Aligning and Trespassing

Abstract

This study examined how Republican and Democratic candidates utilized Twitter to manage their impressions during the 2012 U.S. Senate elections, and examined their discussion of political issues and character traits across three types of tweets: campaign tweets, campaign-selected retweets, and tweets that mentioned their respective opponent. Candidates aligned with and trespassed party-based ownership of issues and traits across the tweet types. In the aggregate, Republicans discussed Republican-owned issues and traits more than Democrats, and Democrats emphasized Democrat-owned issues more than Republicans. This party alignment broke down when examined across winning and losing candidates, yielding varying routes to electoral success.

Keywords

impression management political ownership Twitter elections

Political elections produce a deluge of information. Campaigns and political action committees are pumping out information across every platform available about their favored candidate while trying to trounce the opposition. The news media moves from heavy speculation during the invisible primary and arcs increasingly toward feverishly reporting on every poll as Election Day nears. Constituents in heavily contested races and battleground states are inundated with information from campaigns and the news media, and citizens themselves add to the growing discourse by chiming in on social media, doorbelling for their preferred candidate, and bringing up politics, whether wanted or not, at the dinner table. And the exponential growth in use of social media by candidates, news media, and citizens has compounded this situation. For candidates, each new social media platform supplies them with a new way to reach a segment of the population, but it also burdens them with yet another platform to wrangle. The challenge for candidates, then, is to create a cohesive campaign image that is simultaneously pervasive and coherent across these platforms, while still taking advantage of the unique affordances of each campaign space. In today’s frantic stream of information and access, candidates’ need to manage impressions is at a seemingly all-time high.

This study explores how candidates are tackling this challenge by examining candidates’ campaign Twitter feeds across two important areas of impression management and formation: political issues and character traits. Although a number of studies have assessed candidates’ discussion of issues and traits before (e.g., Benoit & Hansen, 2002, 2004; Conway, Kenski, & Filer, 2014; Petrocik, Benoit, & Hansen, 2003; Rhea & Chattopadhyay, 2014), this study extends this line of work by examining how candidates utilize Twitter to manage their impressions in three ways: campaign tweets, retweets, and opponent mentions. Campaign tweets—tweets directly generated by the campaign—are the most straightforward assessment of a candidate’s impression management strategy and are akin to previous explorations of campaign-generated content, such as campaign websites and advertisements. Retweets—when a candidate reposts another user’s tweet—represent a blending of the old with the new. Surrogates are often employed to bolster a campaign’s communication strategy. These surrogates typically have a clear connection to the candidate, for example, they are family members, longtime friends, or fellow politicians. Candidates undoubtedly retweet these more traditional surrogates, but the interactivity of Twitter, combined with its ability to flatten hierarchical communication structures, means that candidates can easily retweet citizens and supporters. By bringing their voices into the campaign feed by retweeting them, these “everyday surrogates” become part of the campaign’s impression management efforts. Finally, in this information game, it is imperative to define your opponent. Elections are comparative, and impression management often means candidates need to define themselves and their opponent because such characterizations inevitably reverberate back to them and influence voters’ impression formation process. Therefore, assessing what issues and traits candidates discuss in tweets that mention their opponent illuminates an additional component of how candidates use Twitter to control their image. To explore these three types of impression management, this study analyzes 24 U.S. Senate candidates’ campaign Twitter feeds across 12 general election races during 2012. This examination builds upon the intersection of impression management and political ownership scholarship to explore how Republican and Democratic candidates traversed this dynamic platform via their issue and trait discussion.

Impression Management in Political Campaigns

Candidates do not haphazardly reveal information about themselves in a stream of consciousness; rather, they carefully craft a persona that attempts to elevate and polish their “everyday selves” into “electable selves.” This process is called impression management, and it occurs when individuals monitor what information they present to others in an “attempt to control the impressions others form of them” (Leary & Kowalski, 1990, p. 34; McGraw, 2003).

Impression management is often seen as synonymous with self-presentation, which Goffman (1959) defined as “all the activity of a given participant on a given occasion which serves to influence in any way any of the other participants” (p. 25). Candidates certainly strive to influence others, and scholars have provided several models for understanding impression management in politics, including Goffman’s (1959) “frontstage” and “backstage” distinctions between a person’s public and private selves, as well as Fenno’s (1978) “homestyle” and “Washington style” categories that represent politicians’ potentially differing self-presentations when they are with constituents versus when they are on Capitol Hill. This study focuses on Leary and Kowalski’s (1990) two-component model, which states that impression management is comprised of impression motivation and impression construction.

Everyone engages in some form of impression management on a daily basis, but our level of self-monitoring varies based on the situation. When among close friends, our focus on impression management is relatively low. But when people encounter moments of acute public awareness or intense scrutiny, our focus on our self-presentation rises. During an election, candidates are constantly exposed to intense scrutiny, and are motivated to impression-manage. A candidate’s motivation to impression-manage is further heightened by three other factors. First, Leary and Kowalski (1990) noted, “People are more motivated to impression-manage when the impressions they make are relevant to the fulfillment” of a goal (p. 38). These goals can include material goals, such as an employee trying to appear more competent in front of their boss so as to obtain a raise. For candidates, the ultimate material goal is being elected. As candidate communication influences voters’ knowledge about the candidates and which self-presentation they want to vote for (Sides, 2006), each performance matters and candidates are deeply motivated to impression-manage. Second, Leary and Kowalski (1990) noted that “the more public one’s behavior . . . the more motivated one will be to impression-manage” (p. 38). Candidates’ actions are very public during an election, and increasingly so with today’s 24/7 media landscape. Finally, the dependency on others for valued outcomes, for example, depending on citizens for their votes, and the high need for approval, means a person will be more likely to engage in impression management (Leary & Kowalski, 1990). Based on these factors, candidates are in a situation where they are highly motivated to control their image.

Once motivated, individuals move into impression construction. Impression construction involves selecting what image to create and deciding how to construct that image via various communicative means, for example, verbal content, body language, or appearance. Many candidates have employed certain campaign phrases and rhetoric to capture their image. For example, Senator Patty Murrary (D-WA) reappropriated the label “just a mom in tennis shoes” to showcase how she understood everyday citizens’ concerns because of her own life experiences. When candidates engage in impression construction, this process does not exist in a vacuum, but rather takes root in their self-concept because people often project a self-presentation that is in line with how they see themselves (Leary & Kowalski, 1990). The self-concept resists the idea that individuals put on a show and portray characteristics detached from who they are—often called “pretense.” The self-concept does not ignore that self-presentations are performative and strategic; rather, the idea is that in an effort to elevate certain aspects of one’s identity to achieve a goal, these “claimed images often mirror people’s self-concepts . . . [and] People hesitate to claim images that are inconsistent with how they see themselves . . .” (Leary & Kowalski, 1990, p. 40; Schlenker, 1980). As such, though deeply strategic, constructed impressions are tethered to who the person is.

