In recent years, there has been renewed interest in how men's and women's self-presentation in political campaigns may differ. The expectation that such differences exist originates from a concern about the effects of gender stereotypes in politics, where women have been and still are underrepresented. Gender stereotypes assume essential differences between men and women regarding their abilities, character, and behavior (Dolan Reference Dolan2014, 22; Ellemers Reference Ellemers2018, 276–78). They not only are bundles of expectations about how men and women typically behave but also have a strong prescriptive dimension (Prentice and Carranza Reference Prentice and Carranza2002). Beliefs that women should display other qualities and behave in different ways than men can therefore be expected to influence women's self-presentation in political campaigns, either unconsciously or as a strategic reaction to the perceived existence of gender stereotypes in the electorate. In the latter case, candidates may attempt to present themselves according to stereotypical expectations, to overturn gender stereotypes, or to avoid them (Schneider Reference Schneider2014, 265). However, empirical studies have found only partial evidence for gender differences in self-presentation, and researchers increasingly stress how similarly men and women present themselves in political campaigns (Hayes and Lawless Reference Hayes and Lawless2016).
This study is the first to look at gender differences in candidate campaigns in Ireland. The Republic of Ireland has long been notorious for its underrepresentation of women in the legislature. Women's share in the Dáil, the lower house of the Irish parliament, increased by a meager 3 percentage points over the course of the 1990s and the first decade of the twenty-first century and stood at 15.1% after the 2011 general election. The introduction in 2016 of a gender quota of 30% for all party candidates has altered the institutional setting for women's candidacies. An unprecedented number of female candidates, both party affiliated and independent, ran in the 2016 general election. This Irish election is a particularly interesting case, for two reasons. First, the rather sudden introduction of a gender quota created a campaign environment that put the role of women in politics into the spotlight and made gender a particularly salient category. Second, the Irish political system has been devoid of polarized party politics for a long time, which could favor gender differences in self-presentation.
This article uses a unique data set based on video statements of almost 90% of the candidates who ran in this election to analyze gender differences in the self-presentation of male and female candidates. The article first reviews the literature on gender differences in campaign communication and the self-presentation of candidates. It then describes the characteristic features of electoral competition and campaigning in Ireland and the context in which the Irish gender quota was introduced. It highlights factors that render the 2016 election a particularly interesting case when it comes to possible gender differences and proposes four hypotheses regarding these differences. After introducing the empirical material and methods used, the article presents the results of the empirical analysis and discusses them in relation to the hypotheses and the research on gender differences in politics.
GENDER DIFFERENCES IN CAMPAIGN COMMUNICATION
The typification of individuals as representatives of a particular social category or group is part and parcel of social life, as it simplifies interactions with unknown individuals (Schutz Reference Schutz1972). It reduces information costs, especially when the chosen criterion of categorization is easily discernible (van den Berghe Reference van den Berghe1997, 6). Gender is such a primary social category in perception and shapes expectations and behavior even when other categories are also salient (Ridgeway and Smith-Lovin Reference Ridgeway, Smith-Lovin, Chafetz and Kaplan2006).
Yet expectations about how men and women typically behave also have a strong prescriptive dimension (Prentice and Carranza Reference Prentice and Carranza2002). Gender stereotypes not only bundle expectations about how men's and women's personal qualities and behavior differ but also shape expectations about how they should differ. Gender stereotypes can therefore be expected to influence women's self-presentation in political campaigns, either unconsciously or as a strategic reaction to the perceived existence of such stereotypes in the electorate.
Self-presentation in political campaigns encompasses the various ways in which candidates attempt to appeal to voters. Some aspects of self-presentation are specific to the chosen medium—for example, the casting and setting of television ads (Bystrom et al. Reference Bystrom, Robertson, Banwart and Kaid2004). Others can be measured in similar ways in different media, such as issue priorities, the emphasis on specific character traits, or the extent of personalization.
The empirical evidence regarding gender differences in self-presentation is mixed. Studies that compare large numbers of candidates have mainly used data from the United States. Kahn (Reference Kahn1993) analyzes television ads of male and female candidates for the U.S. Senate between 1984 and 1986. She finds differences in issue priorities that correspond to common gender stereotypes, according to which women are more communal, selfless, and concerned with others than men and strong on so-called compassion issues (Eagly and Steffen Reference Eagly and Steffen1984; Huddy and Terkildsen Reference Huddy and Terkildsen1993). Men more often addressed economic issues such as taxes and the federal budget, while women addressed social issues such as education or health. With regard to character traits, both male and female candidates discussed stereotypical male traits such as competence and leadership much more than stereotypical female traits such as compassion and honesty.
