“You Tweet Like a Girl!”

Abstract

We investigate the Twitter activity of all congressional candidates leading up to the 2012 U.S. House elections to assess whether there are significant differences in the tone and content of the tweets from male and female candidates. We argue that the electoral environment will have a significant effect over whether candidates engage in negative tweeting, address political issues, and discuss so-called “women’s issues” on Twitter. We find that gender has both a direct and contextual effect on candidates’ communication style on Twitter. Female candidates tweet significantly more “attack-style” messages than their male counterparts, discuss policy issues at a significantly higher rate, and women representatives focus more on “women’s issues.” We also find strong contextual effects in races with more female candidates: There is significantly more tweeting about political issues as well as significantly more negative attack-style tweets. However, with more female candidates, the number of tweets about “women’s issues” declines.

Keywords

Twitter gender campaign communication social media

Introduction

Over the past decade, women have made significant strides in American politics; however, women still have yet to gain parity to men in the political arena. Scholars have identified a variety of factors that lead to women’s underrepresentation in political institutions. One explanation for the greater number of men in political office is that women have had to fight against gender stereotypes over the years, which has affected their campaigning style (Fox, 1997; Herrnson & Lucas, 2006; Kahn, 1996; Kahn & Gordon, 1997). The media have historically treated male and female candidates differently. Until recently, women have received less coverage than their male counterparts and when they did receive coverage, gender stereotypes were often portrayed (Kahn, 1993; Kahn & Goldenberg, 1991). More recent studies note that these stereotypes have diminished somewhat and when women do run for office, they are often more successful than their male counterparts (Anzia & Berry, 2011; Cook, 1998; Lawless & Pearson, 2008). Despite these recent gains, women largely remain outsiders in the male-dominated political arena.

Because of their out-group status, we argue that women candidates adopt a different style of communication during their campaigns to both combat stereotypes and to distinguish themselves from male candidates. In examining gender differences in congressional campaign communication, we investigate the Twitter activity of all congressional candidates leading up to the 2012 U.S. House elections (September 6 to Election Day) to assess whether there are significant differences in both the tone and content of tweets sent by male and female candidates for office. We argue that the electoral environment will have a significant effect on whether candidates (male and female) engage in negative tweeting. We also examine whether women spend more time addressing so-called “women’s issues” on Twitter than men and whether they are more likely to address policy-specific issues in general.

Our findings lend support for the claim that gender has both a direct effect on candidates’ communication style as well as a contextual effect over campaign communication on Twitter. We find that female candidates tweet significantly more “attack-style” messages than their male counterparts, discuss policy issues at a significantly higher rate than their male counterparts, and focus more on “women’s issues.” We also find strong contextual effects in races with more female candidates: There is significantly more tweeting about political issues as well as significantly more negative attack-style tweets. However, in campaigns with more female candidates, the number of tweets about “women’s issues” declines as this is no longer a characteristic on which women candidates can distinguish themselves from their competitors.

As more women are entering politics and social media is becoming more widespread in campaigns, our research has broad implications. First, our research builds upon the extant literature on how female candidates are perceived by voters. Although some previous research has argued that female candidates are punished when they exhibit behavior counter to gender norms or stereotypes (Kahn, 1996; Trent & Friedenberg, 2008), we find—consistent with some of the more recent literature (see, for example, Brooks, 2011)—that female congressional candidates consistently posted more attack-style, negative tweets, engaging in counter-stereotypical behavior. However, female candidates may also use stereotypes to their advantage, especially as a way to distinguish themselves from their male competitors by focusing on issues in which they have an advantage with the electorate. The decision to focus on “women’s issues” is conditional on the electoral environment, particularly the number of other female candidates.

Theoretical Foundations

Research on gender and congressional elections highlights the unique experiences of men and women pursuing public office. Despite the fact that women candidates have recently fared just as well as their male counterparts in fundraising and vote totals (Cook, 1998; Lawless & Pearson, 2008), many have attributed the slow progression of women to public office to the socialization of gender roles and structural barriers. For instance, studies have shown that citizens are more likely to project feminine characteristics (e.g., compassionate and caring) onto women candidates while projecting masculine traits (e.g., confidence and strength) onto male candidates (Brown, Heighberger, & Shocket, 1993; Leeper, 1991; Rosenwasser & Seale, 1988). These “masculine” traits are associated with greater perceived competency, especially on foreign policy and economic issues, the so-called “male” issues (Huddy & Terkildsen, 1993; Lawless, 2004; Sanbonmatsu, 2002). Regardless of party identification, women are thought of as being more interested in and able to adequately deal with issues such as health care, education, poverty, child care, and the environment, whereas men are seen as being more competent in issues dealing with the military, economy, taxes, and trade (Brown et al., 1993; Dolan, 2010; Huddy & Terkildsen, 1993; Koch, 1999).

Other studies have shown that the media coverage that women receive is quite different from coverage of their male counterparts. Press coverage of female candidates tends to emphasize their appearance or personality, whereas coverage of men is much more likely to focus on their credentials, office-holding experience, and leadership qualities (Braden, 1996; Bystrom, Banwart, Kaid, & Robertson, 2004; Dunaway, Lawrence, Rose, & Weber, 2013). For instance, according to Representative Susan Molinari, female legislators are more likely to be asked superficial questions compared with their male colleagues who are more frequently asked to comment on policies. Rep. Molinari (1998) explains, “There I’d be, in a war zone in Bosnia, and some reporter—usually female—would comment on how I was dressed, then turn to my male colleague for answers to questions of substance” (p. 7).

These stereotypes could hamper female candidates from successfully competing in an arena where leadership and aggressiveness are highly valued by voters (Lawless, 2004). Women often perceive that they must work harder than men. For instance, Lawless and Fox (2005) find that women are less likely than similarly qualified men to believe that they are qualified to run for office. Moreover, women are more concerned than men with their legitimacy as candidates, likelihood of victory, and fundraising abilities (Dodson, 1998; Fulton, Maestas, Maisel, & Stone, 2006; Sanbonmatsu, 2002). These differences in attitudes within the electorate likely translate into differences in how male and female candidates campaign for congressional seats. Drawing from previous studies, we argue that these stereotypes offer both constraints and opportunities for female candidates, and the extent to which female candidates embrace these perceptions and run “as women” will depend on the electoral context, particularly their need to distinguish themselves from their competitors. In other words, women will seek to use their “out-group” status to their advantage. Likewise, when female candidates perceive these stereotypes as constraints or liability, they will engage in behaviors that diminish their “out-group” status, particularly highlighting partisan—rather than gender—differences.

