Are Openly LGBTQ2+ the New Sacrificial Lambs? Campaign Contexts and the Gendered Implications for LGBTQ2+ Candidates

Abstract Recent increases in the number of openly LGBTQ2+ candidates have not resulted in a corresponding rise in the number of LGBTQ2+ politicians elected to the Canadian House of Commons, reviving the hypothesis of the “sacrificial lamb” candidacies. Drawing upon Lovenduski and Norris’ work on political recruitment, we analyze the backgrounds and experiences of the 172 LGBTQ2+ candidates who ran in the 2015, 2019 and 2021 federal elections in Canada. Our approach is based on the idea that LGBTQ2+ candidacies are the new sacrificial lambs of Canadian politics, although some of them seem less likely to be sacrificed than others. Indeed, we highlight how the electoral opportunities (for example, district competitiveness) afforded to LGBTQ2+ cis men are more likely to result in success than those afforded to LGBTQ2+ cis women or gender minority candidates.

Like other Western countries, Canada has seen increases in the number of "out" (or "openly" known to the general/straight public) LGBTQ2+ 1 people running for political office. Numbers rose from 9 in the 2000 federal election to 69 in 2021 (down from 81 in 2019). However, there has not been a corresponding rise in the number of LGBTQ2+ individuals elected. In 2000, winning candidates Svend Robinson and Réal Ménard represented 22 per cent of the out LGBTQ2+ candidates. In 2021 the number of elected LGBTQ2+ MPs was 8, which was only 11 per cent of the LGBTQ2+ candidates. This suggests that, as has been the case for women (Thomas and Bodet, 2013), LGBTQ2+ candidates may be "sacrificial lamb" candidacies-that is, they are running for parties that have little likelihood of winning or in ridings that are less winnable for their parties (Baisley and Albaugh, 2021). However, although women and LGBTQ2+ candidates share challenges as underrepresented identities in Canada, they are different minorities in their nature. Women constitute a quantitative majority, despite maintaining a social minority status; LGBTQ2+ people are, by all criteria, a quantitative minority. Thus, an analysis of the experiences of LGBTQ2+ candidates requires a critical assessment of the intersectional challenges they potentially face, or privileges they may enjoy, as women and men, as well as gender minority members within the LGBTQ2+ community.
Recent progress in diversifying the sexuality and gender identity and expression of political personnel masks a continuity in the gendered nature of our political institutions: when elected, LGBTQ2+ representatives are more likely to be cis male than they are to be cis female, trans/non-binary or 2-Spirit individuals. 2 To date, only four openly LGBTQ2+ women-Libby Davies, Sheri Benson, Pascale St. Onge and Melissa Lantsman-have won seats federally, while 15 men have contested and won office. 3 No trans or non-binary candidates have won federal office, although three that we are aware of have held a seat at the provincial level. 4 There has been some research exploring the argument that LGBTQ2+ candidates suffer as sacrificial lambs for parties with limited likelihood in winning (see Baisley and Albaugh, 2021;Everitt et al., 2019). There are also data from a period in which many candidates were not out in their first election campaigns that suggest that LGBTQ2+ candidate success is not linked to gender (Everitt and Camp, 2014). These findings made it possible to conclude that the lack of LGBTQ2+ cis women or trans/non-binary and 2-Spirit elected officials was primarily due to the deficit of these candidates (Ashe, 2020). Now, with recent increases in the number of lesbian or bi women and gender and sexual minority candidates running for election, it is time to examine the role that gender and sexuality play in LGBTQ2+ candidate success-particularly for the cis women and trans/non-binary candidates who face specific and/or additional intersectional challenges (Haider-Markel et al., 2017;Jones et al., 2018;Jones and Brewer, 2019;Magni and Reynolds, 2021; see also Cao and Gurcay, 2021).
In this article, we examine the idea that cis women (for example, lesbians, gay self-identified and bisexual women) and gender minority (trans/non-binary, queer and 2-Spirit) openly LGBTQ2+ candidates may have become the new sacrificial lambs of Canadian politics, at least when compared to LGBTQ2+ cis men who seem more likely to be electorally successful. We cannot speak to the difference between LGBTQ2+ candidates and straight cis candidates (see Baisley and Albaugh, 2021), 5 but in comparing the experiences of all out LGBTQ2+ candidates in the 2015, 2019 and 2021 Canadian federal elections (N = 172) who are cis women, cis men and trans/non-binary and 2-Spirit, we can highlight how the electoral opportunities afforded to LGBTQ2+ cis men result in greater success than those afforded to LGBTQ2+ cis women or gender minority candidates.
We begin by describing the recent changes in LGBTQ2+ representation and then draw upon a theoretical framework developed by Norris andLovenduski (1989, 1995) and Norris (1996) in their research on women's political experiences, in order to explore the potential barriers facing LGBTQ2+ candidates at each stage of the election process. We conclude that although sexual orientation and gender identity present obstacles for LGBTQ2+ candidates seeking elected office, cis gay or bisexual men still benefit politically from being men. Cis lesbian or bisexual women, as well as candidates who are trans/non-binary and 2-Spirit LGBTQ2+, face a greater hurdle, as they continue to be nominated by parties with lower likelihoods of winning and, within those parties, relegated to ridings with less welcoming social and electoral contexts. In effect, they have become the new sacrificial lambs.