Fox (Reference Fox1997) compares male and female candidates who ran for election to the U.S. House of Representatives in California in 1992 and 1994, based on interviews with campaign managers. He identifies gender differences in issue priorities and highlighted personal traits. Women focused more on both stereotypical women's and stereotypical men's issues than did men and stressed their qualifications for office and their ties to the community. Although they did not refer to their family more often than male candidates, they discussed it more often in terms of personal experience as opposed to appeals to abstract family values. Schneider (Reference Schneider2014) analyzes candidates’ websites in four subsequent election cycles to Congress and measures the congruence of issue priorities and emphasis on the personal traits of leadership and empathy with stereotypical expectations about men and women. Her results indicate that women mostly emphasize issues and traits that are consistent with their gender and choose gender-congruent strategies more often than men.
Dolan (Reference Dolan2005; Reference Dolan2014, chap. 5) does not discern gender-specific issue priorities in electoral campaigns for the U.S. Congress in 2000, 2002 and 2010. Instead, she finds that the television ads and campaign websites of male and female candidates only differed in the attention given to a few less prominent issues and that these differences did not always correspond to gender stereotypes. Sapiro et al. (Reference Sapiro, Walsh, Strach and Hennings2011) also analyze television ads from the 2000 and 2002 congressional elections. They conclude that almost all gender differences were highly dependent on the particular electoral context, except for a more pronounced emphasis on toughness or strength in the ads of female candidates. Hayes and Lawless (Reference Hayes and Lawless2016, chap. 3) analyze television ads and Twitter messages of candidates in the elections for the U.S. House of Representatives in 2010 and 2014; they also find very few differences in the issue priorities of men and women.
Banwart and Winfrey's (Reference Banwart and Winfrey2013) analysis of candidate websites from the 2012 election shows similar issue priorities for male and female candidates. Both genders stressed personal qualities such as past experience, action orientation, and aggressiveness. But women were more likely to highlight their own competence as well as their sensitivity and understanding, thus presenting a blend of stereotypically male and female traits. Fridkin and Kenney (Reference Fridkin and Kenney2014) find that women candidates for the U.S. Senate tended to emphasize their political experience more often than men and their family less often to make sure they appeared qualified enough. Stalsburg and Kleinberg (Reference Stalsburg and Kleinberg2015) also find that women are more reluctant than men to stress their families in their self-presentation, although some research suggests that female candidates can benefit electorally from being known to be mothers, compared to women with no children (Stalsburg Reference Stalsburg2010).
Other research looks at general personalization, that is, the choice to “self-disclose information about their personal lives or use personal experiences to create a connection with a campaign topic” (Meeks Reference Meeks2017, 3). This research finds that men tend to personalize more than women, but also that the specifics of the campaign environment—namely, the competitiveness of a race—can encourage women to personalize (McGregor, Lawrence, and Cardona Reference McGregor, Lawrence and Cardona2017).
Windett (Reference Windett2014) attributes the mixed findings on gender differences at least partly to temporal effects, that is, changes of strategy that occur over the course of a campaign in reaction to other candidates. Hayes and Lawless (Reference Hayes and Lawless2016, 18–19) suggest two reasons for the many commonalities of male and female candidates’ self-presentation in U.S. politics. The first reason is the normalization of female candidacies over time (Center for American Women and Politics 2019). When female politicians are regarded as a normal part of the political landscape, voters may increasingly typify them as members of the category of politicians instead of using gender stereotypes (Schneider and Bos Reference Schneider and Bos2014). Women candidates who perceive such a change may no longer see the need for a self-presentation that takes gender stereotypes into account. The second reason that Hayes and Lawless identify is the increasing party political polarization in the United States, which has made the parties more internally cohesive (Andris et al. Reference Andris, Lee, Hamilton, Martino, Gunning and Selden2015; McCarty, Poole, and Rosenthal Reference McCarty, Poole and Rosenthal2006). Partisanship has therefore become the dominant criterion for the choices that candidates make about their self-presentation.