Gender and Campaigning on (Women’s) Issues

Research on the gender gap has demonstrated significant divisions in the attitudes, priorities, and voting patterns of men and women within the electorate (Box-Steffensmeier, DeBoef, & Lin, 2004; Kaufmann & Petrocik, 1999; Seltzer, Newman, & Leighton, 1997).

Public opinion scholarship has demonstrated consistently that men hold more conservative attitudes than women on criminal justice, foreign policy, and social welfare issues (Kaufmann & Petrocik, 1999; Shapiro & Mahajan, 1986). Women are more strongly in favor of environmental protection (Wirls, 1986) and more supportive of welfare programs aimed at the elderly, sick, and poor as well as government funding of education and health care (Deitch, 1988; Page & Shapiro, 1992). These gender differences in political attitudes within the electorate also translate into differences in how male and female candidates are evaluated by voters. Voters assess women candidates more favorably than their male counterparts on social policy issues involving children, education, health care, and the elderly, whereas male candidates are perceived more capable in handling tax and foreign policy issues (Burrell, 1994; Huddy & Terkildsen, 1993; Sanbonmatsu, 2002). Women are viewed as being more liberal than men within each party (Koch, 2000).

Gender stereotypes can place women candidates in a “double-bind” (Jamieson, 1995). On one hand, vote-seeking candidates have incentives to emphasize issues that benefit them electorally, and to the extent that women are rated more favorably on so-called “women’s issues,” highlighting women’s issues to distinguish oneself from male candidates could prove very beneficial to female candidates. This is consistent with the idea of gender issue ownership. For instance, surveying congressional and state legislative candidates running for office in 1996 or 1998, Herrnson, Lay, and Stokes (2003) demonstrate that women who strategically emphasized “women’s issues” to target female voters gained a significant advantage at the polls. On the other hand, women candidates who run “as women” may be punished by voters who perceive them as ineffectual or only able to speak to a narrow set of “women’s issues” (Larson, 2001; Witt, Paget, & Matthews, 1994). Thus, the extent to which women’s issues are brought to the fore over the course of the campaign should be dependent upon the context of the election. Because campaign messages frequently highlight issues in which the candidate has an advantage, while deemphasizing other issues (Petrocik, 1996; Sellers, 1998), candidates may also decide to target their communication to draw support from particular groups or avoid losing the support of particular groups, such as women voters.

Although previous literature demonstrates that “women’s issues” and the outsider status of women politicians can provide an advantage to female candidates in some contexts (Dolan, 1998; Fridkin & Kenney, 2009), this literature only examines traditional media advertising, such as radio, television, and print advertisements. Scholars have only just begun exploring how candidates use new media, such as Twitter, to gain voters’ support (Evans, Cordova, & Sipole, 2014; Gainous & Wagner, 2014; Hargittai & Litt, 2011; A. Smith & Rainie, 2010). Twitter, a microblogging Internet platform, allows users to communicate with each other in short statements of 140 characters or less. Although the site began in 2006, it was not fully used by most politicians until 2012, when it was the eighth most-visited website during the election (Hendricks, 2014). In October 2012, there were approximately 140 million Twitter users (C. Smith, 2012). As Hendricks (2014) describes, “with that many users, or potential voters, the candidates could not have ignored this social-media platform to communicate with the electorate” (p. 137). Twitter, like all social media websites, allows candidates to communicate directly with their followers in a rapid pace, bypassing the gatekeepers of traditional television and newspaper media.

Although it is undeniable that social media such as Twitter have profoundly changed campaign communication in the United States, scholars have pointed to a variety of ways in which these new media platforms have also presented challenges for candidates and their campaigns. Although candidates are able to bypass the traditional media outlets and launch appeals directly to the public, this may also result in less overall control over the campaign’s message (Gueorguieva, 2008; Johnson & Perlmutter, 2010). Moreover, as some researchers have noted that with the increasing amount of narrowcasting, or targeted communication, employed by campaigns to appeal to particular groups, social media outlets such as Twitter reduces the audience into one—what Marwick and Boyd (2011) refer to as “context collapse” (p. 9). Candidates have many fewer constraints upon the messages they present to potential voters; however, they also must be aware that any of their tweets could be seen by anyone, whether a member of the intended audience or not. Finally, social media outlets also enable candidates to circumvent the usual authoritative hierarchies found with traditional media outlets as well as the party elites (Metzgar & Maruggi, 2009). This suggests some benefits to outsider candidates yet also presents challenges to party elites who wish to maintain some control over their party’s “brand name.”

Given the contrasts of social media outlets and traditional media outlined above, there is good reason to believe that Twitter or other forms of social media could be especially beneficial for female candidates because candidates themselves control the message. As Dolan (2005) describes, like candidate websites, candidates can discuss “as many or as few issues as they want” and this “lets us see the decisions candidates and their staffs make about how they present themselves to the public” (p. 33) by eliminating the bias from the media. During the 2012 campaign, one observer noted that “because Twitter democratizes the delivery of information, tweets can help a candidate by getting out a message that might not be seen on traditional media” (Lever, 2012, p. 1). Gainous and Wagner (2014), for instance, suggest that members of the out-group (those in the minority or political outsiders) may be drawn to Twitter to gain an advantage over the majority group members. Karpf (2012) popularized this “out-group” thesis in his book, The Move-On Effect by showing that minority group members are drawn to new social media to gain an electoral advantage over majority members. He argued that out-party innovation incentives lead candidates and groups to engage in counter-mobilization, which may occur at the organizational, candidate, and party coalition level.

Theoretically, many scholars have argued that actions on the part of the governing group can invoke “political outsiders” to react with counter-mobilization tactics (see, for example, Rosenberg, 1991; Tarrow, 1998; Truman, 1951). Appleton and Ward (1997) and Lowi (1963) also have shown that parties that lose elections are likely to innovate to improve their chances of success in the next election. Twitter offers many advantages over traditional media for all candidates, but particularly those in the minority. First, Twitter is essentially free, which is a benefit to candidates who may have limited campaign resources. Twitter also provides an unlimited space, in terms of the number of tweets, for discussion of issues. Social media also levels the playing field in terms of candidate mentions (Metzgar & Maruggi, 2009). Challengers and minority group members suffer from many media challenges in traditional campaigns, such as the shrinking sound bite and the cost of advertising. As Johnson (2012) describes, however, tweets “are becoming the new ‘sound bite,’ allowing candidates to control more of the coverage of their public images and their campaigns” (p. 55).

