Gay and Straight Men Prefer Masculine-Presenting Gay Men for a High-Status Role: Evidence From an Ecologically Valid Experiment

There is increased acceptance of gay men in most Western societies. Nevertheless, evidence suggests that feminine-presenting gay men are still disadvantaged compared to gay men who present in a more traditionally masculine way. Though gay men themselves may be complicit in perpetuating this bias, studies that demonstrate this possibility are scant. Whereas most studies on perceptions of feminine-presenting gay men have manipulated gender nonconformity via written descriptions, research suggests that behavioural cues such as voice and body-language can mitigate or exacerbate prejudice toward a stereotyped individual. In the current study, audio-visual stimuli were created to investigate how masculine versus feminine behaviour would impact status endowment from other gay and heterosexual men. In total, 256 men (Mage = 42.73, SD = 14.48: half gay; half heterosexual) cast, from a selection of six video-taped candidates, one gay man to play a lead role in a purported ad for a tourism campaign. In the videos, the actors delivered a script related to the tourism campaign in a manner where their voice and body-language was manipulated to come across as either masculine or feminine-presenting. Findings indicated that gay and heterosexual participants showed a significant preference for the masculine videoclips. For heterosexual men, the preference for masculine-presenting actors was predicted by greater anti-gay sentiment, whereas internalised anti-gay prejudice did not predict a preference for masculine-presentation among gay men. Implications of the findings for discourse and education on intraminority prejudice and suggestions for future research are offered.

Social status refers to how much prominence, respect and influence an individual accrues within a social hierarchy (Anderson et al., 2001), and is thought to be a commodity endowed on an individual by the group (Caspi & Bem, 1990). Expectations regarding an individual's potential to contribute to group goals influence their social position (Bernstein, 1981), and such expectations tend to be informed by socially learned stereotypes (Phelan et al., 2014). The "think manager-think male" stereotype (Schein, 1975, p. 340), whereby lay people tend to automatically associate leadership qualities with masculine characteristics, shows how masculinity can favourably influence performance expectations of others (Case et al., 2012;Koenig et al., 2011;Rudman et al., 2012). This association between masculinity and status endowment has complex implications for gay men, given the prevailing stereotype that they are more feminine compared to heterosexual men (Kite & Deux, 1987;Lippa, 2000;Mitchell & Ellis, 2011;Sanchez et al., 2009).

Gay Men and the Feminine Stereotype
Such a stereotype reflects, to some extent, average differences in gender-typicality between gay and heterosexual men. Gay men demonstrate significantly more interest in traditionally feminine hobbies and occupations than heterosexual men (Bailey et al., 2016;Lippa, 2000;Lippa & Connelly, 1990). Further, childhood gender nonconformity remains the best-known predictor of same-sex attraction 1 3 (Bailey & Zucker, 1995;Rieger et al., 2008). Childhood gender nonconformity is also associated with both parental and peer rejection in childhood (Landolt et al., 2004). As a result, gay men often internalise negative beliefs about male femininity ('femmephobia'), and behaviourally defeminise, to some extent, to avoid negative treatment. Policing of masculinity among gay men is not only self-directed; there is also evidence of prejudice toward more feminine gay men from within the gay community (Bailey et al., 1997;Hunt et al., 2016).