If the normalization of female candidates and party polarization tend to reduce gender differences in the self-presentation of political candidates, it will be particularly interesting to look at a case in which these factors are not present. Are the issue priorities of male and female candidates also similar (Banwart and Winfrey Reference Banwart and Winfrey2013; Dolan Reference Dolan2005; Sapiro et al. Reference Sapiro, Walsh, Strach and Hennings2011) in an electoral setting in which parties are not polarized and the share and visibility of women candidates suddenly increases? Are previous findings regarding women's greater reluctance to share personal information in their self-presentation (Stalsburg and Kleinberg Reference Stalsburg and Kleinberg2015) and their stronger emphasis on political experience (Fridkin and Kenney Reference Fridkin and Kenney2014) confirmed in such a setting? The following section outlines why Ireland's 2016 general election is a particularly suitable case to answer these questions.
POLITICAL CAMPAIGNING IN IRELAND AND THE GENDER QUOTA
The Irish political system is characterized by high levels of party discipline in parliament but low party coherence outside it (Farrell, Suiter, and Harris Reference Farrell, Suiter and Harris2017; McGraw Reference McGraw2016). The electoral system combines constituencies with three to five seats and a preference vote (single transferable vote) that allows the ranking of all candidates (Sinnott Reference Sinnott, Coakley and Gallagher2010). This encourages the cultivation of a personal vote based on constituency work (Carty Reference Carty1981; Gallagher and Komito Reference Gallagher, Komito, Coakley and Gallagher2010; O'Leary Reference O'Leary2011), as well as running as an independent candidate (Bowler and Farrell Reference Bowler, Farrell, Marsh, Farrell and McElroy2017). The major parties pursue a catchall approach that makes them very similar in terms of mean party placement between left and right. They produce party manifestos for national elections that highlight key issues, but candidates have great leeway to stress or ignore them in their local campaigns (McGraw Reference McGraw2016). Candidates from the same party therefore take very different positions, depending on the constituency in which they run. The combination of low interparty and high intraparty differences in ideological positioning is a context in which partisanship is unlikely to marginalize gender (Hayes and Lawless Reference Hayes and Lawless2016) as a potential influence on the self-presentation of political candidates.
The share of female candidates for the Dáil did not significantly increase in the 1990s and 2000s. It fluctuated between 15% and 20% and even showed a downward trend from 1997 to 2011 (Buckley et al. Reference Buckley, Mariani, McGing and White2015, 313). In 2016, it rose to almost 30%, following the introduction of a legislative gender quota.
The general election in 2016 was the first election at the national level in Ireland in which parties were required to ensure that at least 30% of their candidates were female. The quota is backed by severe financial sanctions if parties do not comply. Introduced as part of the Electoral (Amendment) (Political Funding) Act 2012, the new rule means that public payments to parties “shall be reduced by 50 per cent, unless at least 30 per cent of the candidates whose candidatures were authenticated by the qualified party at the preceding general election were women and at least 30 per cent were men” (Section 42). Although the rule is phrased in a gender-neutral way, the problem that the act addresses is the notoriously low share of female representatives (Teachtaí Dála or TDs) in the Dáil and the low presence of women in Irish politics in general. The act is a public acknowledgment that the presence of women in politics is not sufficiently normalized to leave the selection of female candidates for general elections unregulated.
The quota was introduced rather suddenly and without a prolonged debate, although the underrepresentation of women in politics had long been recognized as a problem. All parties had made attempts to improve women's representation within their ranks, but the promotion of gender equality in elections had been mostly rhetorical or based on self-imposed targets that were missed in practice (Buckley Reference Buckley2013). A legislative quota appeared on the agenda in a context that stressed a commitment to political reform in the aftermath of the financial crisis. When the devastating fiscal consequences of the crisis led to the election of a coalition of the long-term opposition parties Fine Gael and Labour in 2011, the new government emphasized a need for fundamental political change. In this context, the adoption of a gender quota symbolized a break with past Irish politics. Even the parliamentary party of Fianna Fáil, which had been in government since 1997 and was blamed for the financial crisis, supported the gender quota. At the same time, many ordinary members of this but also other parties rejected it (Buckley Reference Buckley2013), and general support for gender quotas was only moderate (Keenan and McElroy Reference Keenan and McElroy2017).
In such circumstances, a decoupling of formal rules and informal practices can be expected (Brunsson Reference Brunsson1989; Meyer and Rowan Reference Meyer and Rowan1977). Parties could nominate enough women to formally meet the quota but select inexperienced candidates or concentrate female candidates in constituencies where they are unlikely to win a seat (Buckley, Mariani, and White Reference Buckley, Mariani and White2014, 475). Buckley, Galligan, and McGing (Reference Buckley, Galligan, McGing, Gallagher and Marsh2016, 191), however, note that the majority of female party candidates were chosen in selection conventions at the local level and not imposed by a directive of the party headquarters, which suggests that many party members regarded them as qualified candidates.