Because women are a minority in Congress, they should be drawn to this new media as well. Research has shown that females in general are more likely to have accounts on Twitter and use them more frequently (Gainous & Wagner, 2014; Hargittai & Litt, 2011; A. Smith & Rainie, 2010). Moreover, Evans et al. (2014) find that women candidates are more likely to discuss political issues compared with their male counterparts. To date, there has been no research examining whether women candidates are more likely to use Twitter to discuss “women’s issues,” or whether the discussion of issues on Twitter by female candidates is dependent on the context of the campaign and the sex of their opponent(s).

Because female candidates experience greater challenges in establishing their qualifications for office, women should be more likely to use Twitter to combat stereotypical news coverage that focuses on personal traits in favor of getting out the message on political issues. We expect that women will send more tweets focused on policy issues than men. We also expect that in races with more female candidates, we will see much more discussion of policy issues on Twitter to combat stereotypical coverage by traditional media outlets. Thus, we set out the following hypotheses:

Hypothesis 1 (H1): Women candidates will discuss policy matters at a higher rate than their male counterparts on Twitter.

Hypothesis 2 (H2): As the percentage of female candidates in a race increases, there will be significantly more policy-related tweets by candidates.

Because voters often perceive female candidates’ strength to be the so-called “women’s issues” (i.e., health, education, welfare, etc.), female candidates may use Twitter to emphasize “women’s issues” to distinguish themselves from male competitors on issues in which they have an advantage due to voter stereotypes. This leads to the following hypothesis:

Hypothesis 3 (H3): Female representatives will tweet about “women’s issues” at a significantly higher rate to distinguish themselves for their male counterparts.

However, as more female candidates compete for a congressional seat, the strategic use of “women’s issues” may no longer serve as a beneficial strategy for candidates to use to distinguish themselves from their competitors. This leads to the following hypothesis:

Hypothesis 4 (H4): As the percentage of women in the race increases, there will be significantly fewer tweets about “women’s issues” because the salience of gender as a distinguishing characteristic is reduced.

Gender and Negativity Campaign Communication

Voters’ expectations and stereotypes of candidates may also influence candidates’ decisions concerning if and when to engage in negative advertising or negative campaign communication. Some argue that the different campaign tactics taken by male and female candidates are largely due to different socialization processes and are a product of the different expectations of male and female candidates held by voters and reporters (Carroll, 1994; Kahn, 1993; La Cour Dabelko & Herrnson, 1997). Gender stereotypes in society depict women as kind, helpful, sympathetic, and passive, and men are often deemed independent, aggressive, and forceful (Fridkin & Kenney, 2009; Huddy & Terkildsen, 1993). These commonly held beliefs translate into voter expectations about male and female political candidates and likely affect the tone of candidates’ political communication.

Given these gender stereotypes, the costs and benefits of engaging in negative campaigning may differ for female candidates. When female candidates engage in negative campaigning, it runs counter to gender stereotypes because negative campaigning is perceived as an aggressive tactic. They may opt to conform to conventional gender stereotypes and eschew negative campaign communication. Alternatively, they may opt to engage in negative campaigning tactics to dispel stereotypes. Political science research has produced mixed evidence concerning the relationship between candidates’ gender and the use of negative campaign tactics. Female voters in particular tend to dislike negative advertisements (Gordon, Shafie, & Crigler, 2003); therefore, to the extent that female candidates wish to cultivate the “female vote,” they may fear that going negative would backfire. However, King and McConnell (2003) find that negative campaigning is less effective in reducing female candidates’ evaluations compared with male candidates. Other work suggests that negative communication could serve to bolster voters’ evaluation of female candidates’ competency because it runs counter to existing stereotypes (Kelley & McAllistar, 1983; Lau & Pomper, 2004) and that women could benefit from attacking their opponents on male issues (Gordon et al., 2003). By contrast, some scholarship (Kahn, 1996; Trent & Friedenberg, 2008) finds that female candidates could suffer electorally when they violate gender stereotypes.

Although much of the work on gender and negative campaigning has focused on traditional media and advertisements (e.g., television, radio, and news print), scholars are beginning to address how this plays out on social media. Using a small sample of eight candidates, Parmelee and Bichard (2012) show that women were less likely to send negative tweets. Evans et al. (2014), however, using a data set of 1,119 candidates in the 2012 election, demonstrate that women spend more time on Twitter sending negative attack-style tweets than men. However, neither considers how the context of the election—specifically the number of women and men in the race—shapes negative tweeting.

These studies produce divergent expectations for the direct role of gender on negative tweeting:

Hypothesis 5a (H5a): Female candidates will go on the attack through Twitter, tweeting “attack-style” messages at a significantly higher rate than their male counterparts.

Hypothesis 5b (H5b): Female candidates—conforming to stereotypes—will produce significantly fewer negative “attack-style” tweets compared with their male counterparts.

Candidates’ communication style is also dependent upon the other candidates in the election. Kahn (1993) demonstrates that male candidates are less likely to engage in negative campaigning when facing a female opponent for fear of backlash. However, if women are seeking to dispel gender stereotypes, there may be an increase in negative tweets with more female candidates. For these competing theoretical reasons, we also investigate the following research question:

Research Question 1 (RQ1): When there are more women in the race, will there be more attack-style tweets than in contexts with fewer female candidates?

Data and Method

To test our hypotheses, we performed a content analysis of every tweet posted by candidates for the U.S. House of Representatives during the 2 months prior to Election Day (September 6-November 6).¹ The Twitter data were collected by a group of students who were assigned a particular set of candidates to “follow.” They gathered the tweets and compiled a spreadsheet of information, including background information on the candidates as well as classification of every tweet sent by hand. Of the 1,119 candidates, 765 (68.4%) had a Twitter account that they were using for their campaign during this time period. These candidates sent a total of 67,119 tweets.² Female candidates accounted for 21% (162 candidates) of our data set. For more details on the characteristics of those using Twitter versus those who do not, please see online appendix.

Whereas Evans et al. (2014) investigate which candidates are more likely to have a Twitter account, the number of tweets sent, and the number of followers, our focus here is the content of candidates’ tweets. In particular, we coded every tweet along three dimensions: (a) message tone, (b) issue, and (c) “women’s issue.” The first dimension, message tone, indicates whether a candidate’s tweet is negative in the tone of its message. To code message tone of tweets, we relied upon Finkel and Geer’s (1998) classification of positive versus negative advertisements. They classified positive advertisements as that in which “candidates offer to promote themselves on some issue or trait,” and they classified negative advertisements as those in which “attacks [were] leveled at the opposition” (p. 579). Thus, we did not simply code whether or not a competitor was mentioned in a tweet but rather the actual content of the tweet. Negative tweets were those that directly attacked the candidate’s opponent(s). One example of a negative tweet was Rep. Eric Swalwell’s (D-CA, 15) tweet on October 20, “Surprise! Another candidate event and & another absence for Pete Stark. Does he know there’s an election or want the job?” The average number of attack tweets by candidates was 5.17, with a range of 0 to 200. There was not an overwhelming amount of negative tweeting as almost 58% of all candidates did not engage in attacking their opponents on Twitter during this time period.