Status Penalties for Feminine Gay Men
Contemporary theories of effective leadership have challenged the perceived virtues of masculinity. Transformational leadership theories (e.g., Gartzia & van Engen, 2012;Ryan et al., 2011), for example, argue that feminine traits like empathy, nurturance and interpersonal sensitivity are better suited for managing modern organisations, compared to more traditionally masculine and domineering traits. Furthermore, it has been argued that gay men represent potentially strong candidates for leadership because they are more likely to possess such traits (Stein, 2013). Nevertheless, masculinity -as a central characteristic of status potential -continues to permeate contemporary consciousness (Koenig et al., 2011), with negative impacts on gay male status attainment.
A growing body of lab and field studies have demonstrated status penalties against openly gay men. As job applicants, they are less likely to get interviews, are rated less positively, and are offered lower salaries compared to heterosexual men (Ahmed et al., 2013;Badgett, 1995;Horvath & Ryan, 2003;Pellegrini et al., 2020;Tilcsik, 2011). Further, research has observed an apparent "gay glass ceiling effect" (Frank, 2006, p. 485) for upper-managerial positions akin to that found for women (Aksoy et al., 2019;Frank, 2006). Theoretical explanations for these findings consistently focus on the possibility that gay men elicit such discrimination because of the stereotype that they are feminine and are therefore perceived as less equipped to occupy higher-status positions in social hierarchies, such as the workplace (Kite & Deux, 1987;Lord et al., 1984).
Further support for this notion is found in studies wherein gay men appear to avert status-penalties when they adopt a more masculine presentation Morton, 2017;Pellegrini et al., 2020). Across these studies, greater anti-gay attitudes in participants predicted greater discrimination against feminine, but not masculine, gay men. Similarly, Clausell and Fiske (2005) found that subgroup labels for feminine gay men like 'flamboyant' elicited higher ratings of warmth, but lower ratings of competence compared to more masculine subgroup labels like 'straight-acting'. This finding demonstrates that increased representation and positive regard for feminine gay men in Western cultures does not necessarily equate to less discrimination against them when it comes to status opportunities. Thus, the research appears to suggest that feminine gay men are at particular risk of status penalties, especially from individuals who possess anti-gay attitudes.

Anti-Effeminate Sentiment Amongst Gay Men
A further question regarding potential status penalties for feminine versus more masculine-presenting gay men is how complicit gay men themselves may be in perpetuating such prejudice. Whereas most relevant research has used heterosexual samples, both lab and field studies on romantic partner preferences amongst gay men highlight a commonplace desire for masculine over feminine traits in potential partners (Bailey et al., 1997;Clarkson, 2006;Laner & Kamel, 1977;Sanchez & Vilain, 2012;Tayawaditep, 2002). Sanchez and Vilain (2012) showed that antifemininity and masculine idealisation in romantic partner preferences were associated with higher levels of internalised negative attitudes and beliefs about one's same-sexual orientation. Such a connection suggests that the extent to which gay men internalise societal stigma about being gay may influence their treatment of individuals who possess stigmatised traits.
There is a considerable literature demonstrating that gay men discriminate against more feminine gay males beyond the romantic context (Brooks et al., 2017;Ravenhill & de Visser, 2019;Sánchez & Vilain, 2012;Taywaditep, 2002). In one such study by Hunt et al. (2016), when gay men received bogus feedback that they had rated below-average on a masculinity measure, they were more likely to show a decreased desire to associate with a feminine -but not a masculinegay male target. Conversely, gay men who had their masculinity affirmed (i.e., receiving feedback that they were aboveaverage in masculinity) did not show such an effect.
Despite there being increasing research investigating discrimination toward feminine gay men by gay men, studies investigating such discrimination -as specifically measured by status endowment -are sparse. Salvati et al. (2021a) adopted an analogue research design to assess the impact of gender nonconformity on the perceptions of leadership effectiveness (see also Cuddy et al., 2007;Morton, 2017). The perceived femininity/masculinity of gay male targets was manipulated using written descriptions of their traits, interests, and qualifications, which tapped into traditional, stereotypical notions of masculinity (henceforth masculinity for simplicity). Salvati et al. (2021b) found that higher internalised homonegativity led gay men to rate feminine (vs. masculine) gay candidates as less-effective leaders. This effect among gay men mirrors similar findings observed among heterosexual participants (Aksoy et al., 2019;Frank, 2006;Pellegrini et al., 2020) that also used analogue tasks, in which masculinity/femininity of gay male targets were manipulated via written descriptions.
A factor not addressed with analogue designs, however, is that prejudice toward stereotyped individuals can be exacerbated or mitigated by behavioural cues (Johnson et al., 2015;DeGroot & Motowildo, 1999). For example, studies have shown that the more an individual's voice and body language aligns with a negative stereotype, the harsher the status penalties laid against them (Blair et al., 2004a, b;Klofstad et al., 2012). Therefore, analogue designs may lack ecological validity in instances where perceptions of leadership capability involve judgements based on visual and auditory cues of masculinity/femininity; arguably the majority of such instances. Therefore, a more ecologically valid alternative approach to analogue designs is to employ carefully calibrated audio and/or visual stimuli to operationalise masculinity and femininity.
Using a heterosexual sample of men and women, Fasoli et al. (2017) found that recordings of feminine gay voices led to lower ratings of leadership effectiveness compared to masculine gay voices. The same effect was not found when using still pictures of faces. This was possibly due to the absence of cues like facial expressions, eye movements, and body language, which may be needed to adequately distinguish feminine versus masculine-presenting stimuli. Another potential confound may be that, in this study, the individuals portraying masculine voices and faces differed from those portraying the feminine stimuli. Differences in perceived age and attractiveness between stimuli were not controlled for, despite both variables influencing status outcomes for men in other studies (e.g., Anderson et al., 2001). Whereas Fasoli et al. provided important advances in offering ecologically valid demonstrations of the reduction in status bestowed upon feminine men by heterosexual individuals, important unaddressed questions remain about whether gay individuals also show such a bias, using audio-visual stimuli, and what psychological mechanisms might explain such biases. Traditionally, studies investigating the impact of feminine-presentation on gay men's status have used either heterosexuals or gay men in isolation -to date, no study integrated these two populations to facilitate meaningful comparisons. Demonstrating that gay men are as likely to discriminate against feminine gay men as heterosexuals would contribute to the emerging awareness of intraminority prejudice as an area of concern for the gay community.

