Voter Perceptions of LGBTQ Politicians

Public opinion towards LGBTQ rights has become increasingly positive in Western democracies (Ayoub and Garretson Reference Ayoub and Garretson2017), contributing to the expansion of LGBTQ rights provisions, including widespread legalization of same-sex marriage. More LGBTQ individuals have also been elected to political office (Reynolds Reference Reynolds2013). More recently, LGBTQ politicians have risen to the highest positions of political leadership, including former Irish Taoiseach and Fine Gael leader Leo Varadkar, who stepped down in April 2024.

Despite the successes of the LGBTQ rights movement, studies have found that LGBTQ politicians are still penalized by voters (Haider-Markel Reference Haider-Markel2010; Magni and Reynolds Reference Magni and Reynolds2021). This penalty is stronger for religious and conservative voters and for voters without LGBTQ friends or family members (Doan and Haider-Markel Reference Doan and Haider-Markel2010; Magni and Reynolds Reference Magni and Reynolds2021). In part, hostility to LGBTQ candidates comes from the prejudice some voters have towards the LGBTQ community (Magni and Reynolds Reference Magni and Reynolds2021). Beyond outright prejudice, LGBTQ candidates also face electability concerns, even among voters generally supportive of LGBTQ rights (Magni and Reynolds Reference Magni and Reynolds2021). Both conservative hostility to LGBTQ candidates and electability concerns can be partially explained by the relationship between LGBTQ identity and ideology. LGBTQ candidates are often viewed as more left-leaning than their heterosexual counterparts (Golebiowska Reference Golebiowska2003; Magni and Reynolds Reference Magni and Reynolds2021). Thus, sexual orientation and gender identity can serve as a heuristic for voters when determining a candidate’s ideology, with some voters perceiving LGBTQ candidates as too far left to win office.Footnote ²

Recent work, however, has suggested that the penalty incurred by LGBTQ candidates may not be as significant as portrayed above. In analyzing the 2015 UK general election, Magni and Reynolds (Reference Magni and Reynolds2018) found that LGBTQ candidates generally did not have a negative impact on party vote share and performed at least as well as their heterosexual counterparts. Similar results were found in Canada (Everitt and Horvath Reference Everitt and Horvath2021) and the United States (Haider-Markel et al. Reference Haider-Markel, Patrick Gauding, Lewis, Miller, Tadlock and Taylor2020). Finally, Loepp and Redman (Reference Loepp and Redman2022) found that the negative effect of a candidate’s sexual orientation significantly decreases when voters are given information about the candidate’s partisanship. Party identification, then, may trump concerns that arise due to a candidate’s identity.

Although the studies highlighted above provide insight into how voters evaluate gay candidates for legislative office, we still know little about how voters evaluate gay politicians in executive leadership positions. Two studies, one from the United States and one from Canada, provide some initial insights. In their study, Magni and Reynolds (Reference Magni and Reynolds2024) found that voters penalized Democratic presidential candidate Pete Buttigieg for being in a same-sex relationship, although emphasizing his military background mitigated this penalty. Everitt and Camp (Reference Everitt and Camp2009) found that media coverage of Allison Brewer, the openly lesbian leader of New Brunswick’s New Democratic Party, was dominated by stereotypes of lesbians and women in politics. Beyond these studies, however, we do not know much about how voters evaluate gay leaders.

Existing scholarship has shown that voters prefer different traits for politicians in executive office and those in local and legislative office, with voters tending to prefer that legislators possess more communal traits that can facilitate coalition-building and consensus, while executives possess more agentic traits such as assertiveness and decisiveness (Sweet-Cushman Reference Sweet-Cushman2022). Given the increasing presidentialization of party politics (Poguntke and Webb Reference Poguntke and Webb2005) and the personalization of party leadership in Europe (Blondel and Thiebault Reference Blondel and Thiebault2010), where party leaders dominate the campaigns, elections, media coverage, and governance, and where they serve as the national face of the party, in both the electorate and the legislature as they set the party agenda and manage party operations (Garzia et al. Reference Garzia, Ferreira Da Silva and De Angelis2022), it is critical to know how voters react to LG party leaders. Thus, while candidates for legislative office need only compete in one (likely strategically chosen) district, party leaders are meant to oversee the national campaign and represent the party in each constituency, regardless of its ideological bent. While parties may be comfortable running LG candidates in ‘safe seats’ to mitigate any voter bias, selectorates (whether politicians, activists, or party members) may be reluctant to elevate an LG politician to the leadership post so as not to alienate needed voters. Therefore, while MPs win their seats in an election, party leaders are said to have earned the role through their strong commitment to the party organization, legislative successes or robust legislative agendas, and, typically, significant tenure in office. Indeed, party leaders are meant to maintain and build the party organization, oversee party committees and branches, work with grassroots activists and organizations, manage the party whip, and manage and build the party’s resources well beyond what could be expected of a rank-and-file MP. Finally, party leaders are responsible for managing a national legislative agenda for the parliamentary faction and shepherding legislation through the parliament.