We also coded each tweet for whether or not it pertained to a policy issue as opposed to constituency outreach, mobilization, personal messages, or something else. Issue-specific tweets ranged considerably in the types of issues discussed, including health care (e.g., the Affordable Care Act, “Obamacare”), education, foreign policy, and many other important national issues. Each tweet was coded 1 if it dealt with a policy issue and 0 otherwise, and from these data, we summed the results to create a count variable for each candidate indicating how many tweets pertaining to policy issues he or she had during this time period. The mean number of issue tweets by candidates was 9.5 with a standard deviation of 18.03. In 24% of cases, candidates did not have a single tweet discussing policy issues. Betsy Dewey, a Libertarian candidate in the Texas 25th congressional district, had the most issue tweets at 149 during this time period.

Finally, we coded tweets for whether or not they pertained to issues traditionally defined as “women’s issues.” There is an extant literature that has considered both what constitutes a “woman’s issue” and whether women candidates or elected officials are more likely to emphasize “women’s issues” (Bratton, 2002; Dodson & Carroll, 1991; Reingold, 2000; Swers, 2002; Thomas, 1991, 1994; Wolbrecht, 2000). Although there are some differences in their definitions, there is also considerable overlap in how they define “women’s issues.” Each of these classifications define “women’s issues” as a core set of issues that directly and disproportionately affect women as a group. Some scholars include in their definition of women’s issues a broad set of policy issues that have traditionally been associated with women, including health, welfare, education, and the environment; whereas other scholars merge issues of traditional concern with feminist issues, such as equality/equal rights and Gay, Lesbian, Bisexual, Transgender, Questioning, and Allied (GLBTQA) rights. The definition of women’s issues used in this study draws from the definition reflected in this previous scholarship. As such, we focus both on traditional issues associated with women as a group as well as feminist concerns seeking to improve the social, economic, and political status of women as a group. Specific examples are provided in the appendix; however, these include health care, education, welfare, the environment, children, family, equality, gay rights, and poverty. Within these broader categories, we included crimes that disproportionately affect women, including rape and domestic violence, were classified as women’s issues. In addition, we classified Plan B, abortion, pro-choice, and pro-life as “women’s issues.” Although some might argue that these latter issues are more partisan than gender related, we find that both Republican and Democratic women are significantly more likely than their male copartisans to tweet about these issues, suggesting that although they undoubtedly have become partisan issues over time, there remains a salient gender dimension to these issues as well.

The coding for “women’s issues” was not part of the original data collection efforts and unfortunately not every tweet was saved for all the candidates. When the coders returned to capture these data, many of the losing candidates had already deleted their accounts or particular tweets. Therefore, for this subsequent analysis, we were only able to code tweets for U.S. representatives (i.e., the winning candidates in the House races), because their accounts had remained intact over this full time period. The tweets were coded by a keyword search of all the tweets sent. During the 2012 elections, 78 women claimed victory in the House (17.9%). The mean number of women’s issues by candidates was 4.23 with a standard deviation of 9.35. There were 190 representatives who did not discuss women’s issues at all during this time period. Representative Betty McCollum (D-MN) had 149 tweets specifically about women’s issues. Despite some of the issues classified as “women’s issues” also having a partisan leaning to them (e.g., Obamacare, pro-life, pro-choice, Plan B, etc.), these issues still have a salience for women as a group.

To examine whether partisanship overwhelmed the gendered nature of these issues, we first performed a test of whether there were significant differences between the mean number of women’s issues tweeted by Republicans and Democrats. We found that Democrats tweeted on average 4.96 women’s issues and Republicans tweeted on average 3.6 women’s issues; this difference was not statistically significant (p < .065). We then separated the legislators by party to assess whether there were significant differences between women and men within each of the parties. Democratic women posted on average 7.53 tweets pertaining to women’s issues, and Democratic men posted an average of 3.92 women’s issue tweets. The difference in means (between Democratic male and female candidates) is statistically significant at the .05 level. Republican women posted on average 6.6 tweets about women’s issues, and Republican men posted on average 0.36 tweets about women’s issues. This difference (between Republican men and women) is statistically significant at the .01 level. Thus, we find through exploring the difference in means that gender is a salient factor influencing whether or not so-called women’s issues are discussed on Twitter, and this is true despite our inclusion of some issues or topics that have been associated quite closely to the parties (e.g., Obamacare, Plan B, “binders full of women,” etc.). It is also important to note that Republican women were quite active in tweeting about women’s issues even though some have argued that Democrats have “ownership” on some of these issues. If Republican women perceived a disadvantage in discussing some of the Democratic-owned issues that overlapped with “women’s issues,” it did not appear to discourage them from discussing the issues on Twitter during this election cycle.

The two main explanatory variables we include to assess our hypotheses are the candidate’s gender and the proportion of women in the election. The gender variable is coded as 1 for female candidates and 0 for male candidates. The proportion of women in each race is simply a ratio of the number of women in the race divided by the total number of candidates in the race. In addition, we include an explanatory variable indicating whether the candidate is an incumbent (=1) or not. Both these variables (gender and incumbency) were obtained through the official record of U.S. House members found at www.house.gov. We also include a dummy variable for whether the individual was a third-party candidate (=1) versus a major-party candidate (=0).³ Evans et al. (2014) show that during the 2012 election, third-party candidates and challengers were more active on Twitter than major-party candidates and incumbents. Twitter provides an equal space for lesser known candidates (Gueorguieva, 2008). Therefore, we might expect third-party candidates to differ in how they communicate their messages to voters. Because they often receive less coverage from traditional news sources, third-party candidates may use Twitter to talk about the issues more than major-party candidates, for instance. We include a control variable for the number of tweets issued by the candidate to account for the possible bias that could result from simply having more versus less prolific candidates on Twitter. In other words, we want to rule out that the individual candidate differences in the number of policy-related tweets could simply be due to some candidates simply tweeting significantly more than other candidates. In the models involving only the winning candidates, we include a dummy variable for whether the representative is a Republican (=1) versus a Democrat (=0). We also included a dummy variable for whether the congressional seat was open (=1). Finally, we include an indicator for the district competitiveness. Any race listed as a “toss-up” or “leaning Republican” or “leaning Democratic” by the Cook Political Report on September 13, 2012, was coded as competitive.⁴ Evans et al. (2014) find that on average candidates from competitive districts used three more attack tweets than those in uncompetitive races. The descriptive statistics of our key independent variables as well as a breakdown of Twitter activity are provided in Table 1.