The Current Study
The aim of this study is to explore whether a relatively feminine-presentation negatively impacts status attainment for gay men using a more ecologically valid methodology that allows meaningful comparisons of the reactions of gay and heterosexual men. Moreover, the study aims to test psychological mechanisms that may underly the hypothesised reluctance to endow status to feminine-presenting gay men.
Most relevant lab studies to date have measured status attainment using indirect measures, such as subjective ratings of leadership effectiveness or behavioural intentions. However, it could be argued that such continuous measures do not adequately emulate real-world processes of statusendowment, such as recruiting for a high-status position, whereby an individual is either chosen or not. Such measures also do not address the possible issue of an intentionbehaviour gap (Sheeran & Webb, 2016).
The current study introduces a concrete status outcome measure. Participants bestowed an actual representational role on one candidate from a pool of candidates, in line with the contemporary perspective that "representation is power" (Levina et al., 2000, p. 738). This is especially important in the current context as empowered visibility is considered an important reflection of where out-groups exist in social hierarchies (Phelan et al., 2014).
The following hypotheses were proposed: 1) Overall, gay and heterosexual voters will be significantly more likely to choose a masculine-versus femininepresenting individual to represent their country's famous city. 2) Gay men will be as likely as heterosexual men to choose a masculine-versus feminine-presenting individual. 3) For heterosexual men, higher levels of anti-gay sentiment will predict a preference for a masculine-presenting individual. 4) For gay men, higher levels of internalised anti-gay prejudice will predict a preference for a masculine-presenting individual.
Though not informing primary hypotheses, we also examined whether sexism may mediate preference for more masculine gay candidates, given that Sanchez and Vilain (2012) found that antifeminine attitudes predicted a preference for masculine-presenting romantic partners. Sexism amongst gay men has been a long-standing topic of feminist discourse (Bonnet, 2002) and thus, exploratory analysis was employed for this secondary construct.

Participants
To detect a small effect size (Cohen's g = .10), with an alpha of .05 and power of .80, a sample size of 200 participants was required to compare directional binomial proportions according to a G* power calculation. To allow for possible attrition, we aimed to recruit 280 males (half gay, half straight).
Responses were collected via an online Qualtrics link from a national sample of gay (n = 128, 18-77 years old, M = 36.91, SD = 12.09) and heterosexual (n = 128, 18 -86 years old, M = 48.55, SD = 16.86) Australian men, matched for state of residence. Participants were contacted by Qualtrics personnel, based on pre-existing market panels who reimbursed them for their time. Sexual orientation was determined by a single-item pre-screen and only responses indicating a gay or heterosexual orientation were selected. From the original 280 responses, 12 from gay men and 12 from heterosexual men were excluded because of failing an attention test (an embedded audio test) or for not completing the survey.

