In 2016, he tweeted, “The protesters in New Mexico were thugs who were flying the Mexican flag. The rally inside was big and beautiful, but outside, criminals!” This tweet referred to demonstrators outside a Trump rally in New Mexico who set fire to campaign merchandise and clashed with police to protest Trump’s plans to build a wall at the US border with Mexico to keep immigrants out. What is problematic about the tweet is its sweeping stereotypical nature that casts all protestors of the event as criminals, while only some perpetuated the more virulent acts. Other tweets do not overtly use hateful words, but they perpetuate stereotypes by highlighting controversial—but incorrect—information that immigrants are afforded benefits that are withheld from citizens, including veterans. For example, he tweeted in 2016, “Notice that illegal immigrants will be given ObamaCare and free college tuition, but nothing has been mentioned about our VETERANS #DemDebate-2016.” These tweets from 2015 also exemplify this theme: “We will soon be at a point with our incompetent politicians where we will be treating illegal immigrants better than our veterans” and “It’s a national embarrassment that an illegal immigrant can walk across the border and receive free health care . . .”

A subtheme in this theme focuses on treatment of women in Trump’s tweets. It is notable that misogynistic comments were generally missing from tweets between 2009 and 2014, when he merely mentioned women in passing or focused on his female supporters. For example, “I have so much admiration and respect for the 2.4 million men and women of our Armed Forces” Trump tweeted.

But misogyny in his tweets seemed to surface once a 2005 video recording where Trump bragged about grabbing women by the vagina without their consent became public in fall of 2016. Trump came under fire for the video, and some Republican supporters distanced themselves from him temporarily (Burns, Martin, & Haberman, 2016), although overall the depiction of Trump’s misogyny seemed to leave his followers unfazed (Chen, Pain, & Zhang, 2018). In his tweets, women, especially Clinton, became the focus of vicious misogyny. Hyper aggressive and extremely reactive, the tweets that followed demeaned and discredited. He denounced the NYT for being among news organizations that revealed the video, as well as that reported accusations of sexual misconduct by nearly 20 women (Saslow, 2018). He called the Times’ stories “. . . totally phoney stories, 100% made up by women (many already proven false) and pushed big time by press, have impact.” Later, he tweeted, “New polls are good because the media has deceived the public by putting women front and center with made-up stories and lies and got caught.” As Vickery and Everbach (2018) note, Trump appeared to have purposefully used “his personal Twitter account to unapologetically broadcast and draw attention to his atrocious views and behavior” (pp. 3-4). We found the same in our data: “Trump’s misogyny was blatant” (Vickery & Everbach, 2018, p. 4)

Reacting to the accusations against him, he tried to shift the focus on his misdeeds by retweeting supporters’ tweets naming Joe Biden as the one with a “. . . Long History Of Grabbing, Kissing and Groping Women Who Are Cringing.” Yet, he also seemed to acknowledge his misogynistic language in the 2005 video as recording but dismissed it as trivial, calling it merely “locker room talk.” The term “locker room talk” was interpreted as “what men do when they are together in spaces free from women’s ears” (Harp, 2018, p. 198). “I’m not proud of my locker room talk,” he tweeted, but later said “. . . this world has serious problems. We need serious leaders. #debate #BigLeagueTruth.” He also tried to shift the focus onto the well-publicized sexual misdeeds of Clinton’s husband, former President Bill Clinton, a trope Trump’s followers took up with vigor. “There’s never been anyone more abusive to women in politics than Bill Clinton. My words were unfortunate-the Clintons’ actions were far worse” he tweeted in 2016. These tweets offer evidence of Trump’s overall tactic to downplay the importance of the video, a strategy he continued soon after the event by releasing a recorded apology where he attempted to normalize his actions (Harp, 2018). The tweets we found support findings from a discourse analysis of Twitter hashtags, news stories, satire broadcasts from this period in the campaign that demonstrated how Trump framed his sexual improprieties as trivial—a frame reinforced within the campaign and in media reports (Harp, 2018).