Recent research also reveals that LG leaders are often depicted in the media differently from straight leaders because their identity as gay is always part of the narrative (Lalancette and Tremblay Reference Lalancette and Tremblay2019), potentially contributing negatively to the broader narrative of their fit for the top political offices. For instance, former Prime Minister Leo Varadkar faced characterization as an ‘other’, with critiques regarding his suitability as a leader due to his identity as an Irish Indian and his sexuality. While international media framed him as a progressive figure, the Irish media forwarded narratives undermining his legitimacy through anti-gay stereotypes (Kerrigan and Pramaggiore Reference Kerrigan and Pramaggiore2021). Similarly, a study of Canadian Premier Kathleen Wynne by Everitt et al. (Reference Everitt, Raney and Tremblay2019) noted that, despite her high profile among voters, she could not fully escape the stereotypes linked to her being an openly gay politician, facing media portrayals that questioned her electability and whether Ontario was ready for a lesbian leader (McLean Reference McLean2019). In contrast, former Icelandic Prime Minister Jóhanna Sigurðardóttir, the world’s first openly gay prime minister, was primarily portrayed as gay only at the start of her term, with her gender identity serving as a more dominant narrative throughout her leadership (Mundy Reference Mundy2013).

We also know that while there are many examples of LG legislators, there have been relatively few openly LG party leaders. Thus, stereotypes may play an outsize role in citizen perceptions of LG leaders. Outside political science, recent scholarship on LGBTQ CEOs found that the companies they lead are persistently undervalued due to investor discrimination and that LGBTQ individuals are more likely to represent small growth stocks with poor past performance (Shanaev et al. Reference Shanaev, Skorochodova and Vasenin2023).Footnote ³ Might LG leaders face similar perceptions in the political arena?

Ultimately, given the unique role of party leaders, we argue that LG leaders are likely to suffer penalties as compared to heterosexual leaders that go beyond what is experienced by LG legislators. LG party leaders, who are more widely covered in the media, run the electoral campaigns, and dominate the legislative politics in parliamentary systems, may be seen as less likely to have earned their position due to perceptions that they were selected in an attempt to increase diversity, rather than selected based on merit. They may be perceived as less able to unify the party given expectations of homophobia among more conservative factions within the party or perceptions that an LG leader is farther left ideologically than the mainstream of the party. Rank-and-file legislators, however, are often not expected to perform this type of unifying role. Similarly, given that leaders are the face of the party nationwide, rather than simply the party’s representative in a ‘safe’ constituency, they may be perceived as less able to help their party win seats in a subsequent election due to perceptions of their ideological leanings as well as expectations of homophobia among voters. Finally, LG party leaders may be perceived as less effective in advancing the party’s legislative agenda due to anticipated homophobia among other political elites and stereotypes that LG politicians will be overly focused on ‘LGBTQ issues’ (Herrick and Thomas Reference Herrick and Thomas1999; Golebiowska Reference Golebiowska2002).

At the same time, one cannot examine the hostility towards LG party leaders without assessing the intersectional role of gender and gender-based stereotypes. The stereotypes of women as compassionate and empathetic and men as assertive and confident (Hedlund et al. Reference Hedlund, Freeman, Hamm and Stein1979; Alexander and Andersen Reference Alexander and Andersen1993) suggest that men, but not women, are often linked to leader-like behavior (Eagly et al. Reference Eagly, Makhijani and Klonsky1992). More specifically, women are often stereotyped as possessing more communal, rather than agentic, traits, making them more suitable for legislative, rather than executive, office in the eyes of voters (Sweet-Cushman Reference Sweet-Cushman2022). Furthermore, studies have found that heterosexual women perceive lesbians as more masculine and heterosexual men perceive gay men as more feminine (Kite and Whitley Jr. Reference Kite and Whitley1996; LaMar and Kite Reference LaMar and Kite1998; Herek Reference Herek2002). In separating voter evaluations of gay and lesbian congressional candidates, Doan and Haider-Markel (Reference Doan and Haider-Markel2010, p.85) found that ‘gay men are feminized and lesbians are masculinized by heterosexuals’, although their analysis cannot determine if lesbians are helped or gay men are hurt by these stereotypes.Footnote ⁴ Finally, Doan and Haider-Markel (Reference Doan and Haider-Markel2010) found that gay men were more likely to face negative stereotypes than gay women, mainly due to the lower likelihood of heterosexual women negatively stereotyping lesbians. If voters perceive gay men as possessing more feminine traits while gay women are seen as possessing more masculine traits, gay men may receive a stronger penalty than gay women when serving as party leader. Therefore, given the divergent expectations for legislators and executives, and the potential feminization of gay men, the penalty for gay men as leaders may be greater than the penalty found in the literature on LG candidates for legislative office.Footnote ⁵