Script for Fictional Ad Campaign
To demonstrate the direct influence of masculine/femininepresentation on status attainment for gay men, a novel dependent variable was constructed for the current study. Participants, who were naïve to aims of the study, were asked to choose a gay man to represent their country's largest city in an advertisement campaign (i.e., promoting tourism in Sydney) from a pool of six shortlisted candidates. An emphasis on status was reflected in the instructions, that read: "Thanks for agreeing to be a part of this market research. We're currently casting for an advertising campaign but need your help. This campaign aims to sell Sydney -to the rest of Australia and Overseasas one of the world's great cities to visit. To do this we want to include ambassadors from a wide range of communities -not just 'typical' Aussies (although we'll still have them of course!). We are really looking for people an audience can 'Admire' -think 'Leader'. As part of this we're looking to cast a gay man in the campaign. We'd like you to now evaluate our six shortlisted candidates by watching their auditions and then answering a series of questions about them. Please note that they're all using the same script, so focus on them as 'people' as opposed to what they're actually saying. You'll be asked to vote for your favourite at the end, so your feedback is VERY important to us!"

Masculine and Feminine-Presenting 'Audition Videos'
Two sets of video auditions that could be considered identical except for a manipulated difference in the feminine/ masculine-presentation in the actors' behaviour were created. Six cis-male, White-Australian professional actors, 25 to 35 years old (who all identify as gay in real life) were filmed performing an identical vox pop script in two ways; 1) once where they were directed to manipulate their voice and body language (VBL) to be more feminine, and 2) once where their VBL was to be more masculine. Their monologue read: "I guess what my boyfriend and I love most about living in Sydney is… the weather -who wouldn't love a 6-month Summer, right? …The food -so much great choice -and the best coffee in the world (except for Melbourne, of course). Um… and I guess we won't mention the traffic and the rent?" (Actor laughs) The script made no reference to the candidate's qualifications, occupation, skills, education, or hobbies (that is, information that may be construed as gendered by participants; Lippa, 2000), while making the candidate's homosexuality explicit (by mentioning a same-sex partner). No change to the actors' appearance was made between masculine and feminine versions. The actors were selected subjectively by the authors to be of similar ethnicity (Caucasian), age (between 25 to 35 years-old) and attractiveness (operationalised as having an average physicality, full head of hair and mild to no facial hair) -given that these qualities have been shown to indicate status in the literature (Anderson et al., 2001;Berger et al., 1980). Their height differences were controlled for by having them all filmed in a seated position. Each video was 19-23 s in length (with no more than 1 s difference between any given actor's two videos). The actual video sets (see Fig. 1) can be viewed at the following YouTube links, https:// youtu. be/ fWrBC OIDY-I and https:// youtu. be/ rkbxK hntg-4. Pre-ratings from an independent participant pool of 40 gay men were used to validate the VBL in each clip as being masculine or feminine as intended (See the online supplement for method and results of video validation study). As shown in Fig. 1, each participant in the main study was shown a combination of videos that included three masculine-presenting actors and three feminine-presenting actors.

Status Endowment
A single forced-choice item asking participants to select their preferred candidate read as follows: "Please now vote for the actor you think should be cast in the Ad Campaign promoting tourism to Sydney. Your vote will go toward deciding who will get to be our gay ambassador in the ad, so thank you for all your help!" The item was accompanied by a visual cue of all the actors with corresponding letter labels (as shown in Fig. 1, without the m/f designators). Our main dependent variable was comparing the percentage of participants who selected a masculine vs. feminine-presenting actor -from a pool of three masculine and three feminine candidates -to be in the ad campaign.

Internalised Anti-Gay Attitudes (Gay Participants Only)
The 3-item internalised homophobia subscale of the Lesbian, Gay and Bisexual Identity Scale (LGBIS; Mohr & Kendra, 2011) was used to assess negative attitudes toward oneself as a gay person. Using 5-point Likert scale where a score of "0" indicated "Totally agree" and a score of "5" indicated "Totally disagree", gay participants were asked to rate how much they endorsed the items, "I wish I were heterosexual"; "If it were possible I'd choose to be straight"; and "I believe it is unfair that I am attracted to people of the same sex". The scale showed good internal consistency (α = .85). The average of each participant's three responses were calculated to create their Internalised Homonegativity score.

Anti-Gay Attitudes (Heterosexual Participants Only)
To measure anti-gay attitudes we deployed an adapted 6-item version of the Modern Homonegativity Scale (MHS; Morrison & Morrison, 2002), as used by Morton (2017), to exclusively assess contemporary negative attitudes toward gay men. Using 5-point Likert scale, where a score of "0" indicated "Totally agree" and a score of "5" indicated "Totally disagree", heterosexual participants were asked to rate statements such as, "Gay men have all the rights they need"; and "Gay men seem to focus on the ways in which they differ from heterosexuals, and ignore the ways in which they are similar". The internal consistency was acceptable (α = .74). The average of each participant's six responses were calculated to create their Homonegativity score.