Table 1.

Descriptive Statistics.

	M		SD
Female	0.21		0.41
Incumbent	0.44		0.50
Open	0.14		0.35
Total number of tweets	87.74		125.44
Competitive district	0.15		0.36
% female candidates	0.20		0.25
Third party	0.09		0.29
Number of tweets (M)	Incumbent males	Incumbent females	Non-incumbent males	Non-incumbent females
Democrats	6,363 (122.4)	2,330 (105.9)	12,871 (186.5)	7,494 (245.7)
Republicans	9,120 (106.6)	2,580 (234.5)	13,815 (204.7)	2,930 (217.0)
Third-party candidates	NA	NA	7,549 (251.6)	2,067 (318.0)

We estimated three empirical models of candidates’ campaign activities via Twitter that investigate the factors that explain the number of (a) tweets addressing policy issues, (b) tweets focused on “women’s issues,” and (c) negative “attack-style” tweets. The unit of analysis is the individual candidate. Because the dependent variables for each of these models is the number of tweets by a candidate (or representative in the second model), a count model is more appropriate than a linear regression model (Long, 1997). Count models enable us to take into account the discrete and non-negative characteristics of the dependent variable to acquire unbiased estimates of the effects of the independent variables. In some instances of our dependent variables, there are many zero observations in the data, which reflect candidates who did not issue any attack tweets as well as candidates who did not tweet at all about women’s issues. In those scenarios, there may be greater dispersion than in a Poisson count model.⁵ The Negative Binomial Regression Model (NBRM) is an extension of the Poisson Regression Model that enables the conditional variance of the dependent variable to exceed the conditional mean. The NBRM allows us to model heterogeneity in our data while not eliminating the possibility of the data having the Poisson distribution.

Candidates’ political communication via Twitter involves a combination of decisions. First, the candidate must decide whether to adopt and use Twitter, and if so, then the next decision becomes how to effectively utilize this technology to communicate with voters. Because candidates’ decisions to adopt Twitter are not random and may be related to many of our key variables of interest, a failure to account for the selection process could likely produce biased inferences. Evans et al. (2014) find that women, major-party candidates, incumbents, and candidates in competitive races are significantly more likely to adopt Twitter. To estimate this two-stage decision-making process, we first investigate the factors that shape adoption of Twitter (selection), calculate the inverse mills ratio, and use the inverse mills ratio in our final outcome models to account for the selection process. In the selection equation, the covariates included candidate gender, whether the candidate ran under a major-party label, incumbency status, and whether the race was deemed competitive. The predicted probabilities for the model estimating the adoption and use of Twitter are presented in the online appendix.

Results

We estimated a Poisson Regression Model of the number of tweets concerning policy issues posted by each candidate.⁶ We present the results of our model in Table 2. We find strong evidence that supports our hypotheses about the direct and contextual effects of gender on campaign communication via Twitter. In particular, we find support for H1 that women tweet more about policy issues than their male counterparts, and this effect is significant at the p < .01 level. We further find support for H2. The percentage of women in the race also has a highly significant effect on the extent to which candidates discuss policy issues on Twitter (p < .001). To investigate the magnitude of these effects, we calculated the baseline expected number of tweets by holding all continuous variables at their means and the dummy variables at 0. Then, the first difference describes a marginal effect of the covariate in question (i.e., how a change in one variable affects the number of expected bill introductions relative to the baseline). For continuous variables, the change is an increase of one standard deviation, and for dummy variables, the change is from 0 to 1. We find that women candidates have an expected increase of about 17.5 policy-related tweets compared with male candidates. Moreover, a standard deviation increase in the percentage of women in the election is associated with an expected increase of about 9 more tweets about policy issues, all else constant.

Table 2.

Poisson Regression Model of the Number of Policy Issue Tweets.

	Coefficient (SE)
Female	0.157 (0.058)**
Incumbent	0.139 (0.238)
Open	−0.107 (0.233)
Total number of tweets	0.004 (0.000)***
Competitive district	0.290 (0.050)***
% female candidates	0.656 (0.085)***
Third party	−0.236 (0.037)***
Inverse mills ratio	0.630 (0.021)***
Constant	1.900 (0.246)***
Log likelihood	−743.21
Prob > χ²	.000

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

A number of control variables also significantly influenced candidates’ decisions to discuss policy matters on Twitter. The overall number of tweets by a candidate had a strong, positive effect on the number of policy-specific tweets by the candidate. Moreover, candidates in competitive districts also appear to have a significantly higher number of tweets pertaining to policy issues compared with their counterparts in non-competitive elections. Finally, party also emerged as a strong predictor of policy-specific issue tweeting. Third-party candidates posted significantly fewer tweets pertaining to policy issues than their major-party counterparts. Although Twitter offers third-party candidates an alternative avenue to broadcast their policy stances, they do not appear to be using the technology to discuss policy; rather, they have largely focused on personal messages. There were no significant differences between incumbents and challengers in the number of policy-relevant tweets nor were there significant differences for candidates running for open seats.

Discussing Women’s Issues on Twitter

Despite women’s greater focus on issue messages more generally during the campaign, were women also more likely to target women voters as a group through a focus on issues that disproportionately affect women as a group? To examine this question, we coded all tweets by the winning candidates by issue type. In the appendix, we include a list of the issues that we coded as “women’s issues” and those coded as “male issues.” The mean of the “women’s issues” variable was 4.23, and the variable ranged from 0 to 149. A difference of means test finds that women in office had significantly more tweets concerning women’s issues compared with their male counterparts (7.29 vs. 3.56, significant at the .1% level).

In Table 2, we examine how often female representatives and male representatives discussed so-called “female” issues on Twitter during the election. Table 3 presents the results from an NBRM of the number of tweets concerning “women’s issues” by each candidate with inclusion of the inverse mills ratio to account for sample selection. We find strong support for the role of gender in enhancing the attention to issues that disproportionately affect women as a group (H3). In particular, we find that female representatives discussed women’s issues on Twitter at a significantly higher rate than male representatives. Investigating the first difference, we also find that this result is substantively significant: Women representatives are expected to issue about 15.2 more tweets about “women’s issues.” Democratic representatives also addressed “women’s issues” on Twitter at a significantly higher rate than Republican representatives. Democratic representatives are expected to tweet about 7 more “women’s issue” tweets than Republican representatives.