LGBTQ2+ Politicians in Canada It is impossible to determine how many secretly gay, lesbian or bisexual politicians have been elected since 1867 or whether any previous individuals seeking or holding political office would today identify as trans, non-binary or 2-Spirit. However, since Svend Robinson, MP for Burnaby, came out of the closet in 1988, 19 out LGBTQ2+ individuals have held seats in the House of Commons. In recent Parliaments, close to 2 per cent of the seats have been held by LGBTQ2+ representatives, which is half of the 4 per cent that sexual and gender minorities reportedly make up in the general population (Statistics Canada, 2021; see also Perrella et al., 2012Perrella et al., , 2019. At no point have there been more than eight MPs sitting at a time, and there is no guarantee that the number of LGBTQ2+ MPs will rise at a steady rate, since after the 2019 election they dropped from six to four, all of whom were cis gay men. Nonetheless, since the early 2000s, an increasing number of lesbian, gay and bisexual politicians have been elected in legislatures across the country, with some sitting as provincial premier or as federal and provincial cabinets ministers (Everitt and Lewis, 2020;Everitt and Tremblay, 2020).
However, even as the percentage of LGBTQ2+ cis women candidates match the percentage of women candidates in general, the percentage that succeeds in getting elected remains notably lower. The opposite is true for LGBTQ2+ cis men. As Table 1 indicates, in 2021 women made up 37.5 per cent of all candidates for the six major parties and 43.5 per cent of all LGBTQ2+ candidates running for office. Yet while women now occupy 30.5 per cent of the seats in the House of Commons, only two LGBTQ2+ cis women were elected (up from 0 in 2019), representing 25 per cent of these candidates. Furthermore, while trans/non-binary and 2-Spirit candidates are becoming more common in Canadian election campaigns, only one, so far, has been elected to federal office. This disparity between the percentage of candidates and the percentage who are successfully elected leads us to question whether LGBTQ2+ cis women and gender minority candidates face even greater electoral challenges than do their cis male gay or bisexual counterparts.

Theoretical Framework
To explore this question, we draw on the theoretical framework developed by Norris andLovenduski (1989, 1995) and refined by Norris (1996) to study recruitment to the British Parliament. Norris and Lovenduski focus on the variables of gender, race and class, although Lovenduski (2016) later makes the case for the model's applicability to a wide range of issues. This model is most frequently used to study the recruitment of women, with most researchers employing a limited understanding of gender, typically narrowly defined as women/femininity versus men/masculinity. Few are sensitive to sexual and gender diversities, illustrating Tremblay's (2021) criticism that most feminist theoretical perspectives are blind to sexualities, gender identities and expressions. Following Lovenduski's (2016) suggestion, we adapt this model to provide a fruitful understanding of the political recruitment of LGBTQ2+ candidates.
The model identifies four different hurdles that must be overcome in order to become a candidate. These included the social context, the recruitment process, the supply of candidates and the electoral demand for those candidates.
Social context refers to the ideological, economic, socio-cultural and political macrostructures that organize how politics works in a given system. For LGBTQ2+ candidates, social context includes legal requirements for engagement in public life; resources (cultural, financial, and so on) to run for office; openness of political structures (such as the electoral and party system) to candidacies; and, perhaps most importantly, ideologies that govern social relations, such as capitalism, patriarchy and cisheteronormativity. 6 The latter (especially its hegemonic status) is significant in relation to LGBTQ2+ people, as it shapes public acceptance of LGBTQ2+ individuals and attitudes toward sexuality, gender identity and expression. We find the notion of "space invaders" developed by Puwar (2004)  particularly fruitful in illuminating the less obvious and more insidious power of cisheteronormativity to marginalize non-cisheteronormative people from politics: Social spaces are not blank and open for any body to occupy. There is a connection between bodies and space, which is built, repeated and contested over time. While all can, in theory, enter, it is certain types of bodies that are tacitly designated as being the "natural" occupants of specific positions. Some bodies are deemed as having the right to belong, while others are marked out as trespassers, who are, in accordance with how both spaces and bodies are imagined (politically, historically and conceptually), circumscribed as being "out of place". Not being the somatic norm, they are space invaders. (8) The connection between electoral politics and masculinity is well documented (Gidengil and Everitt, 1999;Maiolino, 2015). Likewise, the parliamentary space has been deemed a "male sanctuary" (Gardey, 2015: 218; see also Puwar, 2004: 77-105) imbued with a homosocial culture (Bjarnegård, 2013). That noncisheteronormative male bodies enter these political spaces is one thing, but that noncisheteronormative female bodies might dare to do the same further strengthens their status as outsiders, or invaders, insofar as they violate two of the constitutive boundaries of the somatic norm-that is, to be female and non-cisheterosexual.