Modern Sexism (All Participants)
A 5-item subscale from the Modern Sexism Scale (Swim et al., 1995), assessing contemporary negative attitudes toward women was used. Using 5-point Likert scale, where a score of "0" indicated "Totally agree" and a score of "5" indicated "Totally disagree", participants were asked to rate statements such as, "Discrimination against women is no longer a problem in Australia"; and "It's rare to see sexism against women on TV". The internal consistency was acceptable among heterosexual participants (α = .74), and good among gay participants (α = .85). Modern sexism items were interspersed with five other distractor items (on tourism in Australia and Sydney) to diminish the perceived focus on sexism and reduce the chance of socially desirable responding. The average of each participant's five responses were calculated to create their Modern Sexism score.

Procedure
Participants were provided a Qualtrics link to access the study. The study could be completed on a computer or smartphone. Upon clicking on the study link, participants read a Participant Information Statement and Participant Consent form (The University of Sydney Human Research Ethics Committee approved the study). Upon providing consent, demographics (age and sexual orientation again) were completed. Participants then read the fictional ad campaign casting script, before  being randomly assigned to view one of two videoclip sets (6 videos in each), voting for their favourite representative after viewing all six videos. As the videos contained audio, participants were required to verify that they could hear the spoken content by passing a one-item audio test (an audio-only recording of the number "72" was played and participants were asked to type the number they had heard). After watching the assigned set of videos, participants were presented with still colour photographs of the six actors simultaneously (as in Fig. 1), each photo identified by a letter label. They voted for their favourite candidate by typing the corresponding letter label into a box provided. Finally, participants completed the other measures and were debriefed about the real nature of the study.

Analytic Plan
The critical analyses were binomial tests to assess differences in the frequency of votes for six masculine clips versus the six feminine clips. A chi-square test assessed whether masculine/feminine proportions in voting differed between gay and heterosexual men. Finally, logistic regressions examined whether a preference for masculine videos was predicted by pre-existing levels of internalised homonegativity (for gay participants) and homonegativity (for heterosexual participants), followed by exploratory analyses also using logistic regressions.

Results
Overall, combining the heterosexual and gay participants' responses (N = 256), masculine clips received 150 votes (59%), while feminine clips received 106 votes (41%). As predicted, there was a tendency to choose masculine actors at a rate significantly greater than feminine actors (p = .004, g = .09). Among gay participants specifically, masculine videos received 77 votes (60%), while feminine videos received 51 votes (40%). As predicted, gay men were significantly more likely to cast a masculine actor than a feminine actor (p = .013, g = .10). Figure 2 presents voting rates across all participants as well as for gay participants specifically. Further, as predicted, no difference in the proportion of votes for masculine versus feminine videos between gay and heterosexual participants emerged. χ 2 (1, N = 256) = .258, p = .611. Figures 2 and 3 depict the voting pattern for gay and heterosexual participants across all six actors' masculine and feminine videos. Of particular interest is the pattern that in most (but not all) cases each actor's masculine presentations received more votes compared to their feminine presentations. Further, while a masculine presentation on average received more votes, two actors received a majority of overall votes regardless of their presentation -indicating the influence of other unmeasured factors.
A binomial logistic regression was performed among heterosexual participants to examine the effect of homonegativity on the likelihood to vote for a feminine versus masculine gay male actor. The logistic regression model was statistically significant, χ2(1) = 4.557, p = .033, and examination of the predictor variable demonstrated that the likelihood of picking a feminine gay actor was negatively associated with homonegativity, B = -.497, SE = .240, Wald = 4.294, p = .38. For each one-unit reduction in homonegativity, the odds of selecting a feminine gay actor increased by 1.64 times, Exp (B) = .609, 95% CI [.380, .973].
Contrary to expectations, among gay participants, the logistic regression model examining the effect of internalised homonegativity on likelihood of selecting feminine versus masculine gay actors was non-significant, χ2(1) = 1.023,  1.195], in that we predicted higher internalised homonegativity levels would be associated with a lower likelihood of voting for a feminine gay actor. An exploratory logistic regression analysis was undertaken to examine if modern sexism predicted less likelihood of choosing a feminine gay male (over a masculine gay male) actor, and if this effect was moderated by each participant's sexual orientation. As age differences were present between gay and heterosexual samples, we included age as a covariate in the model.