Despite our expectations that LG leaders, especially men, are negatively evaluated by the public, we also note that context matters for activating stereotypes (as shown by Dolan [Reference Dolan2014] for voter perceptions of women candidates for office). When other cues send voters signals about candidate quality, such as incumbency or party ID, voters may rely less on stereotypes when assessing candidate competence. We therefore argue that voters may base their evaluation of the leader’s competency more on experience in office than on stereotypes of LG political leaders, if a new leader has considerable legislative experience.

Our hypotheses, therefore, are as follows:

• ${H_1}$ : Straight leaders will be seen as more deserving, competent, and electoral viable (that is, have higher evaluations) than LG leaders.

  1. – ${H_{1a}}$ : A straight male leader has higher evaluations than a gay leader.

  2. – ${H_{1b}}$ : A straight female leader has higher evaluations than a lesbian leader.

  3. – ${H_{1c}}$ : A gay leader has lower evaluations than a lesbian leader relative to their straight counterparts.

• ${H_2}$ : The gap between evaluations of LG and straight leaders narrows when LG leaders have more legislative experience.

In assessing citizen perceptions of LG party leaders, we focus on a leader’s perceived deservingness, effectiveness in the various aspects of the position, and electoral viability. Beyond simply capturing the leader’s favorability or popularity, we assess various factors that contribute to individual perceptions of a leader’s performance in the role. As mentioned above, party leaders are expected to effectively manage the party organization, legislative caucus, and electoral campaign. Party leaders, then, are likely to receive higher leadership evaluations when citizens feel they have earned their position and will work hard to maintain and build the party apparatus. Additionally, party leaders are likely to receive higher evaluations when they are seen as contributing to – and responsible for – the party’s legislative and electoral successes. Assessing party leadership evaluations along these dimensions thus provides a more comprehensive assessment of an individual’s perception and evaluation of a party leader. Additionally, not only do evaluations of party leaders along these dimensions depend on the qualities of the leader themselves, they also likely induce individuals to consider potential reactions to LG party leaders from the public and other political elites. As mentioned above, some of the bias faced by LG candidates for legislative office has been linked to concerns that they are unable to appeal to a broader electorate. We further clarify how we measure these dimensions in the next section.

Study Design

We test our hypotheses through a conjoint survey experiment conducted in the United Kingdom in 2023 by the survey firm Bilendi.Footnote ⁶ Our sample consists of 1,198 respondents from Bilendi’s online panel and was designed using quotas to match the population distribution of gender, age, and region.Footnote ⁷ Although none of the major political parties has had an openly LG leader, surveys show that citizens have increasingly favorable views towards homosexuality, all major parties support both marriage equality and anti-discrimination legislation, and over 150 LGBTQ candidates ran for office in the 2015 election (Magni and Reynolds Reference Magni and Reynolds2018). This relatively favorable view of the LGB community suggests that the United Kingdom is a more difficult test for our hypotheses when compared to countries with fewer rights protections and less descriptive representation for the gay community.Footnote ⁸

Upon agreement to participate in the study, respondents were asked to answer a standard battery of demographic questions. Respondents were then informed that they would see five pairs of hypothetical political parties and their recently elected leaders and would be asked about their attitudes towards the party and its leader.Footnote ⁹ Following recent scholarship on conjoint designs, the first four tasks in our experiment consisted of classical conjoint tasks (with attribute levels randomized for each profile), while the fifth task repeated the first task, flipping the order of the profiles, serving as a way to account and correct for intra-respondent reliability (Clayton et al. Reference Clayton, Horiuchi, Kaufman, King and Komisarchik2023).Footnote ¹⁰ Therefore, our design should be conceived as a four-task conjoint, a reasonable number with likely minimal impact on respondents’ survey satisficing (Bansak et al. Reference Bansak, Hainmueller, Hopkins and Yamamoto2018).