Taken as a whole, this type of vitriolic discourse in Trump’s tweets clearly violates the norms of deliberative speech. As the aim of deliberation is to be inclusive of different viewpoints and foster cohesion (Gastil, 2008; Guttmann & Thompson, 1996; Jacobs et al., 2009), the tweets in this theme appear to do the opposite. They have the power to foment discord and division, rather than draw people together. As other scholars have shown (e.g., Kenski et al., 2018; Lee & Lim, 2016), our analyses suggest quite strongly that Trump’s tweet lack attributes of deliberative discourse, answering RQ2a.

Fake News

A third prominent theme in our data was Trump’s assertions that mainstream media outlets were reporting inaccuracies about him—which he calls fake news. In the scholarly literature, fake news is defined as misperception, or a belief in information that is false, misleading, or unsubstantiated (Nyhan & Reifler, 2010). The term fake news is fraught because it means different things to different people, but Trump’s tweets illustrate how he used the term as a weapon to attack discourse that did not fit his worldview or that revealed facts he would rather were kept hidden. His tweets demonstrate that he saw news outlets, such as CNN, NBC, and the Times as clear enemies. Trump’s battle with the media began in earnest when his misanthropic treatment of women began to be covered in detail. He claimed in a tweet from 2016 that “can you believe I lost large numbers of women voters based on made up events THAT NEVER HAPPENED. Media rigging election!” He made this claim even though he, ultimately, won the election. He called media outlets that disagreed with his views “biased.”

The fake news controversy is one that dogged Trump’s campaign and now his presidency. Actor Alec Baldwin’s comedic portrayal of Trump on Saturday Night Life particularly provoked him. “.@NBCNews is bad but Saturday Night Live is the worst of NBC. Not funny, cast is terrible, always a complete hit job. Really bad television!” he tweeted in 2017. Even more troubling, responses to Trump’s tweets suggest his followers seem to believe him, as evidenced by this tweet from 2017: “@levisteveholt: @realDonaldTrump I appreciate your use of Twitter to keep us informed and maintain transparency. Very dishonest media!”

This theme also illustrates that Trump exhibits a populist bent in his tweets that rallies against elites, such as the media, again answering RQ2b. In addition, this theme demonstrates that Trump appears unburdened by the norms of deliberative discourse, which require truthfulness and balance in political conversations. Kenski and colleagues (2018) argue persuasively that accusations that others are lying or willfully misleading others “undermine respect through credibility attacks focused on distrust” (p. 288). Thus, accusations that the media are lying or tweets that support misinformation clearly violate both the spirit and letter of deliberative discourse, answering RQ2a.

In summary, Trump’s tweets across the 8 years we studied demonstrate a strong populist theme that not only challenge conventional authorities, such as the government structure and the media, but also paint Trump as the only savior for the nation. His tweets reach Americans directly and give him an enviable method of talking directly to the people. Thus, they could be a boon of deliberative dialogue. Our analysis shows the opposite. His tweets perpetuate division, misinformation, and lies and lack any semblance of deliberative discourse. He does not use evidence to support points, which could further his deliberative reach, and instead, he attacks in a manner that leads to greater divisiveness and rancor.

Discussion

This study sought to understand the discourses in Trump’s tweets from his pre-campaign days in 2009 to his inauguration in 2017, using the rich data from more than 30,000 tweets analyzed in using an interpretive, qualitative analysis. We sought to uncover meaning in the corpus of tweets by categorizing the tweets, employing the technique of closing reading and comparison (Johnson, 2000) to allow predominant themes to emerge. While other research has examined attributes of his tweets (e.g., Kenski et al., 2018) our goal was to understand the discourses that emanate from these tweets to make broader inferences about the messages the public is receiving from Trump. We employed deliberative democracy (Fishkin, 1991; Guttmann & Thompson, 1996) and technological populism as performance (Baldwin-Philippi, 2018) to interpret how he communicates through Twitter. We begin by contextualizing our findings more generally before offering insights more specifically into the discursive elements of his tweets that fit within the discourses of deliberative democracy and digital populist performance.

