Tweeting during presidential debates: effect on candidate evaluations and debate attitudes.

• Author: Houston, J. Brian

The last several U.S. presidential campaigns have each been marked by different, evolving communication technologies. The 2004 election has been referred to as the "internet election," best characterized by Howard Dean's innovative use of web technologies to organize campaign events, mobilize volunteers, and raise campaign funds during the Democratic presidential primary campaign (Trippi, 2004; Williams & Tedesco, 2006). The next presidential election in 2008 was characterized as the "Facebook" election, as the presidential candidates, particularly Barack Obama, frequently utilized Facebook to engage and connect with supporters and to coordinate campaign information and activities (Johnson & Perlmutter, 2010). The 2012 presidential election continued the trend of harnessing internet and social media technologies for political purposes, but the most recent election was specifically characterized by use of the social networking platform Twitter (McKinney, Houston, & Hawthorne, in press).

During the 2012 campaign, Twitter was used by political campaigns to disseminate information, organize events, and gauge public sentiment; by the news media to inform reporting and promote news content; and by the public to express political opinions and virtually watch political events with others. As a result of the prominence of Twitter during the 2012 election, as well as the variety of Twitter uses, users, and audiences, there is great need to understand the effect of this emerging technology on political outcomes. This study contributes to this area by examining how tweeting during a political event-specifically a 2012 general election presidential and vice presidential debate-affected attitudes about the political candidates participating in the event as well as perceptions of political debates.

TWITTER AND THE 2012 ELECTION

Twitter is a social media technology that allows registered users (also known as tweeters) to create a public or private profile, to post messages (also known as tweeting) of 140 or fewer characters, and to follow other users and see the messages (also known as tweets) posted by those users. Twitter was created in 2006 and currently has over 500,000,000 registered users worldwide (Lunden, 2012). Twitter users generate more than 500,000,000 tweets per day at a rate of approximately 5,800 tweets per second (Terdiman, 2012). During the 2012 election, political candidates, the news media, and the public increasingly took to Twitter to discuss the campaign. Campaign discussion on Twitter took a variety of forms, from candidates advocating for their own election or refuting opponent's attacks, to members of the news media reporting on campaign activities or engaging with their audience, to individual citizens praising, lamenting, mocking, or satirizing various political candidates or occurrences. Tweeted political statements and exchanges occurred throughout the campaign; however, political tweets tended to increase in frequency during major campaign events. For example, during Barack Obama's nomination acceptance speech there were 52,756 tweets per minute related to the event (Groom, 2012). Then during the first 2012 Obama-Romney presidential debate there were 158,690 tweets per minute related to the event. Finally, on election night at the time Obama's victory became apparent, there were 327,452 tweets per minute related to the event (Lee, 2012). At the time, each of these events represented a record for U.S. based politically oriented Twitter activity.

Given the increasing intersection of politics and Twitter, not only in the United States but globally, research has begun to examine the nature of political content on Twitter. For example, research has analyzed Twitter content and network structure in order to compare political/media elite tweets to citizen tweets (Ampofo, O'Loughlin, & Anstead, 2011; Hawthorne et al., in press), to understand what users tweet about during a political event (Anstead & O'Loughlin, 2011; Hawthorne et al., in press), to explore how politicians and governments use Twitter (Lassen & Brown, 2011; Waters & Williams, 2011), to determine if Twitter content can predict election results (Tumasjan, Sprenger, Sandner, & Welpe, 2011), to understand whether tweeters with different political ideologies interact on Twitter (Himelboim, McCreery, & Smith, 2013), and to explain how Twitter is used as part of political protest (Lotan et al., 2011; Tufekci &Wilson, 2012). Although understanding the nature of Twitter political content, networks, and uses is important, research examining the effects of tweeting about politics is also needed. A focus on tweeting effects guides the current project.

Specifically, we examine how tweeting about general election presidential and vice presidential debates influences attitudes about the political candidates and the debate itself. We focus on general election presidential and vice presidential debates for two reasons. First, general election debates are the single most watched political event during U.S. elections (McKinney & Carlin, 2004) and those who watch debates have been found to be better informed as a result of debate exposure (Benoit, Hansen, & Verser, 2003). Therefore, debates are an important component of the U.S. presidential campaign and viewing debates appears to benefit citizens. Second, and as discussed previously, major political events during the 2012 general election debates were associated with the heaviest political Twitter activity. Individuals who tweet during debates or other unfolding political events as they occur are participating in live-tweeting (Hawthorne et al., in press), wherein users watch the live broadcast of the event and tweet about it at the same time. Live-tweeting a political debate expands the debate viewing experience because in addition to watching a live broadcast of the debate, individuals are able to tweet their comments about what is happening; read other tweets from political campaigns and organizations, the media, and other citizens; respond to other tweets; forward tweets (also known as retweeting); and otherwise interact with other tweeters. Thus live-tweeting is a (virtual) form of "social watching" a political event (McKinney et al., in press). Because live-tweeting a general election debate is an emergent form of debate viewing, the effects of this process are unknown. The current project seeks to better understand these effects.

Overall then, the 2012 debates not only captured the attention of the U.S. public but also of the Twittersphere and thus, provide a good opportunity to examine the effects of tweeting during a political event. Moreover, live-tweeting a political event is an emergent phenomenon that is increasing in popularity and necessitates study. Our variables of interest in this study include candidate evaluations and debate attitudes. Examining the effects of media use on candidate evaluations is a mainstay of political campaign research (Pfau, Houston, & Semmler, 2007), and the opportunity to affect citizen vote choice is the ultimate reason for why candidates participate in debates. Debates have been found to be particularly effective at influencing the vote choice of undecided, less educated, and non-partisan citizens (McKinney& Carlin, 2004). Additionally, both news coverage of debates (Chaffee & Dennis, 1979; Lowry, Bridges, & Barefield, 1990) and interpersonal discussion of debates (Carlin & McKinney, 1994) have been found to contribute to debate viewing...