Among other traits, the profiles randomly presented to participants varied the recently elected party leader’s gender, sexual orientation, and previous political experience.Footnote ¹¹ After being presented with each pair of profiles, we assessed respondents’ evaluations of the hypothetical party leaders. Specifically, we focused on a leader’s deservingness, competency in managing the party organization and legislative caucus, and electoral viability. We measured leadership evaluations on these various dimensions using a five questions battery:

1. 1. Which of these leaders earned their position?

2. 2. Which of these leaders would be more effective in passing legislation?

3. 3. Which of these leaders will work harder on behalf of their party?

4. 4. Which of these leaders would be more effective in unifying the party?

5. 5. Which of these leaders would help their party win more seats in the next election?

Thus, in going beyond existing studies of candidates for legislative office, our questions focus on the specific responsibilities of party leaders, including their role in unifying the party, in crafting a successful legislative agenda, and in helping the party grow its parliamentary faction in subsequent elections. As described in the previous section, there is good reason to believe that LG leaders will suffer additional penalties across these dimensions as compared to heterosexual leaders and, due to their different roles and expectations, LG candidates for legislative office.

Additionally, these questions not only capture perceptions of the party leaders themselves, but also how the respondent anticipates the electorate, party activists, and other party elites would react to the new leader (specifically questions 2, 4, and 5). These anticipated reactions are particularly important because, as discussed in the previous section, LG politicians are often stereotyped as unelectable and potentially ineffective, even by supporters of LGBTQ rights, due to expected homophobia from the public and other politicians.

While these questions capture both evaluations of party leaders themselves (questions 1 and 3) and anticipated reactions to party leaders (questions 2, 3, and 5), as well as leader-specific attributes such as the ability to unite the party (question 4) and helping the party win more seats (question 5), and more general political attributes that take on enhanced significance for party leaders (questions 1, 2, and 3), we find participant responses across all five dimensions to be highly consistent. Thus, because Cronbach’s $\alpha $ , which shows how closely a set of items relate to each other as a group, is $0.84$ for our five questions, we combined the results to create a leadership evaluation index ranging from 0 to 5, which serves as the main outcome variable ( $\bar Y = 2.5$ , ${\hat \sigma _Y} = 1.96$ ).Footnote ¹²

We also included an attention check and a manipulation check. The first was presented to participants before the conjoint tasks and asked respondents to pick the (text of) color brown among six color options. Respondents who failed this attention check were not presented with the conjoint tasks and were thus excluded from our sample. The second was presented to respondents after the conjoint tasks and asked in a multiple-choice format which office the politicians in the tasks were elected to, with the correct answer being ‘the position of the party leader’.Footnote ¹³

Robustness

The results presented are robust to a variety of robustness checks provided in the Appendix. First, following the recommendation of Clayton et al. (Reference Clayton, Horiuchi, Kaufman, King and Komisarchik2023), we assessed the intra-respondent reliability by including a repeated conjoint task at the end of the survey. Respondents completed an additional task where the profiles from the first task were repeated with leader profiles flipped. Results demonstrated a high degree of consistency in responses within individuals, reducing concerns about intra-respondent reliability (see, Appendix E.1 for these results).

Second, we validated the leadership evaluation index through disaggregated analyses of its five component questions. These analyses showed that treatment effects were consistent across these five questions. In addition, we validated our results by iteratively removing each question from the index and confirming that the patterns remained unchanged. Further, our results remained stable under alternative approaches to modeling correlation in responses for each individual, such as specifying respondent fixed effects instead of clustering standard errors by respondents (see Appendix E.1 and Appendix E.2).

Third, we replicated our results by excluding inattentive respondents identified through the manipulation check following the conjoint tasks asking the respondents to identify the tasks’ political office of focus. Appendix E.3 shows that our results stay robust when we use only the subsample of those respondents who correctly identified the tasks’ focus as the political party leadership.

Finally, we checked whether the respondents’ ideological orientation and predispositions towards gay individuals condition our results. Appendix E.4 and Appendix E.5 present these findings and demonstrate that penalties for LG leaders are more pronounced among ideologically conservative respondents and those with higher levels of bias against gay politicians, as measured by attitudinal questions included in the survey.Footnote ¹⁷ In contrast, respondents with more supportive attitudes towards gay politicians evaluated LG and straight leaders more similarly. These patterns reinforce the robustness of the findings by demonstrating that the observed penalties align with well-documented ideological and attitudinal divides on LGBTQ+ issues.