This study examines how Republican and Democratic candidates construct their identity via their discussion of political issues and character traits. Issues and traits are crucial to the impression management process because they speak to who the candidate is—what legislation they will push for once in office and the character they will bring to that process—and as a result, they are key evaluative criteria for voters (Louden & McCauliff, 2004). Issues and traits are also important to impression management because they are associated with key identity characteristics of American political parties. Democrats and Republicans are perceived as “owning” certain issues and traits. The idea of political ownership emerges from work by Budge and Farlie (1983), and Petrocik (1996), which suggests that particular political parties are perceived to be better able to “handle” certain matters. “Handling” is the ability to address and resolve problems, and is derived from a “history of attention, initiative, and innovation toward these problems, which leads voters to believe that one of the parties (and its candidates) is more sincere and committed to doing something about them” (Petrocik, 1996, p. 826). Petrocik’s analysis of multiple national surveys showed that Democrats were seen as better at handling “compassion issues” and social welfare issues, such as improving education and health care, protecting social security, and helping the elderly. Republicans were perceived as more successful at handling issues like foreign affairs, national security, crime, and economic matters, such as taxes and reducing the deficit (Petrocik, 1996). A 2014 Pew survey reported similar trends in public perceptions: Republicans were perceived as more capable of handling the budget deficit, economy, and domestic terrorist threats, and Democrats were perceived as better at dealing with health care, education, and reproductive rights. It is important to note that ownership of an issue neither gives one party exclusive rights to campaigning on that issue nor suggests that other parties lack a history with the issue. Rather, ownership simply reflects the public’s perception of which party does a better job at handling an issue, and thus creates an image for voters of what issues they associate more strongly with a party. Based on these perceptions and associations, the remainder of this study will refer to issues and traits as “Democrat-owned” or “Republican-owned.”

Ownership also extends to character traits. Hayes (2005) noted, “Since traits are not directly observable, individuals generally infer attributes of another person’s personality from their actions and behavior” (p. 910). For example, if we witness a stranger help someone who is struggling to carry their groceries, we infer from this behavior that the stranger is a nice person. Based on this line of thought, Hayes (2005) created the theory of trait ownership, which posited that issues and behaviors could be the basis for trait assessments, creating a “carry-over effect.” The scenario plays out as follows: Candidates talk about party-owned issues, and then often campaign in locations that align with those issues. This issue-oriented discussion and action lead the public to attribute certain traits to a party. For example, Democrats own compassion issues that involve taking care of children, the elderly, and the environment. While campaigning on these issues, they may schedule campaign stops in elementary schools, assisted living facilities, or Sierra Club meetings to cement this association. This communication and campaigning help to cultivate a more nurturing image, and Democrats are perceived as owning compassion and friendliness. Republicans own the issues of military and economics and may campaign at defense contractors and business councils, thus portraying traits like toughness and leadership. These trait associations are enduring. Survey research stretching from the 1970s to 2014 has found that the public upholds these perceptions of parties (e.g., Hayes, 2011; Pew, 2014; Winter, 2010). Collectively, parties have distinct connections to issues and traits.

Candidates take note of this ownership and routinely emphasize their party’s issues and traits in their communication (Benoit & Hansen, 2002, 2004; Petrocik, 1996; Petrocik et al., 2003; Rhea & Chattopadhyay, 2014). Their motivation to do so is threefold. First, party membership is an acquired identity, and candidates choose their party affiliation based on their beliefs, on who they see themselves as. Therefore it makes sense that candidates would discuss issues and traits that are affiliated with their self-concept. Second, people impression-manage to be consistent with desired images, but they also aim to be inconsistent with undesired images (Leary & Kowalski, 1990). Scholarship has shown that candidates seek to highlight their “owned” issue strengths while skirting their opponent’s issue assets (Ansolabehere & Iyengar, 1994; Bélanger & Meguid, 2008; Petrocik, 1996). By emphasizing the issues and traits associated with their party, they are distancing themselves from the oppositional party and their related ownership image. Finally, candidates may favor party-owned issues and traits because it comes with rewards. Regarding impression management, Schlenker (1980) suggested that people are driven to maximize expected rewards and minimize expected punishments. In an election, candidates aim to be viewed favorably and get elected. And analysis shows that candidates who carry the party line receive these rewards. Benoit (2007) analyzed presidential TV spots from 1952 to 2004, and found that Democratic and Republican winners stressed party-owned issues more than Democratic and Republican losers. Winter’s (2010) analysis of American National Election Studies data between 1972 and 2004 also found that traits such as strength and confidence were roughly 7 times more likely to be mentioned as a reason to like Republicans than as a reason to like Democrats, and traits such as compassion and friendliness were almost 6 times more likely to be mentioned as a reason to like Democrats than as a reason to like Republicans. Consequently, Republican and Democratic candidates benefit when they portray self-presentations that align with their party’s perceived areas of ownership. Given these motivating factors, the first prediction is as follows:

H1: For campaign-generated tweets, Republicans will discuss Republican-owned issues and traits more than Democrats, and Democrats will discuss Democrat-owned issues and traits more than Republicans.

Impression management, in life and in politics, is not a one-person job. Rather, impression management often includes some collaboration. Goffman’s work on self-presentations is often associated with symbolic interactionism, which “claims that identity and self are constituted through constant interactions with others . . . Individuals work together to uphold preferred self-images of themselves and their conversation partners . . .” (Marwick & boyd, 2011, p. 10). Although this perspective lends itself more toward the maintenance of self-presentations in an interpersonal setting, the idea that impression management is a team effort is applicable to politics. Conway et al. (2014) noted, “In the political arena, campaign members, party members, and even family members contribute to the overall maintenance of candidate reputation” (p. 4). Each of these groups of people regularly acts as surrogates on behalf of the candidate, working throughout the campaign to build, maintain, and propagate the candidate’s image. Conway et al. analyzed tweets of candidates, campaigns, and surrogates, including spouses and party leaders, during the 2012 presidential election to examine impression management tactics involving trait discussions. They found that President Barack Obama’s network focused on discussion of positive traits to maintain Obama’s image, whereas Republican nominee Mitt Romney and his network emphasized more negative traits of Obama. Republicans and Democrats managed impressions differently.

Formal surrogates are invaluable to most campaigns, but the network of people engaged in impression management for a candidate stretches further than these allies and includes volunteers and supporters who actively work to fulfill the same goal as this close ring of surrogates. In most campaigns, the impression management work of these citizens is less visible to candidates because it occurs when they launch extensive canvassing activities or host debate watch parties that campaigns may know exist, but usually do not attend. Twitter, and other interactive social media, change this dynamic to some extent. On Twitter, individual citizens can directly @reply to candidates, retweet a campaign’s tweets, @mention the candidate in tweet, or incorporate a hashtag related to the campaign in a tweet. Campaigns can access this information and see what kind of impression management work supporters are doing on their behalf. Campaigns can also choose to amplify supporters’ impression management efforts by retweeting an individual’s tweet. This action further disseminates the individual’s tweet and enables the campaign to show that others embrace a favorable image of the candidate.