Table 3.

Negative Binomial Regression Model of Women’s Issue Tweets.

	Coefficient (SE)
Female	0.591 (0.244)*
Incumbent	3.678 (0.351)***
Republican	−0.051 (0.152)
Competitive district	0.10 (0.05)*
Total tweets	0.007 (0.001)***
% female	−0.343 (0.352)
Inverse mills ratio	0.721 (0.041)***
Constant	−1.810 (0.478)***
Log likelihood	−868.022
Prob > χ²	.000

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

The percentage of women in the election did not increase candidates’ attention to issues that disproportionately affect women. This is consistent with H4, suggesting that when there are women competing with each other, it is less useful to emphasize women’s issues to distinguish oneself from a female competitor. We also investigated the effect of a number of control variables on attention to women’s issues via Twitter. We found no significant incumbency effect on attention to women’s issues; representatives who previously had served in the U.S. House were not discussing women’s issues at a significantly different rate than those newly elected to the House. However, representatives who were elected in competitive districts discussed women’s issues at a significantly higher rate than those from safe districts at the 5% significance level.

Negative “Attack-Style” Tweeting

Table 4 presents the results of the NBRM of the number of negative tweets by congressional candidates. This model lends support for our key hypotheses regarding both the direct and contextual effect of gender on the negativity of campaign messages. First, consistent with H5a, we find that female congressional candidates were more likely to engage in negative tweeting during the campaign. This variable was marginally significant at the .10 level, and the coefficient in the hypothesized direction. On average, female congressional candidates are expected to post about seven more negative tweets compared with their male counterparts, all else equal. This finding is consistent with the research of Evans et al. (2014), yet departs from the results of Parmelee and Bichard (2012). In addition, we found strong support that the percentage of women in the race influenced “attack-style” behavior on Twitter (RQ1). When more females were present in the race, there were a significantly higher number of negative tweets during the campaign. This effect is significant at the 1% level, with a coefficient of 1.11. We find that for a standard deviation increase in the percentage of female candidates in an election, there are expected to be about nine more negative tweets.

Table 4.

Negative Binomial Regression Model of the Number of Negative Tweets by Candidates.

	Coefficient (SE)
Female	0.318 (0.169)^†
Incumbent	−0.176 (0.411)
Third party	0.436 (0.124)***
Competitive district	−0.145 (0.125)
Open seat	0.161 (0.416)
% female candidates	1.11 (0.255)***
Inverse mills ratio	0.976 (0.032)***
Constant	0.968 (0.513)^†
Log likelihood	−460.862
Prob > χ²	.000

†

p < .10. *p < .05. **p < .01. ***p < .001.

Among the control variables, only partisanship was a significant predictor of the number of negative tweets. In particular, third-party candidates were most actively engaged in attack tweets against their major-party opponents.⁷ Interestingly, we did not find any significant differences between Democratic or Republican candidates in their rates of negative tweeting.⁸ The electoral environment also had a significant effect on negative tweeting. Candidates running for competitive seats issued significantly more negative tweets. However, whether candidates were running for open seats did not significantly influence their decision to “go negative” on Twitter.

Conclusion

This study investigates the content of congressional candidates’ communication via Twitter during the period leading up to the 2012 congressional elections. We argue that gender significantly influences both the tone and types of tweets of candidates. Our analysis demonstrates that gender has both a direct and contextual effect on Twitter behavior. First, we hypothesized that women would present more information about their policy preferences (H1), that they would be more likely to discuss “women’s issues” (H3), and that both of these would be dependent on the context of the election (H2/H4). We find that not only are female candidates discussing policy issues at a significantly higher rate than their male counterparts, but that with the inclusion of more women in the race, more attention is paid to policy matters (H2). We also find that female representatives had a significantly higher rate of tweeting about “women’s issues,” or those that disproportionately affect women as a group (H3). Women are also less likely to discuss “women’s issues” when they are running against other women (H4). In terms of negativity during elections, we find, like Evans et al. (2014), that women candidates issue significantly more “attack-style” tweets than their male counterparts (H5a). In addition, as the percentage of women in the race increases, we find significantly more negative tweets (RQ1).

Our results demonstrate that females seeking congressional office act differently than males online. Women use Twitter in the manner that the “out-party” typically uses new social media (Gainous & Wagner, 2014; Karpf, 2012). Although we suggest here that their status in the legislature affects their tweeting style, there may be other reasons for their Twitter behavior. Only with candidate interviews will we have a clear answer as to why women and men use Twitter differently for their campaigning purposes. We see here, however, that women do send more issue-specific and negative tweets and that their behavior on Twitter is dependent on their opponents. Given the widespread belief that voters generally disapprove of candidates who rely on negativity (Lau & Rovner, 2009; Lau, Sigelman, & Rovner, 2007), the use of negative communication could serve to reduce the attacking candidate’s overall support, and this could be particularly pronounced for female candidates because female candidates’ negativity could be perceived as a counter-stereotypic behavior, which is frequently punished by voters (Fridkin, Kenney, & Woodall, 2009; Herrnson et al., 2003; Hitchon & Chang, 1995; Rudman, 1998). This could have broad implications for the representation of women in elective office, especially if voters penalize negativity in campaigns and may even disproportionately penalize female candidates who “go negative.”

Although these results show gender differences in Twitter behavior, given the rise in this social networking site and the use of the site by candidates, the effects may decline over time. Women were more likely to talk about female issues during the 2012 campaign, but this does not mean that they are more likely to discuss those issues on Twitter when the race is over. Future research should examine the Twitter behavior of members of Congress when an election is not on the horizon.

In future research, it would be useful to chart candidates’ Twitter activity throughout the election cycle to investigate whether the issues discussed change over time in response to events occurring over the course of the campaign cycle. This would provide important insight into the dynamics of campaigns, which we are unable to address because our study focuses on a couple of months leading up to Election Day. It would also be interesting to expand the scope of the study over time to examine whether candidates are responding to their competitors. In particular, when a challenger raises particular issues over the course of the campaign, does the incumbent respond and vice versa? Because Twitter is free and instantaneous, candidates should be better able to adapt their campaign messages to current events and to messages or issues being raised by competitors.

Footnotes

Appendix

Examples:

Examples of male and female issue tweets:

“Female” issue tweet: Niki Tsongas: “Joined colleagues on another amicus brief opposing #DOMA, hope to see this law follow #DADT and become another relic of a bygone era.”

“Male” issue tweet: Tim Huelskamp: “Still no answer from President Obama on why help was not sent to the embassy in #Benghazi http://t.co/diFrncuo.”