Yet even if party leaders and members embraced the chance to increase the share of women in politics, the sudden imposed requirement to nominate more female candidates creates a context that is very different from one of gradual normalization. On the one hand, female candidates were likely to be less restricted by established patterns of politics in Ireland than in previous elections, because of the circumstances in which the gender quota had been introduced. Since the turn to a legislative quota had been rather sudden, parties had little time to make up for insufficient efforts of the past to significantly increase the number of electable female candidates. Moreover, the fallout of the financial crisis had created doubts about the quality of the Irish political system (e.g., Carey Reference Carey2010; Collins Reference Collins2011; O'Toole Reference O'Toole2011). The strict austerity of the outgoing government had alienated many voters and exacerbated many long-known problems; apart from the economy, underfunding in many policy fields, especially health care, was high on the public agenda (Costello, O'Neill, and Thomson Reference Costello, O'Neill, Thomson, Gallagher and Marsh2016). Established expectations about what voters wanted from political candidates were no longer a matter of course. On the other hand, uncertainty generally encourages a recourse to stereotypes (van den Berghe Reference van den Berghe1997). Moreover, the newly introduced gender quota made the candidates’ gender a particularly salient category. Therefore, female candidates are likely to have paid special attention to the possible impact of gender stereotypes in the electoral campaign, which may have resulted in gender-specific choices regarding their self-presentation.
The potential relevance of gender stereotypes is especially plausible considering the central role that a traditional notion of the family has played in Irish politics until the recent past, even backed by the Irish Constitution (Article 41.2) and its enshrinement of the traditional female gender role as the caregiving mother. On the one hand, the relevance of this notion has decreased (Murphy-Lawless Reference Murphy-Lawless2000; O'Sullivan Reference O'Sullivan2012a), which is exemplified by the vote in favor of same-sex marriage in a referendum less than a year before the 2016 general election. On the other hand, a gendered division of labor continues to affect women's choices and responsibilities (Lynch, Baker, and Lyons Reference Lynch, Baker and Lyons2009; O'Sullivan Reference O'Sullivan2012b). Data from the Organisation for Economic Co-operation and Development indicate that the traditional notion of family and the gender roles it implies continue to be relevant (OECD n.d.). Ireland's fertility rate of 1.9 is one of the highest in Europe. Until today, divorces, which were only legalized in 1997, have remained at a very low level compared with other European countries. The employment rate of mothers who live with a partner is lower than in most European countries. At the same time, the amount of unpaid work that Irish women do is among the highest.
Based on the research on gender differences in self-presentation and characteristics of the Irish political setting, I expect differences in the issues that male and female candidates prioritized. The 2016 electoral campaign took place in a context in which a share of around 30% female candidates was not normal and in which parties were not polarized in terms of policy. Therefore, it seems likely that there were gender-specific issue priorities congruent with stereotypical female, care-related strengths (H 1).
I also expect differences between men and women with regard to the extent to which candidates addressed their political and personal background. I expect that women more often than men mentioned their political experience to stress that they were qualified for political office (H 2). Furthermore, I expect that men addressed aspects of their personal background more often than women (H 3). In addition, I look specifically at the role of family as a part of this personal background. I expect differences in the extent to which male and female candidates referred to their family (H 4). The hypothesis is nondirectional since the literature is not conclusive with regard to which gender tends to emphasize family more. Some effect is likely, however, considering that the traditional notion of family still seems to have considerable normative relevance for Irish women's life choices.
EMPIRICAL MATERIAL AND METHODS
The empirical analysis uses video statements of candidates during the run-up to the Irish general election in 2016. The Irish public broadcaster RTÉ had invited all candidates in the election to record a short statement. When the election was called, the videos were published on a website (http://www.rte.ie/news/election-2016/candidates). Videos are available for 488 of the 550 candidates who ran in the election. The website provided short profiles of the candidates who did not record a statement.