The question is then, what lies within these retweets? One could imagine that campaigns would retweet tweets that further endorse candidates’ claimed party images—Republicans would retweet content that associates them with Republican-owned matters, and Democrats would follow suit for Democrat-owned content. As such, campaign-generated tweets and campaign-selected retweets would align in their emphasis. However, it is also possible that retweets could diverge from campaign tweets, resulting in ownership trespassing. There are a few reasons for why campaigns may select this route. While parties are seen as owning certain issues and traits, and receive rewards when they fulfill these expectations, candidates are always trying to get more votes, and that means they need to appeal to as many voters as possible. By emphasizing party-owned elements, many candidates may do well with their base, but there are other factors at play. First, Republican-owned issues and traits, for example, the economy and strength, are typically perceived as more important than those owned by Democrats (Hayes, 2011; Meeks, 2012; Smith, Paul, & Paul, 2007). Democrats may want to compliment their focus on party-owned content with some references to Republican-owned content so as to appeal to the public’s interests. Alternately, voters also want candidates who demonstrate integrity and empathy, traits typically owned by Democrats (McGraw, 2003). Republicans may also want to incorporate these traits into their feeds. Campaigns may look to surrogates, via retweets, to do this ownership-trespassing work for them because by coming from someone besides the candidate, it may be perceived as more credible. Furthermore, impression management often focuses on maximizing rewards and minimizing punishments (Schlenker, 1980). If a candidate trespasses, it may come off as pretense and inauthentic, and thus may be punished. However, if a candidate retweets someone else saying it, they may be able to capture the rewards of stepping outside party lines and appealing to broader interests. For example, wives of male politicians are often tasked with showing the softer side of their husband because if the man did it himself, he may be accused of not being “manly enough.” To examine these possibilities, this study proposes two contrasting hypotheses for retweets:

H2: The ownership-alignment hypothesis mimics H1 and predicts that Republicans will discuss Republican-owned issues and traits more than Democrats, and Democrats will discuss Democrat-owned issues and traits more than Republicans.

H3: The ownership-trespassing hypothesis predicts that Republicans will discuss Democrat-owned issues and traits more than Democrats, and Democrats will discuss Republican-owned issues and traits more than Republicans.

Candidates’ efforts to define themselves often include trying to define their opponent. This effort was witnessed in Conway et al.’s (2014) analysis, which showed that candidates and their surrogates used positive trait mentions to characterize themselves and negative trait mentions to damage their opponents’ image. Furthermore, Benoit and Wells’s (1996) Functional Theory of Political Campaign Discourse argues that candidates use acclaim, attack, and defend to manage their and their opponent’s image, and they use these functions to discuss policy/issues and character traits. The current study takes a broader approach and examines what issues and traits are discussed in tweets that mention an opponent. Tweets that mention an opponent can take on several forms. They can be banal, for example, Candidates A and B will be in a debate tonight. They can also be more comparative in an attempt to create distinction between the two candidates’ images. In such tweets, it would seem likely that candidates would discuss owned issues and traits because candidates try to highlight their areas of strength in an attempt to increase their rewards. Such a strategy would help the candidate reinforce their image of ownership. However, the candidate could also try to attack the opponent on his or her turf by suggesting that the opponent is not competent on the elements they are perceived as owning in an attempt to damage the opponent’s image of ownership. Williams (2006) examined U.S. presidential candidates George W. Bush and John Kerry’s campaign emails in 2004 and found that in opponent-based issue framing, issue trespassing was present: Kerry focused on Bush’s failures regarding the Republican-owned issues of foreign policy and economy, and Bush emphasized Kerry’s shortcomings on the Democrat-owned issue of health care. However, Williams also found issue alignment in opponent-based emails from both candidates. As with retweets, this study offers two contrasting expectations for opponent mentions:

H4: The ownership-alignment hypothesis predicts that Republicans will discuss Republican-owned issues and traits more than Democrats, and Democrats will discuss Democrat-owned issues and traits more than Republicans.

H5: The ownership-trespassing hypothesis predicts that Republicans will discuss Democrat-owned issues and traits more than Democrats, and Democrats will discuss Republican-owned issues and traits more than Republicans.

Thus far, the conceptual framework has focused on what impression management strategy would appear to be most beneficial based on scholarship. To directly assess what strategies worked in 2012, and to determine whether winners and losers implemented an ownership-alignment or ownership-trespassing strategy, the research question asks the following:

RQ1: What issues and traits do Republican and Democratic electoral winners and losers incorporate in their tweets?

Whether a candidate is celebrating after Election Day depends on a number of factors, and as a content analysis, this study cannot claim causation. That said, candidate communication contributes greatly to electoral success, and with candidates’ goal of “staying on message” to create and maintain cohesive impressions, what a candidate says on Twitter is likely indicative of their broader campaign communication strategy. Therefore examining electorally successful and unsuccessful images may illuminate what strategies were used in 2012, and how these Twitter impression management strategies compare with previous scholarship.

Method

To examine these hypotheses and research questions, this study features a quantitative content analysis of U.S. Senate candidates’ Twitter feeds during the 2012 general elections. The need to construct effective online communication strategies is especially high in statewide elections given the geographic distribution of one’s constituency. By focusing on Senate elections, this study can examine how candidates utilized the reach and functionality of Twitter to present themselves in such contexts. In total, 24 candidates—12 Republicans and 12 Democrats—were examined across 12 elections, which is roughly a third of the 33 Senate elections held in 2012 (see Table 1).¹ This sample represented a variety of elections, including geographic diversity, Republican and Democratic winners, open versus closed elections, and all-male, all-female, and male versus female elections. These factors created a diverse cross-section of elections.

Table 1.

Selected Senate Elections and Number of Total Tweets.

	Republican	Democrat	Election dates
Arizona	Jeff Flake (n = 83)^a	Richard Carmona (n = 253)	August 29 to November 6
California	Elizabeth Emken (n = 2,347)	Dianne Feinstein (n = 52)^a,b	June 6 to November 6
Connecticut	Linda McMahon (n = 754)	Chris Murphy (n = 307)^a	August 15 to November 6
Hawaii	Linda Lingle (n = 2,303)	Mazie Hirono (n = 338)^a	August 12 to November 6
Indiana	Richard Mourdock (n = 1,101)	Joe Donnelly (n = 121)^a	May 9 to November 6
Massachusetts	Scott Brown (n = 170)^b	Elizabeth Warren (n = 337)^a	September 7 to November 6
Missouri	Todd Akin (n = 448)	Claire McCaskill (n = 157)^a,b	August 8 to November 6
Nebraska	Deb Fischer (n = 132)^a	Bob Kerrey (n = 866)	May 16 to November 6
Nevada	Dean Heller (n = 1,053)^a,b	Shelley Berkley (n = 669)	June 13 to November 6
New Mexico	Heather Wilson (n = 471)	Martin Heinrich (n = 454)^a	June 6 to November 6
New York	Wendy Long (n = 818)	Kirsten Gillibrand (n = 695)^a,b	June 27 to November 6
Pennsylvania	Tom Smith (n = 423)	Bob Casey Jr. (n = 310)^a,b	April 25 to November 6

Winner of the election.

Incumbent candidate in a closed race.