Discussion
As predicted, participants demonstrated a greater preference to cast masculine-presenting over feminine-presenting actors as representatives in a tourism campaign. This preference was significant amongst gay men -at a rate no different from heterosexual men. As predicted, stronger levels of anti-gay attitudes predicted a stronger preference for the masculinepresenting actor amongst heterosexual men. The hypothesis that stronger internalised anti-gay sentiment would predict a stronger preference for the masculine-presenting actor amongst gay men was not supported. Finally, an exploratory analysis indicated that modern sexist attitudes predicted a stronger preference for the masculine-presenting actor among our participants (regardless of participants' sexual orientation).
The observed preference for the masculine-presenting actor supports the proposition that masculinity remains an advantageous characteristic in the pursuit of high-status opportunities (Berger et al., 1977;Berger et al., 1980;Kite & Deux, 1987;Lord et al., 1984). By using audio-visual stimuli and real-life (perceived) consequences, the current findings provide greater ecological validity in support of the phenomenon, which was explored previously (Fasoli et al., 2017;Morton, 2017;Pellegrini et al., 2020). The finding that stronger anti-gay negativity predicted preference for the masculine-presenting actor amongst heterosexual men also replicates previous studies (Morton, 2017;Pellegrini et al., 2020), offering further evidence for the connection between feminine-presentation among gay men and the increased risk of status-penalties from individuals who harbour anti-gay attitudes, even under circumstances of affirmative action (i.e., casting a gay man in a campaign).

Limitations and Future Research Directions
The finding that gay men were complicit in the bias against other gay men extends the findings of Salvati et al. (2021a), with more ecologically valid audio-visual stimuli. Along with Salvati et al. (2021b) these are the first known results to demonstrate in-group status-penalties against gay men who are feminine-presenting. Salvati et al. (2021a) found that stronger internalised anti-gay sentiment predicted masculinity-bias -in line with the proposition that the more shame one feels about their sexuality, the less likely they will want to be represented by a fellow group-member who perpetuates negative stereotypes.
In the current study, however, a preference for masculinepresenting actors amongst gay participants was not significantly predicted by levels of internalised anti-gay sentiment. A potential explanation for this non-significant relationship may be that the measure of internalised anti-gay prejudice used in the current study lacked sufficient sensitivity. The three-item internalised homonegativity subscale (Mohr & Kendra, 2011) included items, such as "I wish I was heterosexual," that may have suffered from a floor-effect, whereby the items were too extreme to be endorsed by a contemporary sample of gay men (particularly if such gay men were motivated to conceal obvious in-group prejudice). Other recent studies (see Hunt et al., 2020;Salvati et al., 2021a, b;Sanchez & Vilain, 2012) have used more comprehensive measures to operationalise internalised anti-gay prejudice, such as the 7-item Reaction to Homosexuality Scale (RHS: Smolenski et al., 2010). Given the robust theoretical rationale for internalised anti-gay sentiment as a mechanism underlying masculinity-bias amongst gay men, future studies should continue to investigate its role in status-penalties against feminine gay men (using comprehensive measures). Several related questions should be addressed by future research. Aside from masculinity, attractiveness has also been identified as a status characteristic for men (Anderson et al., 2001). Along with wealth, both attractiveness and masculinity have been identified as markers of high-status in gay communities (Pachankis et al., 2020). To isolate the effects of masculinity, the current design eliminated differences in attractiveness by utilising the same six actors varying only their feminine-masculine presentations. Participant ratings of actor attractiveness were not assessed to avoid its overt influence on voting decisions. However, future studies could investigate how masculinity and attractiveness interact in contributing to status attainment for gay male targets, especially to determine whether attractiveness is protective against status-penalties for feminine-presenting gay men.
The design of the current study did not allow for direct assessments of the underlying reasons for a masculinity bias to avoid raising participants' suspicion, but future research with a different design may benefit from tapping such reasons more directly. Whereas the mediation analyses tested antigay sentiment, internalised anti-gay prejudice, and modern sexism as potential explanations, the cross-sectional design limits drawing casual conclusions for those explanations (Bullock & Green, 2021). Additional possible explanations may relate to previous experiences with femininity-based stigmatisation or discrimination (e.g., Meyer, 2020), one's perception of their own masculinity/femininity (Salvati et al., 2021a(Salvati et al., , 2021b, and internalisation of societal positive biases of masculine-stereotyped traits (e.g., Aube et al., 1994;Hunt et al., 2020) among other possibilities to be examined by future research.
The current study did not investigate whether women would also levy status-penalties against feminine-presenting gay men. Fasoli et al. (2017) found no difference in the behavioural intentions of heterosexual women toward masculine versus femininesounding gay voices, however women rated feminine-sounding gay voices lower on leadership effectiveness. Considering these preliminary findings, future investigation into status-penalties against feminine-presenting gay men should also include heterosexual and sexual minority women.
A potential confound in the current study may be that, on average, heterosexual participants were slightly older than gay men. While a positive relationship between age and prejudice has been observed in literature (Danigelis & Cutler, 1991;Gonsalkorale et al., 2009), age was not related to masculinity-bias in either population in the current study. Further, a younger pool of gay versus heterosexual participants would theoretically suggest that the present results may have underestimated masculinity-bias amongst gay men compared to heterosexual men. Moreover, in our exploratory analysis, when controlling for age, sexism played a similar role in predicting masculinity-bias in both gay and heterosexual men. The post-hoc finding that modern sexism (toward women) predicted the masculinity-bias, highlights the potential importance of such attitudes in this context. Thus, modern sexism may represent an additional theoretically sound variable of interest for future research assessing masculinity bias, especially given that significant relationships were observed between anti-effeminacy and idealisation of masculinity in past research (Sanchez & Vilain, 2012).
Finally, power could have been increased with an ordinal, as opposed to dichotomous, dependent variable. To achieve this, participants could have been asked to rank the six candidates in order of preference. However, only being able to select one candidate is arguably more representative of the process of status-endowment in many real-world scenarios where a person is either awarded a position or not.