The (partial) decriminalization of gay sex in 1969, LGBTQ2+ activism in Canadian society since the 1970s and more recently on the electoral scene (see Atkins, 2019;Rayside, 1998: 105-211;Tremblay, 2015) and the willingness for parties to nominate and run LGBTQ2+ individuals as candidates reflect a social context in recent elections that is more favourable than ever before. However, not all environments are as welcoming as others, and the social context in parts of the country may remain challenging to LGBTQ2+ candidates. Haider-Markel (2010: 80; see also Haider-Markel et al., 2020) argues that in the United States in the early 2000s, LGBT "candidates strategically select where and when to run and thereby reduce, and perhaps negate, the role of sexual orientation in elections." These locations tended to be larger urban centres with higher levels of social diversity and more liberal attitudes, as opposed to smaller or more rural communities, which are more likely to hold more conservative views about homosexuality (see also Wilson, 1995). Button et al. (1999: 204) observed that the chances for success of lesbian and gay candidates "are greatest in large, diverse communities with sizable gay populations and electoral structures that are hospitable to minority candidates." Similarly, Thompson (2022) demonstrates that living in a metropolitan area is conducive to greater support for LGBT rights, not least because this environment provides queer social infrastructure and more opportunities for intergroup contacts, thus fuelling the notion of metronormativity (Halberstam, 2005: 36).
Given arguments by Haider-Markel (2010), we expect that LGBTQ2+ candidates would be more successful in ridings with large LGBTQ2+ populations. While only a small proportion of Canadian ridings would have such a demographic base, we anticipate that a higher than average number of candidates would run in such ridings (Tremblay, 2022: 220-36), particularly as urban areas inclusive of a gay community, such as those found in Toronto, Montreal or Vancouver, enhance LGBT candidates' chances for electoral success (Everitt and Camp, 2014). Not only would there be more acceptance in "gay villages," but LGBTQ2+ friendly environments provide an electoral asset to LGBTQ2+ candidates in terms of identity empowerment, availability of campaign volunteers and voting networks (Nash and Gorman-Murray, 2015;Podmore, 2015;Thomlinson, 1997).
We also expect that LGBTQ2+ candidates would have greater success in ridings where the population holds higher levels of education, lower levels of religious affiliation and are more employed in service industries, government or white-collar jobs, as these backgrounds are highly correlated with more accepting views toward sexual and gender minorities (Bailey and Nawara, 2017;Magni and Reynolds, 2021;Smith, 2011). As Tremblay (2022: 225, 233-34) found in her recent study of LGBQ politicians in Canada, these are in fact the locales where most successful Canadian LGBTQ2+ politicians are elected.
The second component of Norris and Lovenduski's model is the recruitment process. This includes the rules that govern how a political party chooses its electoral candidates, which may make some parties more open than others to LGBTQ2+ candidates. In Canada, each party uses different rules, with differing degrees of emphasis on the importance of social and descriptive representation. For example, until the 2015 election, the leftist New Democratic Party (NDP) and, to a lesser extent, the centrist Liberal Party have been most receptive to nominating LGBTQ2+ candidates (Everitt, 2015;Everitt and Camp, 2014;Everitt et al., 2019). For the NDP, this is because rules for their nomination meetings promote affirmative action for underrepresented candidates such as LGBTQ2+ individuals (although these rules have met with resistance within the party and were weakly enforced before 2015, with the main beneficiaries being white women; for more details see Ashe, 2020). As Figure 1 indicates, the party typically fields more LGBTQ2+ candidates than other parties. Until recently, the Liberals' efforts to recruit LGBTQ2+ candidates were driven by the attitudes of the party leader. However, in 2016 the party changed section 7.1.a.i of the National Nomination Rules so that local associations now need to demonstrate that they have done thorough searches for equity-seeking candidates (including women, Black, Indigenous and people of colour, the LGBTQ2+ community, and people with disabilities) before holding their nomination meetings (Liberal Party of Canada, 2016). The combination of this new rule and the leadership of Justin Trudeau, an ally of LGBTQ2+ people and communities, has resulted in a slight increase in the number of LGBTQ2+ candidates that the Liberal Party runs.
Right-wing parties, such as the Conservative Party of Canada and the People's Party of Canada (PPC), have traditionally made no efforts to recruit candidates from underrepresented groups, arguing that everyone should have equal opportunities to step forward (Boily and Robidoux-Descary, 2019). That said, in recent years, these parties have followed the lead of several right-wing parties internationally that have opened up to LGBTQ2+ candidates. According to Reynolds (2019: 225), a "mix of socially liberal and economically conservative policies is increasingly the norm in the parties of the right in developed democracies." Indeed, as tensions within the Conservative Party of Canada illustrate, particularly since Stephen Harper left, conservatism is by no means monolithic but is rather traversed by different trends-social, economic and fiscal-with the latter two more prone to endorsing the principle of individual freedoms (including the choice of a non- cisheteronormative sexuality and identity). Social conservatism has little sympathy for LGBTQ2+ people and communities (Farney, 2012: 21-26, 98-126), although social conservatism may likewise be divided between "rights issues" and "life-and-death issues" trends (Tremblay, 2022: 194-95).