To analyze communication during these campaigns, each candidate’s campaign Twitter feed was downloaded after the election was completed via the official Twitter application programming interface (API) using a computer script that was verified by a professional software engineer.² The API limits access to 3,200 historical tweets per account, and this easily accommodated the time frame of the candidates’ general elections. To ensure that the script collected all of the tweets, a manual check was conducted on a subset of the candidates in which tweets from the download were periodically compared with tweets on the candidate’s actual Twitter page. The manual check revealed a complete match between the two sources. Each Twitter account was verified by matching them against the accounts specified on the candidate’s campaign website. For each candidate, the entire census of general election tweets was coded, ranging from the day after the appropriate primary election to Election Day on November 6, 2012. In total, there were 14,662 tweets, and Republicans (n = 10,103) tweeted more than twice as much as Democrats (n = 4,559).

For the analysis, the following variables and operationalizations were created. The unit of observation and analysis was the individual tweet. Dimensions of each composite variable were coded as “Absent” (0) or “Present” (1), and then collapsed so that “0” indicated complete absence of the variable and “1” indicated that at least one dimension of the variable was present. Political issues and character traits were categorized as “Democrat-owned” or “Republican-owned” based on previous studies (e.g., Petrocik, 1996; Pew, 2014; Winter, 2010). Democrat-owned issues were coded as present when there was any discussion of any of the following dimensions included in an individual tweet: education, health care, environment, or “women’s issues,” which included issues that are typically seen as more important to women, for example, equal pay, abortion, and contraception. Republican-owned issues were coded as present when there was any discussion of military/national security, crime, economy, or taxes. Democrat-owned traits were coded as present when there was any discussion of compassion, honesty, congeniality, or collaboration. Republican-owned traits were coded as present when there was any discussion of leadership, strength, confidence, or decisiveness.

There were three types of tweets: campaign-generated, retweets, or opponent mentions. Campaign-generated tweets were the baseline of categorization and included any tweets that were not subsequently coded as a retweet or opponent mention. Campaign-selected retweets were coded as present when a candidate reposted another user’s tweet and “RT” or “MT” appeared at the beginning of the tweet. MT stands for “modified tweet,” and occurs when a user reposts another user’s tweet, but has modified the original text in some way and places “MT” at the beginning of the tweet. Retweets and modified tweets that came from an account affiliated with the campaign were coded as campaign-generated tweets, not campaign-selected retweets. Accounts affiliated with the campaign included a candidate’s in-office account, for example, the account of a sitting Senator running for reelection, or specialized campaign accounts, for example, the @BigBlueMO2012 account, which featured tweets while Claire McCaskill was touring Missouri in a campaign bus. Opponent mentions were coded as present when a candidate’s tweet referenced his or her electoral opponent in the tweet via his or her @username, a variation of the given name, or the content of the tweet references the opponent but does not use the name and says something to the effect of “my opponent.” For example, Heather Wilson (R-NM) mentioned her opponent, Martin Heinrich, when she tweeted: “#FLASHBACK #FACTS @Heinrich4NM said he was proud to vote for cap and trade, which would kill 11,000 NM jobs. #nmsen #debate.” Furthermore, if a candidate @replied to their opponent, the tweet was categorized as an opponent mention. All other @replies to users who were not opponents, and were not retweets, were categorized as campaign-generated tweets.

The coding process was manual and was as follows. First, the codebook was constructed and then employed in two pilot tests, with each test including the coding of 100 tweets, for a total of 200 tweets. The codebook was revised based on these pilot tests. Second, training for two coders included first reviewing the codebook, and then independently applying the codebook to 100 tweets in a practice round of coding. Revisions were made to the codebook based on this practice round. Third, a formal test of intercoder reliability was conducted on a random sample (n = 735) by the two coders. The size of the random sample met and exceeded the minimum required based on Riffe, Lacy, and Fico’s (1998) equation for selecting an appropriate size. Calculated using Krippendorff’s alpha, coefficients met appropriate acceptance levels (Krippendorff, 2004): retweets (α = .99), opponent mentions (α = .92), Democrat-owned issues (α = .86), Republican-owned issues (α = .83), Democrat-owned traits (α = .83), and Republican-owned traits (α = .81).

Except where noted, the hypotheses and research questions were investigated using descriptives to obtain the percentages for each comparison. Inferential statistics are not necessary because analysis included the entire census of tweets for each candidate’s general election.

Results

Before delving into the results, it is helpful to examine some general contours of the data. Of the 14,662 total tweets, 8,870 (60.5% of the total tweets) were campaign-generated and did not reference the opponent or feature a campaign-selected retweet. Of these tweets, 15.4% discussed Republican-owned issues, 7.7% discussed Democrat-owned issues, 7.3% discussed Republican-owned traits, and 4.9% discussed Democrat-owned traits. There are two notable findings within these percentages. First, candidates were more likely to discuss Republican-owned issues and traits than Democrat-owned issues and traits. Second, based on the percentages, it appears as though candidates spent the majority of their time discussing matters besides issues or traits. As shown later, this heavy focus on matters besides issues and traits is also present in retweets and opponent mentions. Although not categorized for the sake of this study, these other tweets typically included updates on campaign stops, links to a candidate’s latest advertisement or a news article on the candidate, @replies to users thanking them for their support, or personal information on the candidate. For example, Linda McMahon (R-CT) tweeted a campaign update: “Had a great time today saying thanks to supporters in Danbury who mobilized voters on primary day! #ctsen.” Kirsten Gillibrand (D-NY) tweeted a photo with her sons and sent holiday greetings: “Enjoying time with the boys today. >From my family to yours, wishing you a happy & safe #4thofJuly!” This heavy focus on topics besides issues and traits has also been found in other Twitter studies of politicians. Golbeck, Grimes, and Rogers (2010) categorized 27% of tweets as updates on location and activities; Parmelee and Bichard (2012) categorized more than 40% of tweets as campaign trail updates; Graham, Jackson, and Broersma (2014) found that roughly 40% of tweets featured campaign trail updates or campaign promotion; and Evans, Cordova, and Sipole (2014) found that nearly 25% of tweets were about campaign stops, almost 30% were personal information, and just more than 10% focused on issues.

Given this focus of the tweets, the remaining analysis only focused on tweets that mentioned at least one issue or trait, which allowed the analysis to zero in on the tweets most relevant to this study. The first hypothesis predicted that for campaign-generated tweets, Republicans would discuss Republican-owned content more than Democrats, and Democrats would discuss Democrat-owned content more than Republicans (H1). Of the 8,870 campaign tweets, 1,882 (21.2%) mentioned an issue. Within these issue-oriented campaign tweets, the results were mixed for H1. Starting with results on the left side of Table 2 for issues, Republicans were slightly more likely than Democrats to discuss Republican-owned issues: 73.9% of Republicans’ issue-oriented tweets versus 71% of Democrats’ issue-oriented tweets. Furthermore, Democrats (40.1%) were more likely than Republicans (33.6%) to discuss Democrat-owned issues. Campaign tweets that mentioned a trait (n = 1,045, 11.8%) occurred half as much as those mentioning an issue. And within these trait-oriented tweets, there was no party alignment: Democrats (70.1%) were more likely to discuss Republican-owned traits than Republicans (57%), and Republicans (47.7%) were more likely to discuss Democrat-owned traits than Democrats (33.6%). Ultimately, H1 was only partially supported: Republicans and Democrats discussed party-owned issues more than the opposing party, but they crossed party lines on trait discussion.

Table 2.