Acknowledgements

We would like to thank the Fall 2012 Media and Politics class of Dr. Heather Evans at Sam Houston State University, Victoria Cordova, Savannah Sipole, Andrew Anderson, and Johnny Nguyen for their assistance in the data collection used in this project.

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 Biographies

Heather K. Evans is an associate professor of Political Science at Sam Houston State University.

Jennifer Hayes Clark is an associate professor of Political Science at the University of Houston.

References

Anzia

S. F.

Berry

C. R.

(2011). The Jackie (and Jill) Robinson effect: Why do congresswomen outperform congressmen? American Journal of Political Science, 55, 478-493.

Appleton

Ward

(1997). Party response to environmental change. Party Politics, 3, 341-362.

Box-Steffensmeier

J. M.

DeBoef

Lin

(2004). The dynamics of the partisan gender gap. American Political Science Review, 98, 515-528.

Braden

(1996). Women politicians and the media. Lexington: The University Press of Kentucky.

Bratton

K. A.

(2002). The effect of legislative diversity on agenda setting: Evidence from six state legislatures. American Politics Research, 30, 115-142.

Brooks

D. J.

(2011). Testing the double standard for candidate emotionality: Voter reactions to the tears and anger of male and female politicians. The Journal of Politics, 73, 596-615.

Brown

Heighberger

N. R.

Shocket

P. A.

(1993). Gender-based differences in perceptions of male and female city council candidates. Women & Politics, 13, 1-17.

Burrell

(1994). A woman’s place is in the house: Campaigning for congress in the feminist era. Ann Arbor: University of Michigan Press.

Bystrom

D. G.

Banwart

M. C.

Kaid

L. L.

Robertson

T. A.

(2004). Gender and candidate communication: Videostyle, webstyle, and newsstyle. New York, NY: Routledge.

10.

Carroll

S. J.

(1994). The politics of difference: Women public officials as agents of change. Policy Review, 5, 11-20.

11.

Cook

E. A.

(1998). Voter reaction to women candidates. In Thomas

Wilcox

(Eds.), Women and elective office (pp. 56-72). New York, NY: Oxford University Press.

12.

Deitch

(1988). Sex differences in support for government spending. In Mueller

C. M.

(Ed.), The politics of the gender gap: The social construction of political influence (pp. 192-216). Newbury Park, CA: Sage.

13.

Dodson

D. L.

(1998). Representing women’s interests in the U.S. house of representatives. In Thomas

Wilcox

(Eds.), Women and elective office (pp. 130-149). New York, NY: Oxford University Press.

14.

Dodson

D. L.

Carroll

S. J.

(1991). Reshaping the agenda: Women in state legislatures. New Brunswick, NJ: Center for the American Woman and Politics.

15.

Dolan

(1998). Support for women’s interests in the 103rd congress: The distinct impact of congressional women. Women & Politics, 18, 81-84.

16.

Dolan

(2010). The impact of gendered stereotyped evaluations on support for women candidates. Political Behavior, 32, 69-88.

17.

Dolan

(2005). Do women candidates play to stereotypes? Do men candidates play to women? Candidate sex and issues priorities on campaign websites. Political Research Quarterly, 58(1), 31-44.

18.

Dunaway

Lawrence

R. G.

Rose

Weber

C. R.

(2013). Traits versus issues: How female candidates shape coverage of senate and gubernatorial races. Political Research Quarterly, 66, 715-726.

19.

Evans

H. K.

Cordova

Sipole

(2014). Twitter style: An analysis of how house candidates used twitter in their 2012 campaigns. PS: Political Science and Politics, 47, 454-462.

20.

Finkel

Geer

(1998). A spot check: Casting doubt on the demobilizing effect of attack advertising. American Journal of Political Science, 42, 573-595.

21.

Fox

R. L.

(1997). Gender dynamics in congressional elections. Thousand Oaks, CA: Sage.

22.

Fridkin

K. L.

Kenney

P. J.

(2009). The role of gender stereotypes in U.S. senate campaigns. Politics & Gender, 5, 301-324.

23.

Fridkin

K. L.

Kenney

P. J.

Woodall

G. S.

(2009). Bad for men, better for women: The impact of stereotypes during political campaigns. Political Behavior, 31, 53-77.

24.

Fulton

S. A.

Maestas

C. D.

Maisel

L. S.

Stone

W. J.

(2006). The sense of a woman: Gender, ambition, and the decision to run for congress. Political Research Quarterly, 59, 235-248.

25.

Gainous

Wagner

K. M.

(2014). Tweeting to power: The social media revolution in American politics. New York, NY: Oxford University Press.

26.

Gordon

Shafie

D. M.

Crigler

A. N.

(2003). Is negative advertising effective for female candidates? An experiment in voters’ uses of gender stereotypes. Harvard International Journal of Press/Politics, 8, 35-53.

27.

Gueorguieva

(2008). Voters, MySpace, and YouTube: The impact of alternative communication channels on the 2006 election cycle and beyond. Social Science Computer Review, 26(3), 288-300.

28.

Hargittai

Litt

(2011). The tweet smell of celebrity success: Explaining variation in twitter adoption among a diverse group of young adults. New Media & Society, 13, 824-842.

29.

Hendricks

J. A.

(2014). The new-media campaign of 2012. In Denton

R. E.

Jr. (Ed.), The 2012 presidential campaign (pp. 133-156). Lanham, MD: Rowman & Littlefield.

30.

Herrnson

P. S.

Lay

J. C.

Stokes

A. K.

(2003). Women running “as women”: Candidate gender, campaign issues, and voter-targeting strategies. The Journal of Politics, 65, 244-255.

31.

Herrnson

P. S.

Lucas

J. C.

(2006). The fairer sex? Gender and negative campaigning in U.S. elections. American Politics Research, 34, 69-94.

32.

Hitchon

J. C.

Chang

(1995). Effects of gender schematic processing on the reception of political commercials for men and women candidates. Communication Research, 22, 430-458.

33.

Huddy

Terkildsen

(1993). The consequences of gender stereotypes for women candidates at different levels and types of offices. Political Research Quarterly, 46, 503-525.

34.

Jamieson

K. H.

(1995). Beyond the double bind: Women and leadership. New York, NY: Oxford University Press.

35.

Johnson

(2012). Twitter Bites and Romney: Examining the rhetorical situation of the 2012 presidential election in 140 characters. Journal of Contemporary Rhetoric, 2, 54-64.

36.

Johnson

T. J.

Perlmutter

D. D.

(2010). Introduction: The Facebook election. Mass Communication and Society, 13, 554-599.

37.