The material has unique, hybrid characteristics compared with the three main remote communication channels—media coverage, television ads, and online sources—through which voters may learn about political candidates (Bystrom Reference Bystrom, Carroll and Fox2018). The website was part of the media coverage of the election, and RTÉ decided the videos’ format, camera angle, and maximum length of one minute. Unlike with campaign ads, the candidates did not have full control over the nonverbal content and production content (Bystrom et al. Reference Bystrom, Robertson, Banwart and Kaid2004, 30–31). The camera was placed in front of the sitting candidates and remained static; the background was bluish and displayed a pattern with the RTÉ logo. However, candidates had full control over the verbal content of their statement within the time constraint. The actual length of the videos varies between 29 and 62 seconds; the average length is 55 seconds. Since all videos were published simultaneously after the election was called, there are no temporal effects (Windett Reference Windett2014) that could influence possible gender differences.
The material makes it possible to compare the self-presentations of a broad range of candidates, including many independents (see overview in Table 1).
Table 1. Empirical material and share of women candidates in the election
* See Buckley, Galligan, and McGing (Reference Buckley, Galligan, McGing, Gallagher and Marsh2016, 188).
A check revealed that the group of candidates without a video includes a disproportionate number of candidates with no prior political experience at the national or local level (62.3% as opposed to only 39.1% among the candidates with a video). However, the share of women in this group of candidates is only 20%, which is lower than the share of women without prior political experience among all the candidates (34%). It is unclear whether more women without political experience saw the video as an attractive campaign medium or whether RTÉ made a special effort to include as many female candidates as possible. In any case, the lack of videos for a number of candidates is unlikely to affect the analysis of gender differences in how candidates presented themselves in these videos.
The strong focus of Irish electoral campaigning at the constituency level may raise the question of how relevant the videos were compared to other campaign efforts and consequently how informative the data are that they provide. It is likely that candidates valued the opportunity to record a video statement differently, depending on their previous political experience and public exposure. Even for unknown candidates, the video was not necessarily the main channel for addressing potential voters, considering the importance of canvassing and personal contact in the constituency.
At the same time, there are several reasons to believe that candidates regarded the video statement as a welcome medium for their campaign message. First, almost 90% of candidates decided to record one. Second, the format left it entirely to the candidates to decide whether they wanted to use the video to remind viewers of their past work for the constituency, refer to pressing matters at the level of the constituency, or talk about issues relevant for the whole country. The videos’ placement on a website that sorted them according to constituency as well as party meant that the actual audience was unknown but likely to be predominantly constituency specific. Third, Irish politicians have made ample use of local radio since the late 1990s and seen huge benefits in being present in a medium that is free for both the candidates and their audience, which largely comes from their constituency (Kavanagh Reference Kavanagh, O'Brien and Beacháin2014, 48–50). Fourth, political parties had already experimented with video messages on their campaign websites in the 2011 general election (Molony Reference Molony, O'Brien and Beacháin2014).
Considering these points, there is no reason to think that candidates who chose to record a video did not carefully consider what they wanted to say. The strict time limit certainly forced them to be selective with regard to content, but similar restrictions are typical for political ads on television, which have frequently been used in research on gender differences in the self-presentation of candidates.
The content analysis of the statements is based on their written transcripts. Passages in Irish were not transcribed for practical reasons. Most candidates who used Irish did so only briefly and no one spoke exclusively Irish, which is why it seems acceptable to analyze only English-language content. The content was coded using a mix of deductive and inductive categories. The categories relevant for the hypotheses are the policy issues that the statements addressed, whether the statements referred to previous political experience, personal CV, and the family of the candidates.
The data are based on statements from 88.7% of election candidates. Conventional measures of statistical analysis, such as significance levels, do not therefore have a probabilistic meaning for this material. Especially with regard to multivariate analyses, they can be interpreted as a general measure of robustness of the results. Apart from gender as an independent variable, I included party and political status as control variables, with Fianna Fáil (as the hegemonic party until the financial crisis) and incumbents used as the reference categories.
RESULTS
Before looking at the findings for the four hypotheses, it is important to establish whether the actual election results suggest that female candidates were overall less successful than male candidates in winning a seat in parliament. If that were the case, it could indicate that women were disproportionately token candidates for the parties, which means that gender differences in self-presentation could originate from a perceived hopelessness of winning. In fact, 31.7% of all male candidates were eventually elected, but only 21.6% of all women. However, once we control for political status and party, the impact of gender on success is negligible compared with the other two variables (Table 2). This indicates that the lower success rate is not an effect of a concentration of female candidates in races with unwinnable seats.
Table 2. Influence of gender, party, and political status on electoral success
Note: Coefficients from binary logistic regression model with standard errors in parentheses.