Percentage of Issue- or Trait-Oriented Tweets for Republican and Democratic Candidates.

	Campaign generated		Retweets		Opponent mentions
	Republicans	Democrats	Republicans	Democrats	Republicans	Democrats
	Issues n = 1,116	Issues n = 766	Issues n = 243	Issues n = 63	Issues n = 1,042	Issues n = 515
	Traits n = 623	Traits n = 422	Traits n = 266	Traits n = 51	Traits n = 502	Traits n = 181
Republican-owned issues	73.9%	71.0%	74.5%	55.6%	78.8%	67.8%
Democrat-owned issues	33.6%	40.1%	32.5%	47.6%	28.3%	47.4%
Republican-owned traits	57.0%	70.1%	72.6%	68.6%	54.4%	51.4%
Democrat-owned traits	47.7%	33.6%	33.5%	33.3%	54.2%	54.1%

Note. n values represent the number of tweets that discussed an issue or trait. Percentages represent the proportion of issue- or trait-oriented tweets that discussed a Republican- or Democrat-owned issue or trait.

The next set of hypotheses examined retweets and pitted the ownership-alignment (H2) and ownership-trespassing (H3) perspectives against one another. Of the 14,662 total tweets, 2,707 (18.5% of total tweets) were campaign-selected retweets. Of these retweets, 8.4% discussed Republican-owned traits, 8.0% discussed Republican-owned issues, 4.0% discussed Democrat-owned issues, and 3.9% discussed Democrat-owned traits. Once again, Republican-owned matters predominated issue and trait discussions. Results in the middle two columns in Table 2 directly assess H2 and H3, and present the proportion of issue- or trait-oriented retweets that included a Republican- or Democrat-owned issue or trait. Some of the patterns from H1 were mimicked in this analysis. Of the 2,707 retweets, 306 (11.3%) mentioned an issue. Within these issue-oriented retweets, Republicans (74.5%) retweeted discussion of Republican-owned issues more than Democrats (55.6%), and Democrats (47.6%) retweeted discussion of Democrat-owned issues more than Republicans (32.5%). Results were somewhat different for retweets that mentioned a trait (n = 317, 11.7%). Within these trait-oriented retweets, Republicans (72.6%) highlighted Republican-owned traits slightly more than Democrats (68.6%) in retweets, and Republicans and Democrats were similar in their retweeted discussion of Democrat-owned traits (33.5% vs. 33.3%). In sum, Republicans and Democrats emphasized party ownership of issues in retweets, Republicans emphasized Republican-owned traits more, and all candidates were similar on Democrat-owned traits. In turn, the ownership-alignment perceptive received more support than the ownership-trespassing perspective for retweets.

The next set of analysis delved into tweets that referenced a candidate’s opponent. Of the 14,662 total tweets, 4,031 (27.5% of total tweets) mentioned the candidate’s respective opponent. Of these tweets, 29% discussed Republican-owned issues, 13.4% discussed Democrat-owned issues, 9.2% discussed Democrat-owned traits, and 9.1% discussed Republican-owned traits. Overall, issue and trait discussion was higher in this category as compared with the previous two categories. And while Republican-owned issues more than doubled that of Democrat-owned issues, for the first time, discussion of traits was similar across both sets of party-owned traits. To further examine these opponent-oriented tweets, the final set of hypotheses tested the ownership-alignment (H4) versus ownership-trespassing (H5) perspectives. Results in the two rightmost columns of Table 2 show that candidates once again aligned with their party on issues. Of the 4,031 tweets that referenced an opponent, 1,557 (38.6%) mentioned an issue. Within these issue-oriented opponent mentions, Republicans (78.8%) discussed Republican-owned issues more than Democrats (67.8%), and Democrats (47.4%) discussed Democrat-owned issues more than Republicans (28.3%). Discussion of traits when mentioning the opponent (n = 683, 16.9%) produced somewhat different results. Within these trait-oriented opponent mentions, Republicans (54.4%) discussed Republican-owned traits slightly more than Democrats (51.4%), and Republicans and Democrats were nearly identical in their discussion of Democrat-owned traits (54.2% vs. 54.1%). Collectively, Republicans and Democrats emphasized party ownership of issues when mentioning the opponent, and Republicans focused slightly more on owned traits. Therefore, the ownership-alignment perspective received full support regarding discussion of issues, and partial support regarding traits.

The final set of analysis examined what impression management strategies were more electorally successful in 2012, and was guided by RQ1: What issues and traits do Republican and Democratic electoral winners and losers incorporate in their tweets? Tables 3 and 4 provide the analysis for Republican and Democratic candidates by electoral success. Before examining the results of Table 3, it is important to take into consideration some differences between winning and losing candidates. In 2012, Democrats won 23 of the 33 Senate elections, Republicans won 8, and Independents won 2. Republican Senate candidates, then, won roughly a quarter of the races. This sample reflects those conditions and includes three winning Republicans. This means that the data for winning Republicans represent three candidates, and for losing Republicans, they represent nine candidates. At times, this can create a large discrepancy in tweet volume. Using percentages enables a comparison of the proportions regardless of the number of candidates or tweets in each category. It is also important to keep this consideration in mind when examining data for Democrats in Table 4.

Table 3.

Percentage of Issue- or Trait-Oriented Tweets for Republican Candidates.

	Campaign generated		Retweets		Opponent mentions
	Winners	Losers	Winners	Losers	Winners	Losers
	Issues n = 217	Issues n = 899	Issues n = 4	Issues n = 239	Issues n = 133	Issues n = 909
	Traits n = 91	Traits n = 532	Traits n = 5	Traits n = 261	Traits n = 25	Traits n = 477
Republican-owned issues	84.3%	71.4%	75.0%	74.5%	71.4%	79.8%
Democrat-owned issues	20.3%	36.8%	25.0%	32.6%	36.1%	27.2%
Republican-owned traits	50.5%	58.1%	60.0%	72.8%	36.0%	55.3%
Democrat-owned traits	53.8%	46.6%	40.0%	33.3%	64.0%	53.7%

Table 4.

Percentage of Issue- or Trait-Oriented Tweets for Democratic Candidates.

	Campaign generated		Retweets		Opponent mentions
	Winners	Losers	Winners	Losers	Winners	Losers
	Issues n = 488	Issues n = 278	Issues n = 35	Issues n = 28	Issues n = 329	Issues n = 186
	Traits n = 242	Traits n = 180	Traits n = 30	Traits n = 21	Traits n = 120	Traits n = 61
Republican-owned issues	70.5%	71.9%	62.9%	46.4%	72.9%	58.6%
Democrat-owned issues	39.8%	40.6%	40.0%	57.1%	45.0%	51.6%
Republican-owned traits	75.2%	63.3%	70.0%	66.7%	50.8%	52.5%
Democrat-owned traits	28.1%	41.1%	33.3%	33.3%	56.7%	49.2%

Across the three tweet types in Table 3, winning Republicans were less likely than losing Republicans to discuss Republican-owned issues and traits in four of six comparisons, and winning and losing Republicans were similar in their discussion of owned issues in the retweet category. In turn, winning Republicans only dominated losing Republicans in one comparison of owned content: Republican-owned issues in campaign-generated tweets. Conversely, winning Republicans discussed Democrat-owned issues and traits more than losing Republicans in four of six comparisons—supporting an ownership-trespassing perspective. Comparatively, then, winning Republicans favored Democrat-owned content more than losing Republicans, and losing Republicans favored Republican-owned content more than winning Republicans.