Kahn

K. F.

(1993). Gender differences in campaign messages: The political advertisements of men and women candidates for U.S. Senate. Political Research Quarterly, 46, 481-502.

38.

Kahn

K. F.

(1996). The political consequences of being a woman. New York, NY: Columbia University Press.

39.

Kahn

K. F.

Goldenberg

E. N.

(1991). Women candidates in the news: An examination of gender differences in U.S. Senate campaign coverage. Public Opinion Quarterly, 55, 180-199.

40.

Kahn

K. F.

Gordon

(1997). How women campaign for the U.S. Senate: Substance and strategy. In Norris

(Ed.), Women, media, and politics (pp. 59-76). New York, NY: Oxford University Press.

41.

Karpf

(2012). The move-on effect. New York, NY: Oxford University Press.

42.

Kaufmann

K. M.

Petrocik

J. R.

(1999). The changing politics of American men: Understanding the sources of the gender gap. American Journal of Political Science, 43, 864-887.

43.

Kelley

McAllistar

(1983). The electoral consequences of gender in Australia. British Journal of Political Science, 13, 365-377.

44.

King

J. D.

McConnell

J. B.

(2003). The effect of negative campaign advertising on vote choice: The mediating influence of gender. Social Science Quarterly, 84, 843-857.

45.

Koch

(1999). Candidate gender and assessments of women candidates. Social Science Quarterly, 80, 84-96.

46.

Koch

(2000). Do citizens apply gender stereotypes to infer candidates’ ideological orientations? Journal of Politics, 62, 414-429.

47.

La Cour Dabelko

Herrnson

P. S.

(1997). Women’s and men’s campaigns for the US house of representatives. Political Research Quarterly, 50, 121-135.

48.

Larson

S. G.

(2001). Running as women? A comparison of female and male Pennsylvania assembly candidates’ brochures. Women & Politics, 22, 107-124.

49.

Lau

R. R.

Pomper

G. M.

(2004). Negative campaigning: An analysis of U.S. senate elections. Lanham, MD: Rowman & Littlefield.

50.

Lau

R. R.

Rovner

I. B.

(2009). Negative campaigning. Annual Review of Political Science, 12, 285-306.

51.

Lau

R. R.

Sigelman

Rovner

I. B.

(2007). The effects of negative political campaigns: A meta-analytic reassessment. The Journal of Politics, 69, 1176-1209.

52.

Lawless

J. L.

(2004). Women, war, and winning elections: Gender stereotyping in the post September 11th era. Political Research Quarterly, 57, 479-490.

53.

Lawless

J. L.

Fox

R. L.

(2005). It takes a candidate: Why women don’t run for office. New York, NY: Cambridge University Press.

54.

Lawless

J. L.

Pearson

(2008). The primary reason for women’s underrepresentation? Reevaluating the conventional wisdom. The Journal of Politics, 70, 67-82.

55.

Leeper

M. S.

(1991). The impact of prejudice on female candidates: An experimental look at voter inference. American Politics Research, 19, 248-261.

56.

Lever

(2012). Twitter shakes up US election campaign. Retrieved from http://phys.org/news/2012-08-twitter-election-campaign.html

57.

Long

J. S.

(1997). Regression models for categorical and limited dependent variables. Advanced quantatitive techniques in the social sciences vol. 7. Thousand Oaks, CA: Sage.

58.

Lowi

(1963). Toward functionalism in political science: The case of innovation in party systems. American Political Science Review, 57, 570-583.

59.

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.

60.

Metzgar

Maruggi

(2009). Social media and the 2008 U.S. presidential election. Journal of New Communications Research, 4, 141-165.

61.

Molinari

(1998). Representative mom: Balanced budgets, bill, and baby in the U.S. Congress. New York, NY: Doubleday.

62.

Page

B. I.

Shapiro

R. Y.

(1992). The rational public: Fifty years of trends in Americans’ policy preferences. Chicago, IL: University of Chicago Press.

63.

Parmelee

J. H.

Bichard

S. L.

(2012). Politics and the Twitter revolution. Lanham, MD: Lexington.

64.

Petrocik

J. R.

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

65.

Reingold

(2000). Representing women: Sex, gender, and legislative behavior in Arizona and California. Chapel Hill: University of North Carolina Press.

66.

Rosenberg

(1991). The hollow hope: Can courts bring about social change? Chicago, IL: University of Chicago Press.

67.

Rosenwasser

S. M.

Seale

(1988). Attitudes toward a hypothetical male or female presidential candidate: A research note. Political Psychology, 9, 591-598.

68.

Rudman

L. A.

(1998). Self-promotion as a risk factor for women: The costs and benefits of counterstereotypical impression management. Journal of Personality and Social Psychology, 74, 629-645.

69.

Sanbonmatsu

(2002). Gender stereotypes and vote choice. American Journal of Political Science, 46, 20-34.

70.

Sellers

P. J.

(1998). Strategy and background in congressional campaigns. American Political Science Review, 92, 159-172.

71.

Seltzer

Newman

Leighton

M. V.

(1997). Sex as a political variable: Women as candidates and voters in U.S. elections. Boulder, CO: Lynne Rienner.

72.

Shapiro

R. Y.

Mahajan

(1986). Gender differences in policy preferences: A summary of trends from the 1960s to the 1980s. Public Opinion Quarterly, 50, 42-61.

73.

Smith

Rainie

(2010). Who tweets? Washington, DC: Pew Research Center.

74.

Smith

(2012). How many people use the top social media, apps & services? Digital Marketing Ramblings. Retrieved from http://expandedramblings.com/index.php/resource-how-many-people-use-the-top-social-media/

75.

Swers

(2002). The difference women make: The policy impact of women in congress. Chicago, IL: University of Chicago Press.

76.

Tarrow

(1998). Power in movement. New York, NY: Cambridge University Press.

77.

Thomas

(1991). The impact of women on state legislative policies. Journal of Politics, 53, 958-976.

78.

Thomas

(1994). How women legislate. New York, NY: Oxford University Press.

79.

Trent

J. S.

Friedenberg

R. V.

(2008). Political campaign communication: Principles and practices. Lanham, MD: Rowman & Littlefield.

80.

Truman

(1951). The governmental process: Political interests and public opinion. New York, NY: Alfred A. Knopf.

81.

Wirls

(1986). Reinterpreting the gender gap. Public Opinion Quarterly, 50, 316-330.

82.

Witt

Paget

K. M.

Matthews

(1994). Running as a woman: Gender and power in American politics. New York, NY: The Free Press.

83.

Wolbrecht

(2000). The politics of women’s rights: Parties, positions, and change. Princeton, NJ: Princeton University Press.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.17 MB