The video statements address a broad range of issues, but many of them are mentioned by only a few candidates, often from small parties or running as independents. For example, only 35 candidates address the controversial issue of abortion—or, more specifically, whether an amendment to the Irish Constitution that effectively banned abortion should be repealed. None of those candidates belonged to one of the four major parties. Table 3 lists the 10 most frequently mentioned issues as well as the number and share of male and female candidates who referred to them.
Table 3. Gender and frequency of references to the 10 most important issues
The literature that has found gender differences in priorities (Fox Reference Fox1997; Kahn Reference Kahn1993; Schneider Reference Schneider2014) would suggest that three of these issues (health care, education, child care) tend to be emphasized by women and five of them (job creation, taxes and charges, economy, business incentives) by men. The only gender difference in the data that is in line with these assumptions concerns child care, to which a higher percentage of female than male candidates refer in their statements. There is also a difference with regard to health care, but with more men than women addressing the issue, it is the opposite of what the literature suggests. Another issue that men mention more frequently than women is infrastructure. There are no clearly discernible gender differences for the other seven issues. Checking how many of the 10 issues the statements mention on average reveals a negligible difference between women (2.32) and men (2.49). All in all, there is very limited support for H 1.
Table 4 compares the frequency of references to political experience (both formal roles and grassroots activities) and personal background or CV in the videos of male and female candidates. There is no gender difference when it comes to mentioning previous political experience; about one-third of the candidates of both genders point to it. H 2 is therefore not supported. For male candidates, references to their personal background or CV are less frequent than references to political experience. The opposite is true for women, since they give details of their personal background much more frequently than men. Almost half of the women but only about one-quarter of the men address their personal backgrounds in some way. This is contrary to what H 3 predicted.
Table 4. Gender and frequency of information on political and personal background
The higher share of women using personalization is due to three types of information, which they give more frequently than men. The first is information about a spouse and/or children; the second is information about previous or current occupations and career; the third (less pronounced than the other two) is information about their current place of residence or the place and circumstances in which they grew up. A binary logistic regression that controls for party and political experience confirms that gender has an independent effect on the likelihood of candidates addressing their personal background, both when references to a spouse and/or children are included in this category and when they are left out (Table 5). Female candidates are almost twice as likely as men to personalize their statements, even when controlling for political status and party.
Table 5. Influence of gender, party, and political status on personalization
Note: Coefficients from binary logistic regression model with standard errors in parentheses.
* p <.05.
It is important to note that the lack of experience at the national level has an even larger effect than gender. Local councilors are more than six times more likely to refer to their personal background, and candidates with no prior formal political role more than nine times more likely than incumbents to refer to personal backgrounds. Membership in certain parties also significantly impacts the likelihood of candidates to refer to their personal CV. Nevertheless, there is a clear gender effect and the more frequent personalization by female candidates contradicts H 3.
H 4 focuses on a special aspect of personalization, namely, references to a candidate's family. It is only meaningful for candidates who have a family. Unfortunately, it was not possible to establish the family status of all election candidates to distinguish between candidates who chose not to mention their family and candidates who did not have one. However, I was able to gather these data for 131 of the 144 successful candidates who recorded a video (Table 4). In this group, there is almost no difference in the share of men (83.2%) and women (80.0%) who are married. When it comes to children, the picture is a little different, with 90.0% of the women being mothers and only 77.2% of the men being fathers. It is likely that the actual share of childless candidates is somewhat higher than these numbers suggest since it may be precisely such candidates for which it was not possible to find public information on the matter.
Focusing on the 117 cases of successful candidates who actually were married and/or have children, the choice to mention this fact in the video statement is clearly gendered (z-test significant at the 0.05 level): 29.6% of the women in this group mention a spouse and/or children in their statement, but only 13.3% of the men. This finding supports H 4.
DISCUSSION
The analysis of candidates’ issue priorities found few gender-specific differences, and only one of them conforms to the literature that predicts and finds such differences (Huddy and Terkildsen Reference Huddy and Terkildsen1993; Kahn Reference Kahn1993). Women mention child care more frequently than men, but the opposite is the case for health care, the second of the top-10 issues that is clearly care-related and stereotypically linked to women. Health care was an important issue in the election (Costello et al. Reference Costello, O'Neill, Thomson, Gallagher and Marsh2016). Austerity had exacerbated the problem of hospital trolley waits, which had been a very visible indicator of failure of the health care system for a long time (O'Ferrall Reference O'Ferrall2009, 157), and the incumbent government had backtracked on a commitment to universal health coverage, a recurring topic in Irish politics for many decades (Darker, Donnelly-Swift, and Whiston Reference Darker, Donnelly-Swift and Whiston2018, 147–48). In a way, the issue of health care therefore had both a stereotypical female aspect, related to compassion and care, and a stereotypical male aspect, related to financing. Although the presence of both aspects suggests that health care is not necessarily a compassion issue in the Irish context, it remains unclear why it should be a particularly male issue instead. Consequently, it remains an open question why men addressed health care more frequently than did women.