It is also beneficial to compare winning Republicans against themselves across tweet types. In every tweet type, winning Republicans filled their issue-oriented feed with a heavy focus on Republican-owned issues. Winning Republicans’ ratio of owned-to-unowned issue discussion ranged from 2:1 in opponent mentions to 4:1 in campaign-generated tweets. Trait discussion was a slightly different story. Winning Republicans were equitable in their discussion of Republican- and Democrat-owned traits in campaign-generated tweets (50.5% vs. 53.8%). They then shifted to a heavier focus on Republican-owned traits in retweets, and then a heavier focus on Democrat-owned traits in opponent mentions. In turn, their issue-oriented construction was similar across the board and supported the ownership-alignment perspective, but their trait-oriented construction differed based on the tweet type.

Across the tweet types in Table 4, winning Democrats were equally or less likely than losing Democrats to discuss owned content in five of six comparisons. The only time winning Democrats dominated the discussion on an owned content item was regarding Democrat-owned traits in opponent mentions. Conversely, winning Democrats were equally or more likely than losing Democrats to discuss unowned content in all six comparisons—supporting an ownership-trespassing perspective. Comparatively, winning Democrats were more likely than losing Democrats to fill their feed with discussion of Republican-owned content, and losing Democrats were more likely than winning Democrats to emphasize Democrat-owned content.

As with Republicans, it is also valuable to compare winning Democrats against themselves across tweet types. Winning Democrats emphasized Republican-owned content over Democrat-owned content in five of six comparisons. The only time winning Democrats shifted their emphasis to Democrat-owned content over Republican-owned content was regarding traits in opponent mentions. Therefore, winning Democrats maintained a similar impression construction strategy across the three tweet types and favored an ownership-trespassing strategy.

To further examine how winning and losing Republicans and Democrats differed in how they discussed issues and traits, Table 5 presents the results of two binomial logistic regressions that used the issue and trait categories across the three tweet types to predict electoral success (loss = 0, win = 1) for Republicans and Democrats separately. A candidate’s gender and incumbency were included as control variables because previous studies have found that both of these identity aspects can affect a candidate’s tweeting strategies (Evans et al., 2014; Parmelee & Bichard, 2012).

Table 5.

Binary Logistic Regression of Electoral Success.

	Republicans	Democrats
	n = 3,182	n = 1,706
	B (SE)	B (SE)
Campaign generated
Republican issues	0.081 (.318)	−0.453 (.174)**
Democratic issues	0.506 (.335)	−0.459 (.179)*
Republican traits	0.962 (.339)**	−0.601 (.185)**
Democratic traits	0.302 (.344)	−0.287 (.206)
Retweets
Republican issues	0.878 (.805)	−0.048 (.439)
Democratic issues	−0.649 (.898)	0.521 (.437)
Republican traits	0.058 (.772)	−0.240 (.442)
Democratic traits	−0.577 (.888)	−1.043 (.432)*
Opponent mentions
Republican issues	−0.250 (.382)	−0.411 (.196)*
Democratic issues	−0.290 (.388)	0.131 (.192)
Republican traits	0.165 (.505)	−0.295 (.225)
Democratic traits	0.642 (.427)	−0.023 (.240)
Pseudo-R²	.761	.474
Constant	−5.241 (.802)***	1.737 (.372)***

Note. n values represent number of tweets that discussed an issue or trait. Electoral success was coded as loss = 0, win = 1. Control variables not shown but included in the analysis: Candidate gender and incumbency.

p < .05. **p < .01. ***p < .001.

According to the results for Republicans in Table 5, only one variable was a significant predictor of electoral success. Republican candidates who emphasized Republican-owned traits in their campaign-generated tweets were more likely to win. Notably, one of the control variables produced significant results: Republicans were more likely to win if they were incumbents (B = 6.036, SE = .303, p < .001). For Democrats, there were five significant predictors that varied between alignment and trespassing. Democrats who focused on Republican-owned issues and traits, as well as Democrat-owned issues, in campaign-generated tweets were less likely to win. Also, Democrats who retweeted others regarding Democrat-owned traits were less likely to win, and Democrats who tried to cross party lines and discuss Republican-owned issues when tweeting about their opponent were less likely to win. Finally, one of the control variables produced significant results: Democrats were less likely to win if they were women (B = −0.765, SE = .131, p < .001).

Discussion

Everyone engages in some form of impression management, and political candidates are certainly no exception. In fact, candidates may be more likely than most to impression-manage because they undergo close public scrutiny in an effort to win an election. To better understand how Republican and Democratic candidates managed their impressions on Twitter, this study built upon impression management and political ownership literature, and analyzed the campaign Twitter feeds of 24 candidates across 12 U.S. Senate elections in 2012. Specifically, this analysis examined candidates’ discussion of political issues and character traits across three types of tweets. Results indicate that candidates aligned with and diverged from party-based ownership of issues and traits, and these routes corresponded with varying levels of electoral success. Ultimately, this analysis yielded three overarching patterns.

First, Republicans and Democrats maintained fairly consistent proportions of discussion across the tweet types. In other words, their impression management across campaign-generated tweets, campaign-selected retweets, and opponent mentions converged. Therefore, when they decided to raise the voice of another Twitter user via a retweet or discuss their opponent, they typically approximated the impression management strategy found in their campaign tweets, for example, emphasizing the same proportion of issue and trait discussion. As discussed previously, candidates on either side of the aisle have some incentive to present themselves to some extent as capable of handling the other party’s matters. The functionality of retweets enables campaigns to highlight other aspects of the candidate, perhaps unowned aspects, at a distance by having someone else showcase their ability. As trespassing can result in punishment, candidates can have others do this work for them and either receive the rewards of being a more well-rounded candidate or offload the punishment to the source of the retweet if the public sees this line of impression management as pretense. Candidates, in this study, did not discuss unowned matters more in retweets and thus did not appear to capitalize on this unique Twitter functionality.

Second, in the aggregate, Republican and Democratic candidates constructed impressions that aligned with their political parties across the three tweet types. These findings provide partial support for H1, H2, and H4, and foster support for the ownership-alignment perspective. Republicans stressed Republican-owned issues more than Democrats across all tweet types, and Republicans emphasized Republican-owned traits more than Democrats in retweets and opponent mentions. For example, Republican Dean Heller painted a positive image of his ability to manage the Republican-owned issue of economy and created a negative image of Democrat Shelley Berkley when he tweeted: “When Nevada was losing jobs, @DeanHeller voted against bailouts; @Berkley4Senate voted to bailout Wall St. #HomeMeansNevada.” Democrats also stressed Democrat-owned issues more than Republicans across all tweet types. This pattern upholds that of previous ownership studies of Republican and Democratic candidates (e.g., Benoit & Hansen, 2002, 2004; Petrocik, 1996; Petrocik et al., 2003; Rhea & Chattopadhyay, 2014). This pattern is also consistent with scholarship on impression construction and political ownership: Candidates most likely aligned with their party to be consistent with desired images of party ownership and to be inconsistent with undesired images of the opposition party’s ownership areas.