More recently, the Green Party has demonstrated a willingness to nominate LGBTQ2+ candidates, running 23 in the 2019 election, but that number dropped to nine in 2021, as the party struggled to find last-minute candidates in a poorly organized campaign. The Greens ended up with only 252 candidates running in the 338 seats. Finally, the very low number of LGBTQ2+ candidates from the Bloc Québécois (BQ) is surprising, given its self-proclaimed social-democratic ideology.
The third component of Norris and Lovenduski's model is the supply of candidates seeking to be elected. This includes the eligibility requirements for being a candidate, as well as the background and experience that makes for a worthy and credible candidate that parties want to run under their banner. In Canada, legal requirements to be a candidate include being at least 18 years of age on Election Day and being a Canadian citizen. Being a non-cisheteronormative individual has never been a formal criterion of electoral ineligibility, although until recently it was certainly an informal one. Having a criminal record 7 is a hindrance to running for office, and the importance of fitting the cisheteronormative model to enter politics has "trapped gay and bisexual men in the straitjacket of conforming to traditional views of masculinity and marriage" (Reynolds 2019: 237), a requirement that many may have been unwilling to submit to. 8 Indeed, as Tremblay (2022: 4) and Lalancette and Tremblay (2019) argue, complying with the requirements of cisheteronormativity means "being respectable," as measured by a commitment to a stable, long-term relationship. Norris and Lovenduski (1995: 122) argue that one reason for the descriptive representation deficit in parliaments is the lack of diversity of people who put forward their candidacy. Ashe (2020: 311) makes the same point for Canada: "It is likely that if more LGBTQ+ aspirants throw their hat in the ring they would get selected." Ajzen and Fishbein's (1980) theory of reasoned action provides an explanation for why more people do not come forward to run for office; it suggests that a person could come to the conclusion that considering her/his situation, getting involved in politics simply would not pass the test of normative behaviour.
Certainly, the state of public opinion regarding non-cisheteronormative identities enters into this equation. Support for homosexuality is quite recent in Canada (Poushter and Kent, 2020), although there is now significant support for the participation of LGBTQ2+ people in politics (Angus Reid Institute, 2017). However, not all LGBTQ2+ candidates are equally welcomed by voters. US studies have shown that trans candidates face more resistance than their cis counterparts (Haider-Markel et al., 2017;Jones et al., 2018;Jones and Brewer, 2019;Magni and Reynolds, 2021); such attitudinal studies have yet to be conducted in Canada, although we expect that trans/non-binary and 2-Spirit individuals, as well as those from racialized minority backgrounds, might perceive politics to be a less welcoming environment. Certainly work by Wagner (2021) suggests that LGBTQ2+ people do not enter into public life due to a desire not to be subjected to moral regulation regarding sexuality, gender identity and physical appearance.
In any case, Norris and Lovenduski found that this stage of the pathway to Parliament involves indirect (or systemic) rather than direct discrimination-that is, the qualifications sought by parties for candidates (for example, education, occupation, ethnicity/race) exclude many people. For example, past ideal candidates tended to have a post-secondary level of education and were employed in a legal, entrepreneurial or public service occupation such as education or healthcare (Coletto, 2010;Morden, 2019). Additional supply side factors that make someone appear a stronger candidate include being from a religious denomination that is dominant in the riding; sharing a racial, ethnic or linguistic identity with a large proportion of a riding's electorate; or having experiences in politics in past elections or at other levels of office. But as Norris and Lovenduski (1989: 94) argue: "By defining the appropriate qualifications for a career in politics in such a way [that is, people who are articulate, well educated, professional, who have high level of time, etc.] then certain types of candidates will tend to be successful. As a result women, working-class candidates and those from the ethnic minorities will tend to be consistently disadvantaged." In her research on the NDP's candidate recruitment process, Ashe (2020: 314) adds sexualities, gender identity and expression to Norris and Lovenduski's above-quoted observation: "That greater efforts are not directed towards recruiting more people from the LGBTQ+ community suggests the party's selection process still prioritises heteronormativity." The final factor affecting a candidate's success-electoral demand-reflects both the demand from a party for a candidacy in a given constituency and the support from the electorate for a candidacy. Candidates may be well qualified for office, but their chances of being nominated decline if the party is not open to having LGBTQ2+ representatives, if the riding is already held by the party's incumbent or if the riding is a party stronghold and competition for the nomination is more intense.