The supposedly female issue of education is similarly important to male and female candidates, which leaves child care as the only issue for which the results show a stronger presence in the self-presentations of female candidates. Among the 10 top-priority issues in the video statements, it is the one most closely related to the question of gender roles in the family. Considering the elements of tradition that are still present in Irish society, the provision of child care facilities and other kinds of support is likely to be the one issue among the top ten that affects women more directly in their daily lives than men (Devitt Reference Devitt2016). The more pronounced focus of female candidates on this issue may have less to do with an attempt to respond to stereotypical expectations about female politicians than with the gendered division of labor in the family (Lynch, Baker, and Lyons Reference Lynch, Baker and Lyons2009; O'Sullivan Reference O'Sullivan2012b) and the perceived need to stress an issue that is still more important to women than to men.
The more pronounced focus of men on infrastructure could tentatively be interpreted as conforming to a gender stereotype that sees men as agentic (Eagly and Steffen Reference Eagly and Steffen1984) and more entrepreneurial than women (Gupta et al. Reference Gupta, Turban, Arzu Wasti and Sikdar2017). The results regarding health care, however, suggest that such ascriptions can at best play a minor role in determining men's issue priorities. Overall, the findings are in line with Dolan's (Reference Dolan2014) observation that gender differences in issue priorities are limited and do not always correspond to expectations based on gender stereotypes.
Contrary to expectations based on the literature (Banwart and Winfrey Reference Banwart and Winfrey2013; Fridkin and Kenney Reference Fridkin and Kenney2014), women did not refer to their political experience more often than men. Since there were fewer female than male candidates who had experience in a formal political role, a comparison of references to political experience only makes sense when it includes informal political engagement. About one-third of the candidates of both genders mentioned political experience in that sense, be it as a TD, local councilor, assistant to a TD, or grassroots activist. There is thus no indication that women felt more need to justify their candidacy and show their qualification for political work. Many candidates, whether male or female, clearly considered previous political experience as important when justifying their candidacy. At the same time, women apparently considered aspects of their personal life as even more important.
Women candidates gave information on their personal background much more often than men. Controlling for political experience and party reduces the effect but does not erase it. All else being equal, female candidates were 90% more likely than men to include such an element of personalization in their statements. The candidates for which their family status could be ascertained showed the same pattern, with women offering information about their spouse and/or their children much more often than men.
The overall picture is almost the opposite of what the literature, drawing on data from the United States, has suggested (Fox 1997; Fridkin and Kenney Reference Fridkin and Kenney2014; Stalsburg and Kleinberg Reference Stalsburg and Kleinberg2015). Instead of women candidates who emphasize their own political experience and are reluctant to share information about their family and to personalize their message, the study of Irish candidates finds women who personalize and mention their family much more than men and refer to previous political experience only as frequently as men.
These results indicate that women were not afraid that references to their personal background or their family life would make them appear less competent or less suitable for political office, as other studies (Bystrom et al. Reference Bystrom, Robertson, Banwart and Kaid2004; Fridkin and Kenney Reference Fridkin and Kenney2014) would suggest. The fact that female candidates put more emphasis on their family than men suggests the contrary—that many expected a positive effect from a self-presentation as wife and mother. This pattern in the data points in the same direction as the more pronounced focus on the issue of child care that female candidates displayed: women expect the traditional notion of family, and more specifically, the role of women as caretakers in the family, to play a decisive role for the way in which they are perceived as electoral candidates. They do not seem to worry that the emphasis on the stereotypical female trait of caregiving could hurt their campaign.
Yet perhaps they do worry, and their disproportionate willingness to give information on occupation and professional career is a reaction to the perceived threat of being restricted to the stereotype of caregiver. The higher frequency, compared with men, with which women address their occupation or professional career could be a result of trying to avoid this trap. Yet a strong emphasis on previous political experience (Fridkin and Kenney Reference Fridkin and Kenney2014) would seem the more obvious means to this end. Moreover, the group of women candidates who mention family and those who mention career only partly overlap. Both points suggest that references to career do not merely fulfill a compensatory function.