The third pattern, however, prompts a more nuanced discussion of desired and undesired images across party. RQ1 sought to examine what issues and traits Republican and Democratic electoral winners and losers incorporated in their tweets. Results indicated that losing candidates were more likely to stress owned content as compared with winning candidates, and winning candidates were more likely to stress unowned content as compared with losing candidates. Therefore, these results supported the ownership-trespassing perspective and contrasted Benoit’s (2007) study, which found that Republican and Democratic winners stressed party-owned content more than Republican and Democratic losers. This finding may be attributable to a few factors, each of which will be discussed by party.

Across most of the analysis, winning Democrats were more likely to trespass. Winning Democrats discussed unowned content more than or similarly to Democratic losers in six of six comparisons, and Democratic losers tweeted about owned content more than or similarly to Democratic winners in five of six comparisons. The regression analysis produced mixed results, with Democrats being less likely to win if they discussed owned and unowned content in some tweet types. But when the analysis focused solely on winning Democrats, winning Democrats favored unowned content over owned content in five of six comparisons. Winning Democrats may have been motivated to trespass for two reasons. First, winning Democrats may have stressed the Republican-owned issues of the economy and taxes more because the U.S. economy in 2012 was improving. Vavreck (2009) asserted that the health of the economy acts as the context in which campaigns perform. For example, if the economy is experiencing growth, candidates can clarify their role in fostering these good conditions. According to Vavreck’s analysis of presidential elections, pursuing this route can be advantageous for candidates. This strategy is particularly good if the candidates are incumbents, but it is also possible that it applies to Senate candidates with the same party affiliation as the current administration because they can claim that their party brought about these prosperous conditions. Therefore, it is plausible that Democrats who took advantage of their party’s attachment to a good economic climate and trespassed party lines to construct economy-focused impressions were more likely to win, whereas Democrats who played down these issues because they are unowned were more likely to lose.

Second, Democrats could have been trying to seize their opponent’s ownership of these issues because Republican-owned issues and traits are seen as more important than Democrat-owned issues and traits (Hayes, 2011; Meeks, 2012, Smith et al., 2007). Perhaps Democrats were angling their impression management more toward favoring public perceptions that privilege Republican issues and traits and, comparatively speaking, shying away from discussing owned matters that are perceived as less important. For example, Democrat Kirsten Gillibrand, who defeated Republican Wendy Long, tweeted about a Republican-owned issue: “Proud to serve on Armed Services Cmte. One reason I came to Congress was to strengthen natl security & be a voice for troops & mil families.” Impression management and political ownership literature suggests that candidates favor owned issues and skirt opponents’ issues as a way to maximize rewards and minimize punishment. It is possible that Democrats view issue trespassing as a way to accomplish both. If the public perceives Republican-owned issues as more important, then Democrats may reap certain electoral rewards by focusing on such important issues. Furthermore, if a Democrat is seen as focusing too much on less important, Democrat-owned issues, and not enough on more important, Republican-owned issues, then the public may punish them for not prioritizing “correctly.” They can avoid this potential punishment by focusing proportionally more on Republican-owned issues. Ultimately, Democrats running for U.S. Senate may need to cross ownership lines to address matters of perceived national import.

Results for Republicans were less clear-cut. First, when the analysis compared winning and losing Republicans, the results supported an ownership-trespassing perspective. Republican winners were more likely than Republican losers to discuss unowned content in four of six comparisons, and Republican losers were more likely or as likely to discuss owned content as compared with Republican winners in five of six comparisons. These findings suggest that the ownership-trespassing perspective is a more electorally successful route for Republicans. Regression results did little to dissuade this pattern because only one predictor was significant. Other work has found trespassing to be beneficial. Hayes (2005) examined American National Election Studies data for presidential candidates between 1980 and 2004, and found that for respondents who rated both candidates equally on traits, “The likelihood of a Republican vote is much higher when the candidates tie on Democratic traits than when they tie on Republican ones” (p. 919). For example, when Republicans tied Democrats on Republican-owned traits, they received 41% to 45% of the vote, but when Republicans tied Democrats on Democrat-owned traits, they received 61% to 65%. Trespassing on traits can lead to a bump in votes. Trespassing on issues may have also been beneficial in 2012 based on Vavreck’s (2009) perception of the economy as an evaluative context. As the economy under a Democratic administration was improving in 2012, winning Republicans, as compared with losing Republicans, may have placed less emphasis on this owned issue so as to not showcase a positive attribute of the Democratic Party. It is possible that effective trespassing is context-dependent. Second, when comparing winning Republicans with themselves, the results supported an ownership-alignment perspective. Winning Republicans favored Republican issues over Democratic issues across all tweet types. This mix of results based on the analytic approach may suggest that a mixed approach is best for Republicans. If Republicans want to win, they may need to align and trespass. Future research is needed to better understand what interplay of alignment and trespassing works best in different contexts.

This study’s findings are not without limitations. First, the results represent 12 Senate general elections in 2012. It is possible that these results are not indicative of other Senate elections or different time periods of the campaign. In particular, Republicans’ relatively low win rate in 2012 prompts the need for future research to examine other election cycles. Second, this study captured tweets after the election ended, and did not include any tweets deleted by the candidates/campaigns during the course of the election. It is possible that the content of those tweets differed from those the candidates did not delete.³ Third, this study focused on a candidate’s party identification, but beyond adding gender and incumbency as controls to the regressions, it did not thoroughly consider other aspects of a candidate’s identity that could have affected his or her impression motivation and construction processes. Other studies have examined how other aspects of a candidate’s identity have affected his or her tweeting strategies, including gender, incumbency, age, and competitiveness of the race (Conway et al., 2014; Evans et al., 2014; Parmelee & Bichard, 2012), as well as how the context of the campaign affected issue emphases (Vavreck, 2009). This work has not examined these differences across this study’s three tweet types, which presents a rich venue for further research. Fourth, much of the ownership literature focuses on presidential elections, which also means the literature is focused on all-male elections with candidates seeking to impression-manage on a national scale with an increasingly polarized public. This study’s elections included female candidates and focused on impression management at the statewide level. Either of these factors may mitigate candidates’ decision to embrace party-aligned images. Future work should examine impression management and ownership at varying levels of office and for male and female candidates. Finally, the differences in this study could be attributable to the medium. Twitter became a prime campaigning space in America in 2012, and it is possible that candidates were using this emerging platform to create impressions that diverged with scholarship based on more established campaigning platforms. Future research could compare candidates’ discussion of issues and traits across traditional and emerging campaigning platforms.