In the past, women have often found themselves recruited by parties in "lost cause" ridings or by parties whose electoral success is low (Thomas and Bodet, 2013). While there was little evidence that LGBTQ2+ candidates were run as sacrificial lambs in earlier elections (Everitt, 2015), there is some suggestion that this may have changed in recent elections as the number of candidates has risen (Baisley and Albaugh, 2021). This is particularly the case for the NDP (and the Green Party in 2019), which ran slates of LGBTQ2+ candidates that are notably disproportionately larger than their presence in the Canadian population. While the NDP have managed to get some LGBTQ2+ candidates elected, they remain the third party in Canada and have limited federal electoral success in many regions of the country. The Greens are even less successful and in 2021 could not even field a full slate of candidates. Thus, while these parties run more LGBTQ2+ candidates, they tend to be run in non-winnable ridings (Baisley and Albaugh, 2021). 9 Although the federal Conservative Party of Canada has run a handful of LGBTQ2+ candidates since 2000, it is only in 2019 when Eric Duncan won the nomination in Stormont-Dundas-South Glengarry-a traditional Conservative stronghold-that one was elected. In 2021, Melissa Lantsman, a Conservative party insider, was elected after winning the nomination in Thornhill, after Peter Kent, a long-term sitting Conservative MP, decided not to run again. Other LGBTQ2+ Conservative candidates ran in ridings that were far less competitive. While many Liberal LGBTQ2+ candidates have lost their elections, the Liberal Party does slightly better than others at running these candidates in ridings that are party strongholds-or at least competitive. Thus, they have had more of their candidates elected. Finally, since the BQ took office in Ottawa, it has elected only two LGBTQ2+ MPs-in fact, two cis gay men-in constituencies where they were more than likely to win.
Thus, differing degrees of electoral success speak to the need to better understand the factors that contribute to the success of LGBTQ2+ candidates, of which the degree of their party's electoral competitiveness in the ridings in which they run is of critical importance.

Methodology
In the past, identifying LGBTQ2+ candidates was difficult, as not all of them were out publicly to their constituents or even to their riding association. More recently, candidates who step forward are open about their identities, making them easier to track. In addition, advocacy organizations (such as ProudPolitics) and news media (for example, Xtra's "Rainbow Votes" newsletter) now work with political parties to record the number of LGBTQ2+ candidates running. This analysis is based on a record of all out candidates who ran in the 2015, 2019 and 2021 federal elections identified through various websites and media sources, including ProudPolitics, the LGBTQ news outlets Fugues and Xtra, the Canadian Broadcasting Corporation (CBC), web searches, and various candidate biographies. No candidate was included who had not specifically self-identified as LGBTQ+. Our list was compared to those produced by other organizations and other academics working in this field (see Baisley and Albaugh, 2021) and updated accordingly. While this dataset has likely not captured all LGBTQ2+ candidates in all parts of the country, it is the most complete record we have for this information and is large enough to make substantive observations that would not differ from the total population of LGBTQ2+ politicians.
For each LGBTQ2+ candidate, we recorded the social and economic context of the ridings in which they ran. This enables us to test arguments that urban centres, those with larger university populations, higher economic status or lower religious affiliations are more accepting and supportive toward LGBTQ2+ individuals (Bailey, 1998;Haider-Markel, 2010;Wilson, 1995). For example, to measure the urban or rural nature of a riding, we relied on a measure of district urbanity developed by Armstrong and his colleagues (2022) that captures multiple dimensions of urbanization. 10 This measure ranges from a score of +1.5 for the most urban district to a score of −1.5 for the most rural district. We also included measures for the percentage of the population that has more than a high school diploma, that worked in the tertiary sector 11 and that claimed no religious affiliation.
We also wanted to include a measure of the degree to which a riding was home to a significant LGBTQ2+ population, since these citizens might provide important electoral resources through campaign volunteers or votes that can be mobilized to support a candidate (Button et al., 1999). Since Statistics Canada does not provide this information disaggregated to the riding, we created a proxy measure taking into account those ridings that encompass "LGBTQ2+ villages" and "LGBTQ2+ spaces." An LGBTQ2+ village is publicly recognized as such because of well-defined boundaries, a history, institutions and a population that identifies as LGBTQ2+. An LGBTQ2+ village is also known to the straight population. Davie Street in Vancouver, Church and Wellesley Streets in Toronto, or the block of streets Ontario, Papineau, René-Lévesque and St-Hubert in Montreal are examples. An LGBTQ2+ space refers to a place that is much smaller in size, quite anonymous and less known to the straight public, even sometimes reserved to a few insiders. Our notion of LGBTQ2+ space builds on the argument of Puwar (2004) referred to earlier about the embodiment of institutions (namely, how masculinity-and we added cisheteronormativity-is embedded in the Commons), as well as the concept of space developed by Lefebvre (1991). Lefebvre argues that space is a place (a physical one, but it could also be an incorporeal and imaginary one, such as a Pride Parade) through which people define themselves, position themselves in relation to others and shape their identity-in short, develop and maintain their subjectivity. It could be an LGBTQ2+ bar or cafe, an advocacy organization, a community or health services centre, or a media outlet. Of course, an LGBTQ2+ village contains a host of LGBTQ2+ spaces, but an LGBTQ2+ space also unfolds its full potential of empowerment when it stands alone, in a sea of cisheteronormativity. A gay bar, however tiny and discreet it might be, is a space where people can drop the masks that cisheteronormative norms impose and can explore their subjectivity. As Millward (2015: 31)  To measure recruitment, we recorded the party that the candidate represented. Ideological predispositions of a political party (that is, its left or left-to-centre positioning) are highly correlated to its openness to running an LGBTQ2+ candidate. In Canada, it has been the NDP and, to some extent, the Green Party and the Liberal Party that have been most open to LGBTQ2+ candidates, while more right-of-centre parties have been more resistant .