The slightly more frequent use of references to a specific place of residence or origin that women candidates display could be interpreted as a special emphasis on ties to the community and thus in line with previous findings (Fox Reference Fox1997). However, considering that the political culture in Ireland creates strong incentives for all candidates to stress their local pedigree (Gallagher and Komito Reference Gallagher, Komito, Coakley and Gallagher2010), this interpretation is not entirely convincing.
Alternatively, references to occupation and a place of living or upbringing can be seen as part of a self-presentation that stresses individuality. It is possible that women candidates in the 2016 election perceived particular advantages of such personalization against the backdrop of the gender quota. Considering its recent and sudden introduction, female candidates may have feared being perceived as women above all else and may have tried to use personalization as an antidote to being perceived in terms of a category (Schutz Reference Schutz1972; van den Berghe Reference van den Berghe1997).
In sum, the self-presentation of female candidates does not indicate that they were worried about being perceived as less competent than men or felt the need to restrict themselves to issues that are congruent with supposedly gender-specific strengths. The major exception is aspects directly related to the family. The stronger overall personalization of female candidates differs from previous findings (McGregor, Lawrence, and Cardona Reference McGregor, Lawrence and Cardona2017; Meeks Reference Meeks2017) in the United States. It could be an effect of the lack of normalization of women's candidacies (Hayes and Lawless Reference Hayes and Lawless2016), which the introduction of the gender quota had highlighted. In these circumstances, many female candidates may have felt the need to both conform to the traditional notion of women as caregivers by referring to their own family and compensate for this. Alternatively (or concomitantly), women may have been concerned that the gender quota itself would encourage voters to perceive them as members of a category and may have used personalization to stress their individuality.
CONCLUSION
This study has analyzed video statements of almost 90% of the candidates who ran in the Irish general election 2016 and looked for gender differences in candidates’ self-presentations. It confirms more recent findings, which mostly come from research on candidates in the United States, that the issue priorities of male and female candidates are similar in most regards. The one exception that conforms to gender stereotypes is the stronger focus of female candidates on child care. It corresponds with women's more frequent references to their own family and suggests that female candidates in Ireland perceive the need to acknowledge the continuing normative power of a notion of family that ascribes the role of caretaker to women.
At the same time, women did not make more effort than men to present themselves as qualified for political work, which suggests that they were not worried about being perceived as less competent. Contrary to what other research has found, female candidates personalized their self-presentation much more often than men, not only by mentioning a spouse and/or children but also by giving information about their professional career and their place of residence or upbringing. Women candidates may partly have intended to counterbalance the effect of mentioning their own family by also referring to a previous or current occupation, trying to avoid a stereotype that would limit them to the caregiver role. At the same time, women candidates’ strong personalization may also be related to the fact that the 2016 election was the first in which a legislative gender quota required parties to nominate at least 30% female candidates.
An important limitation of this study is that it cannot actually determine the effect of a recently introduced gender quota on candidates’ self-presentation. This would require a comparison with other elections, for which comparable data are not available. Moreover, the empirical material used covers only one aspect of campaign communication, and one that is novel in Ireland, where canvassing and face-to-face contact with voters play a large role in campaigning. At the same time, the material makes it possible to compare self-presentations of a large number of candidates, which would otherwise hardly be possible, and there is no reason to think that candidates presented themselves completely differently on video than in other parts of their campaign.
As the first study to analyze gender differences in candidates’ self-presentation in Ireland, this research offers important new insights into the role of gender in Irish politics at a critical point in time, when the gender quota has raised public awareness of women's underrepresentation. The findings also suggest that a predominant focus on the United States, which has characterized studies of gender differences in campaigning that include many candidates from the same election so far, is a serious limitation. The field will benefit from conducting similar studies in other countries, where—as this study shows—results may contradict established wisdom about the existence and patterns of such differences.
In the Irish case, research on gender differences in future electoral campaigns would be extremely valuable for tracking the impact of the gender quota over time. Ideally, a project on a much larger scale would follow the example of studies in the United States (Dolan Reference Dolan2014; Hayes and Lawless Reference Hayes and Lawless2016) and combine the analysis of candidates’ self-presentation with an analysis of gender differences in media coverage and voter perceptions. The findings in this article are a first step in that direction.