The public rarely gets to meet candidates and engage in interpersonal conversations with them. Instead, most of the public relies on campaign communication and news coverage to guide their impression formation process. In turn, how a candidate constructs and manages their image, and the mediated versions of their image, plays a crucial role in whether they are celebrating or commiserating on election night. With Twitter, candidates have a high degree of control over their impression management because this platform establishes a direct line of contact with voters that bypasses news gatekeepers. Even with media gatekeepers, candidates’ impression management on Twitter does exert some control over the news. Parmelee (2014) interviewed political journalists during the 2012 elections and found that tweets from political leaders contributed to first- and second-level agenda building for the journalists. These factors highlight how candidates can construct their images and reach voters directly via their tweets and indirectly via the news media’s incorporation of their tweets. Given this interplay, our understanding of modern impression management in campaigns would greatly benefit from continued work on how candidates navigate emerging platforms to manage impressions generally, and based on the affordances of these platforms, as well as the effects of this management on the news and voters.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biography

Lindsey Meeks is an assistant professor in the Department of Communication at the University of Oklahoma, with a research focus on gender, political communication, and media.

References

Ansolabehere

Iyengar

(1994). Riding the wave and claiming ownership over issues: The joint effects of advertising and news coverage in campaigns. Public Opinion Quarterly, 58, 335-357. doi:10.1086/269431

Bélanger

É.

Meguid

B. M.

(2008). Issue salience, issue ownership, and issue-based vote choice. Electoral Studies, 27, 477-491. doi:10.1016/j.electstud.2008.01.001

Benoit

W. L.

(2007). Own party issue ownership emphasis in presidential television spots. Communication Reports, 20, 42-50. doi:10.1080/08934210601182818

Benoit

W. L.

Hansen

(2002). Issue adaptation of presidential television spots and debates to primary and general audiences. Communication Research Reports, 19, 138-145. doi:10.1080/08824090209384841

Benoit

W. L.

Hansen

(2004). Issue ownership in primary and general presidential debates. Argumentation and Advocacy, 40, 143-154. Retrieved from http://www.americanforensics.org/AA/

Benoit

W. L.

Wells

(1996). Candidates in conflict: Persuasive attack and defense in the 1992 presidential debates. Tuscaloosa: University of Alabama Press.

Budge

Farlie

(1983). Explaining and predicting elections: Issue effects and party strategies in twenty-three democracies. London, England: Allen & Unwin.

Conway

Kenski

Filer

(2014, May). Impression management on Twitter: Strategic use by candidates and their campaigns. Paper presented at annual National Communication Association convention, Chicago, IL.

Dolan

(2014). Gender stereotypes, candidate evaluations, and voting for women candidates: What really matters? Political Research Quarterly, 67, 96-107. doi:10.1177/1065912913487949

10.

Evans

H. K.

Cordova

Sipole

(2014). Twitter style: An analysis of how house candidates used Twitter in their 2012 campaigns. Political Science & Politics, 47, 454-462. doi:10.1017/S1049096514000389

11.

Fenno

R. F.

(1978). Home style: House members in their districts. Boston, MA: Little, Brown.

12.

Goffman

(1959). The presentation of self in everyday life. New York, NY: The Overlook Press.

13.

Golbeck

Grimes

Rogers

(2010). Twitter use by the U.S. Congress. Journal of the American Society for Information Science and Technology, 61, 1612-1621. doi:10.1002/asi.21344

14.

Graham

Jackson

Broersma

(2014). New platform, old habits? Candidates’ use of Twitter during the 2010 British and Dutch general election campaigns. New Media & Society. Advance online publication. doi:10.1177/1461444814546728

15.

Hayes

(2005). Candidate qualities through a partisan lens: A theory of trait ownership. American Journal of Political Science, 49, 908-923. doi:10.1111/j.1540-5907.2005.00163.x

16.

Hayes

(2011). When gender and party collide: Stereotyping in candidate trait attribution. Politics & Gender, 7, 133-165. doi:10.1017/S1743923X11000055

17.

King

Matland

(2003). Sex and the grand old party: An experimental investigation of the effect of candidate sex on support for a Republican candidate. American Politics Research, 31, 595-612. doi:10.1177/1532673X03255286

18.

Krippendorff

(2004). Content analysis: An introduction to its methodology. Thousand Oaks, CA: SAGE.

19.

Leary

M. R.

Kowalski

R. M.

(1990). Impression management: A literature review and two-component model. Psychological bulletin, 107, 34-47. doi:10.1037/0033-2909.107.1.34

20.

Louden

McCauliff

(2004). The “authentic candidate”: Extending candidate image assessment. In Hacker

(Ed.), Presidential candidate images (pp. 85-103). Lanham, MD: Rowman & Littlefield.

21.

Marwick

A. E.

boyd

(2011). I tweet honestly, I tweet passionately: Twitter users, context collapse, and the imagined audience. New Media & Society, 13, 114-133. doi:10.1177/1461444810365313

22.

McGraw

(2003). Political impressions: Formation and management. In Sears

D. O.

Huddy

Jervis

(Eds.), Oxford handbook of political psychology (pp. 394-432). Oxford, UK: Oxford University Press.

23.

Meeks

(2012, May). Meeting expectations: Issues, traits, party, and gender in a woman-versus-woman election. Paper presented at the annual meeting of the American Association of Public Opinion Research, Orlando, FL.

24.

Parmelee

J. H.

(2014). The agenda-building function of political tweets. New Media & Society, 16, 434-450. doi:10.1177/1461444813487955

25.

Parmelee

J. H.

Bichard

S. L.

(2012). Politics and the Twitter revolution: How tweets influence the relationship between political leaders and the public. Lanham, MD: Lexington Books.

26.

Petrocik

(1996). Issue ownership in presidential elections, with a 1980 case study. American Journal of Political Science, 40, 825-850. doi:10.2307/2111797

27.

Petrocik

Benoit

Hansen

(2003). Issue ownership and presidential campaigning, 1952-2000. Political Science Quarterly, 118, 599-626. Retrieved from www-jstor-org.web.bisu.edu.cn/stable/30035698

28.

Pew. (2014). As midterms near, GOP leads on key issues, Democrats have a more positive image. Available from http://www.people-press.org/

29.

Rhea

Chattopadhyay

(2014, May). Issue ownership in the 2012 US presidential election—General campaign advertisements. Paper presented at the annual meeting of the International Communication Association, Seattle, WA.

30.

Riffe

Lacy

Fico

(1998). Analyzing media messages: Using quantitative content analysis in research. Mahwah, NJ: Lawrence Erlbaum.

31.

Schlenker

(1980). Impression management: The self-concept, social identity, and interpersonal relations. Monterey, CA: Brooks/Cole.

32.

Sides

(2006). The origins of campaign agendas. British Journal of Political Science, 36, 407-436. doi:10.1017/S0007123406000226

33.

Smith

Paul

(2007). No place for a woman: Evidence for gender bias in evaluations of presidential candidates. Basic and Applied Social Psychology, 29, 225-233. doi:10.1080/01973530701503069

34.

Vavreck

(2009). The message matters: The economy and presidential campaigns. Princeton, NJ: Princeton University Press.

35.

Williams

(2006). Self-referential and opponent-based framing: Candidate e-mail strategies in campaign 2004. In Williams

Tedesco

(Eds.), The Internet election: Perspectives on the web in campaign (pp. 83-98). Lanham, MD: Rowman & Littlefield.

36.

Winter

(2010). Masculine republicans and feminine Democrats: Gender and Americans’ explicit and implicit images of the political parties. Political Behavior, 32, 587-618. doi:10.1007/s11109-010-9131-z