To assess selection barriers, we measured candidate background demographics, including gender and gender identity (cis man, cis woman, or trans/non-binary or 2-Spirit), age, education, occupation, and whether the individual was a racialized or Indigenous candidate; studies often show that these factors are correlated with candidates' electoral success and with perceptions that candidates are well suited for electoral office. Included as well was information about political experience, such as the year in which the candidate had run, whether it was their first election campaign at this level or not, whether they had won or lost the election, and whether they had previous electoral experience at another level of office. This information was captured in order to determine if the more successful candidates were the ones with greater political experience.
We also included several measures of the electoral competitiveness of the riding. The first measure is an indication of whether a candidate was a challenger (running for a party who did not hold the seat), an incumbent (a candidate who had won in the last election) or inheritor (a candidate in an open seat their party holds). We also coded candidates by where they ranked among others on the ballot and created a variable that identified them as either the first-or second-place candidate or as falling third or lower in the returns. Finally, we included a measure of riding competitiveness, calculated by subtracting the LGBTQ2+ candidate's party's poll results in the previous election from those of the winning candidate-or from the next closest candidate if the LGBTQ2+ candidate or their party won the riding. We then categorized ridings as a stronghold (that is, a candidate's party received more than 10 percentage points more than the next closest candidate), a competitive seat (in which the difference is between −10 points and +10 points of the winner or next candidate), a weak constituency (in which a candidate falls between 10 and 30 percentage points behind the winner) or a lost cause (where the candidate loses by more than 30 percentage points to the winner).

Analysis
Of the 172 identified candidacies during the 2015, 2019 and 2021 elections, only 18 (10.7%) were successful. This number includes the three candidates who won in each of these elections: Randall Garrison, Rob Oliphant and Seamus O'Regan. It also includes Sheri Benson, Randy Boissonnault and Scott Brison, who won in 2015 but either chose not to run again (Brison) or lost their seats (Benson and Boissonnault) in 2019. Eric Duncan won for the first time in 2019 and was elected again in 2021. Boissonault was re-elected in 2021, along with newcomers Melissa Lantsman, Pascale St-Onge and Blake Desjarlais. Eighty-one (47.1%) of the LGBTQ2+ candidates were cis men, 65 (37.8%) were cis women, and another 26 (15.1%) were trans/non-binary or 2-Spirit. The poor showing of the women and trans/non-binary or 2-Spirit candidates hints that gender plays some role in the likelihood of a candidate winning a seat and that non-cisheteronormative female or gender non-conforming bodies are more likely to be perceived as space invaders, making it more difficult for them to enter into the heteronormative male sanctuary of Parliament. To assess this situation, we look at the four factors determining a candidate's success (social context, recruitment process, supply, and demand), with a control for gender to see if LGBTQ2+ candidates who are cis women or trans/non-binary or 2-Spirit candidates run under conditions that are less opportune than those for cis men who are gay or bisexual.
We begin by examining the social and economic context of the ridings. Of the 172 LGBTQ2+ candidacies, 35.5 per cent were in Ontario, 16.9 per cent in Quebec, 8.7 per cent in BC, 25.8 per cent in the Prairies and 12.9 in Atlantic Canada. As Figure 2 indicates, over half of the LGBTQ2+ candidates ran in districts that rank in the top 25 per cent in terms of urbanity (median = .750). However, if we look at the overall mean score for this measure, which is much lower (mean = .432), it becomes obvious that these candidates also run in parts of the country that are more rural and less urban, a result that challenges assumptions that urban ridings are more open to LGBTQ2+ candidates. Interestingly, cis women (median = .770) were more likely than cis men (median = .280) to run in the more urban ridings, as were trans/non-binary and 2-Spirit candidates (median = 1.00).
Not surprisingly, as Figure 3 demonstrates, trans/non-binary or 2-Spirit candidates were also more likely to run in those ridings with an LGBTQ2+ presence, including either a gay village or space. This was less the case for the cis men and cis women. Again, this observation forces a qualification of the idea that LGBTQ2+ candidates run in constituencies with a significant LGBTQ2+ population. Other differences between the riding contexts were minor. These results provide little support for arguments that riding profile plays a significant role in the more limited success in the election of the cis women or trans/non-binary and 2-Spirit LGBTQ2+ candidates, as the ridings in which they run should be more welcoming. Even more important is that additional analysis suggests that there is no difference in the success rates of the candidates who run in urban versus non-urban ridings.