Overall, our analysis demonstrated that Trump tweets frequently and make frequent use of grammatical devices, such as words in all capital letters or insults, that are linked to incivility (Chen, 2017). He uses a colloquial, conversational style of language that seems to illustrate his attempts to be authentic and unfiltered, which underscores his populist message (Baldwin-Philippi, 2018). He does not encourage two-way communication, which would encourage true deliberative discussion. Instead, he either highlights praise from his followers or attacks his detractors in a manner that creates division that challenges any potential for deliberative discourse. For example, in 2017, he tweeted, “Congratulations to @FoxNews for being number one in inauguration ratings. They were many times higher than FAKE NEWS @CNN—public is smart!” These findings dovetail with earlier research that has found supportive quotes or endorsement of Trump comprised nearly 40% of his tweets, followed by criticism or attacks, which encompassed 25% (Lee & Lim, 2016). Our study expands on this earlier research by providing the rich context from excerpts from his tweets to demonstrate this trend.

When we delved more deeply into the themes that surfaced in our data, we found a very prominent theme was that Trump, through his tweets, cast himself as a political outsider who can alone save America. His tweets showed that he abandoned diplomatic language and focuses on controversial issues that would resonate with his populist following. Populist language is rampant in Trump’s tweets in this theme, as he emphasized that if elected he would transfer government power to the electorate and attenuate the power of political elites. His use of sometimes exaggerated or uncivil language highlights this populism, as he attacks former presidents, the media, and Hillary Clinton, his political opponent for the presidency. On the subject of his political rival, Hillary Clinton, he commonly uses the term “Crooked Hillary” to refer to her. For example, in 2016 he tweeted, “Crooked Hillary has been fighting ISIS, or whatever she has been doing, for years. Now she has new ideas. It is time for change.” Thus, his populist message cast Clinton as part of the elite and situates himself as the anti-Clinton. His tweets fit squarely into a performance of populism through the technology of Twitter, much as his campaign did through Twitter and other forms of digital media (Baldwin-Philippi, 2018).

Of course, one can be populist without being acerbic, but our analysis showed that Trump does not strike this balance. Not only does he seldom employ the deliberative technique of using evidence to support his views, the racist and sexist language and hate speech in many tweets challenge any potential for deliberative discourse. Our second theme examined these types of uncivil speech, showing how he eschewed deliberation and instead embraced a defensive attack tone. For example, his views on how the media covered him are clear, and he uses gendered antagonistic language in his criticism of journalists, especially women journalists. “There are many editorial writers that are good, some great, & some bad. But the least talented at all is frumpy Gail Collins of the NYTimes,” he tweeted in 2016. These findings support research that has found Trump framed his sexual improprieties as trivial (Harp, 2018) and instead employed blatant misogynist language (Vickery & Everbach, 2018).

Our third theme, which focuses on Trump’s false claims in tweets that news organizations are perpetuating “fake news,” also fit into the technological performance of populism and lack of deliberation. He is free in his attacks on the media, which from a populist viewpoint are among the elites that should be overthrown. He also takes no heed of deliberative norms that require truthful respectful discourse and, instead, employs a confrontational style in his tweets. For example, in a 2012 tweet, he attacked Time Magazine for doing a story about attachment-style parenting that included a picture of a women breastfeeding: “The TIME Magazine cover showing late age breast feeding is disgusting—sad what TIME did to get noticed. @TIME.”

In summary, Trump’s tweets demonstrated quite conclusively how he perpetuates a populist messages that is both aversive and uncivil and lacks any normative attributes of deliberation that one would expect in the leader of a powerful nation, such as the United States. He strikes a colloquial tone that appears to be authentic and unfiltered and cast him as at odds with both the government system and elites, such as the media. In addition, the way Trump raises himself up in his tweets as a savior of America demonstrates the populist trope that “populism is not just a reaction against power structures but an appeal to a recognized authority” (Canovan, 1999, p. 4)—namely him. He is free with his criticism, uncivil language, and tone, and he takes particular aim at women, people of color, and the media. Taken as a whole, the analysis of this corpus of tweets reveals a political leader who revels in conflict, embraces self-promotion, and eschews normative standards of behavior. His tweets offer a microcosm of palpable tension ongoing in our culture between normative behavior and an all-out free-for-all online.

Limitations and Future Research

Like all studies, this one has limitations. While we explored a large corpus of tweets, our analysis ended with Trump’s inauguration. Future research should continue this work, examining his tweets throughout his presidency and comparing them to his earlier tweets. We also focused on the technology of populist performance and deliberation as theoretical frameworks. Fruitful research could employ other theoretical lenses, such as feminist theory and critical race theory to understand in different the discourses that emanate from Trump’s tweets.