The second hurdle candidates need to surmount is the willingness of a party to nominate them. The party that ran the most LGBTQ2+ candidates in the three elections was the NDP (90). The Greens (36) came next, followed closely by the Liberals (29). The PPC (3) and Conservatives (9) ran far fewer LGBTQ2+ candidates than the other parties. These results are not surprising, as they are consistent with parties' ideological positions and commitments toward diversity and representation. The lower numbers for the Liberals and Conservatives also reflect their smaller number of open seats where there is not already a sitting incumbent. Incumbency has long been pointed to as contributing to the limited opportunities for women and other underrepresented groups to gain nominations in winnable ridings (Pruysers and Cross, 2016;Tolley, 2019). Thus, given their lower incumbency rates, we might also expect that the NDP and Greens to run not only more LGBTQ2+ candidates but also more who identify as cis women or trans/nonbinary or 2-Spirit. As Figure 4 below indicates, this is indeed the case.
While the 79.3 per cent of the Liberal and 77.8 per cent of the Conservative LGBTQ2+ candidates in these elections were cis men, the proportion of cis men and cis women running for the NDP was slightly more balanced, despite Ashe's (2020) finding that the party had a greater difficulty recruiting lesbians than gay men. The Greens ran more cis men as candidates than cis women, but it was also the party to run the largest proportion of trans/non-binary or 2-Spirit candidates. The PPC and the BQ both ran only a few candidates and only in the most recent elections.
The third component of Norris and Lovenduski's model is the supply of candidates. Political candidates are often drawn from older and better-educated segments of society, employed in white-collar positions in professions such as law, business, education or government. The LGBTQ2+ candidates who ran between 2015 and 2021 fit this profile, with only minor differences appearing in the backgrounds of cis men, cis women and trans/non-binary or 2-Spirit candidates (see Figure 5). LGBTQ2+ people may be space invaders, but only with regard to their  sexuality; in other socio-demographic traits, they conform to informal, but very real, expectations of what is required to access political office.
Because we were only able to find the exact age for one-third (59) of the candidates, we use information on years spent at university or first jobs found in candidate biographies or LinkedIn profiles to identify rough age cohorts for all but seven of the rest. 12 Based on these age ranges, we can confirm that a greater proportion of the cis male LGBTQ2+ candidates were over the age of 40, while women or trans/ non-binary or 2-Spirit candidates were younger, and in both cases almost half were under the age of 30. This would have made them less experienced but potentially more comfortable in their identity as members of the LGBTQ2+ community, as younger Canadians are more likely to be out than are older Canadians.
There were few differences in candidates' educational or occupational background, with 75 per cent having university-level educations. However, 40 per cent of the cis men, 21.3 per cent of the cis women and 33.3 per cent of the trans/non-binary or 2-Spirit candidates held graduate level degrees. There was little difference in the percentage of cis women and cis men in white-collar occupations, but trans/non-binary and 2-Spirit candidates were slightly more likely to be found in these types of careers. A closer examination indicates that cis men were more likely to be found in management positions and in business, while cis women and the trans/non-binary and 2-Spirit candidates were more likely to be found within education, law and social or government services or in arts, culture and recreation. One other difference was that cis women and trans/non-binary and 2-Spirit candidates (15%) were also more likely than cis men (5%) to be students. Nonetheless, it is clear that LGBTQ2+ candidates are often well educated and employed in fields from which political candidates are typically drawn. Finally, the cis women and trans/non-binary and 2-Spirit candidates 13 were more likely to be racialized or Indigenous and may therefore face additional barriers, while the cis men candidates in these three elections were more likely to be white.
In terms of previous political experience, of the 70 cis male candidates, 23 had past experience running in federal election campaigns, whereas only seven of the 61 female candidates did. All but two of the trans/non-binary or 2-Spirit candidates (both of whom ran in 2021) were running for the first time in a federal election campaign. Similarly, 10 of the cis men had elected experience in politics at either a band council, municipal, provincial or federal level prior to their election, whereas one of the cis women did and none of the trans/non-binary or 2-Spirit candidates did. This lack of experience may have made it difficult for the women or gender minority candidates to draw on previous networks of campaign support or financial aid that translate into winning campaigns (Coletto, 2010). In addition, as Magni and Reynolds (2021) have shown, past political experience mitigates the party elites' and electorate's electability concerns toward lesbian, gay, bisexual and transgender candidates (see also Haider-Markel, 2010: 149-50).
These results suggest that the cis women and gender minority LGBTQ2+ candidates were younger, more likely to be racialized, and have more limited experience running in federal election campaigns or holding political office than the cis men candidates. These qualities might have presented additional hurdles to their election that cis men might not have had to face. However, they were better educated and were slightly more likely to be employed in white-collar occupations than the cis men candidates. This makes it hard to conclude that they were less qualified candidates and thus less likely to win their seats.
The final hurdle that candidates confront in getting elected is the actual election itself and the support their candidacy and party receive from the electorate. As Figure 6 shows, there are clear differences when we compare the types of ridings in which LGBTQ2+ cis men versus women or gender minority candidates run. In these three elections, cis men won their seats 14 times, while a cis woman only won in three instances. This can be attributed to different levels of riding competitiveness. Fourteen of the men and only three of the women ran as either incumbents or as a candidate in a seat their party had previously held: 82.7 per cent (67) of the cis men and 95.4 per cent (62) of the cis women were challengers. All 26 of the trans/non-binary or 2-Spirit candidates were challengers. While odds were not good for the cis men, they were even worse for the women and gender minority individuals. In seven of the cases in which men (Scott Brison, Rob Oliphant, Seamus O'Regan, Randall Garrison and Eric Duncan) won their seat, they did so in one of their party's strongholds, a riding in which their party won by more than 10 per cent in the previous election. Of the two women who ran in seats previously held by their party prior to the 2021 election, one, Melissa Lantsman, did so in a party stronghold. Pascale St-Onge ran in a much more competitive Liberal riding, where the final vote required a judicial recount. Sheri Benson was a challenger in a competitive seat when she won in 2015 and an incumbent in a competitive riding when she lost in 2019.
However, 87.5 per cent of the cis women and 100 per cent of the trans/nonbinary and 2-Spirit candidates (who all ran as challengers) found themselves running in weak or lost cause ridings in which their party had lost the previous election by more than 10 per cent of the vote. 14 A larger proportion of cis women (69.2%) than cis men (60.5%) were ranked third or lower on the ballot. The situation for trans/non-binary or 2-Spirit candidates was even more dismal (76.9%). While six of these candidates appeared to be ranked in either first or second place on the ballot, this figure is deceptive, as many who came in second place fell more than 20 percentage points behind the top-ranked candidate.
To confirm that riding competitiveness is the most important explanation for the lower levels of electoral success of LGBTQ2+ cis women and trans/non-binary and 2-Spirit candidates, we created a regression model that tests the impact of all these variables on a candidate's likelihood of winning. As might be expected from these bivariate results, a multivariate model demonstrates that the only factor that contributes to whether a candidate wins or loses is the party's competitiveness in the past election. This holds true for whether the candidate is a cis man, cis woman or trans/non-binary or 2-Spirit.

Discussion
These results help to provide context for LGBTQ2+ candidates' electoral success and, in particular, the differences in the experiences of LGBTQ2+ cis men, cis women and gender minority candidates. Even though the number of LGBTQ2+ MPs doubled in 2021, they are still underrepresented in Parliament. More concerning is that when we focus on the number of candidates who run in recent elections, it becomes clear that LGBTQ2+ individuals are overrepresented in the candidate pool compared to their electoral success. This imbalance raises questions about their competitiveness as candidates and about the factors that might contribute to it. This discrepancy is even more notable among the cis women or the trans/nonbinary or 2-Spirit candidates who have been almost completely shut out of office in most recent elections.
Our results support the argument that political parties run LGBTQ2+ candidates as sacrificial lambs in lost cause ridings (Baisley and Albaugh, 2021). More importantly, it highlights the fact that this discrimination particularly confronts the cis women and the trans/non-binary or 2-Spirit candidates. A credible explanation of this disadvantage seems to lie in the strength of the party support in the ridings in which they run. Our findings suggest that cis men are more likely to run in their parties' strongholds, or at least competitive ridings, than are cis women or trans/non-binary or 2-Spirit candidates. This is not to say that LGBTQ2+ cis men do not run as sacrificial lambs. Rather we argue that when a party is willing to run an LGBTQ2+ candidate in one of its strongholds, they are more likely to run a cis man than a cis woman or trans/non-binary or 2-Spirit candidate. This pattern is similar to that found for women and men candidates more generally (Thomas and Bodet, 2013). Why is this so? It is possible that despite digressing from the cis-heteronormative paradigm, LGBTQ2+ cis men nonetheless enjoy the privileges of being men, including a male homosociability that confers on them an "ontological complicity" with the rules of the established political game that may enable them to be perceived as "familiar strangers" rather than "space invaders" (Puwar, 2004: 126-38). This is a hypothesis that research should explore further.
Another future avenue for research would look more closely at the internal dynamics of candidate nomination processes to learn more about what factors affect party choices. This might compare steps taken by electoral district associations to recruit LGBTQ+ candidates, the composition of electoral district recruitment committees, the openness and timing of the nomination contest itself and the diversity of candidates who seek the nomination. Such a study is beyond the scope of this article but becomes more important given the evidence provided here that it is parties' decisions about where to run LGBTQ+ candidates that contribute to their success.
Thus, while the number of LGBTQ2+ candidates has risen dramatically in recent elections, their success has been limited by the fact that they are mostly running for parties such as the NDP or the Greens whose likelihood of winning is lower than it is for the Liberals or Conservatives. Furthermore, even when they do run for the more competitive parties, LGBTQ2+ candidates seldom find themselves running in their party's strongholds. Those that do are typically cis men. This leaves us to conclude that despite a trend that increasingly sees political parties speaking about the importance of diversity among candidates, it may well be the parties and their choices about where to nominate LGBTQ2+ candidates that continue to present the greatest barrier to the diversification of our legislative assemblies.