Practice Implications
While gay men appear to enjoy increasing equality and representation in Western cultures, they are still disadvantaged in pursuing high-status opportunities, compared to heterosexual men (Aksoy et al., 2019). The current study adds support to emerging evidence suggesting that gay men who fail to sufficiently project traditional masculine traits are at particular risk of status-penalties. Affirmative action, for example in the U.S. military where the recruitment of samesex attracted folk should meet diversity quotas (Kamarck, 2017), represents contemporary structural attempts to mitigate discrimination (Holzer & Neumark, 2006). The current study indicates, however, that even with such processes in place, feminine-presenting gay men may still be denied equal access to opportunities, particularly at the hands of individuals who harbour anti-gay sentiments. That a more masculine presentation was enough to elicit preferential treatment, in the absence of information regarding qualifications, highlights that greater awareness is needed regarding how outdated biases unfairly impact feminine-presenting gay men. To this end, more work is required to challenge the prevailing association between masculinity and high status, especially in light of contemporary transformational leadership theories, which posit that feminine traits like warmth can be more effective in managing modern work places (Alimo-Metcalfe & Alban-Metcalfe, 2005). The current study's findings further highlight how this contemporary shift in attitudes toward leadership and status continues to represent a tension for feminine-presenting gay men who have been stereotypically perceived as 'warm' but not 'competent' compared to masculine-presenting gay men (Clausell & Fiske, 2005).
The current study is also among the first to show that gay men themselves are complicit in status-penalties against feminine-presenting members of their own community. The finding is troubling as it represents an intraminority pressure amongst gay men to assimilate (as passably straight), perpetuating the inhibition of empowered representation from a diverse range of gay men. An emerging body of research has also shown how status-based rejections from within the gay community can negatively impact mental health, as much as discrimination from heterosexuals (Pachankis & Hatzenbuehler, 2013;Pachankis et al., 2020). One concrete step to counter the phenomenon of intraminority bias observed in this study, is greater representation of feminine gay males in high status positions in popular media. To facilitate this, diversity training could first be implemented within organisations, to educate and raise awareness about potential biases toward feminine-presenting men, in order to counter prejudice and status-penalties faced by them in hiring and promotional decision-making.

Conclusion
The current study demonstrated that both gay and heterosexual men prefer masculine-presenting over femininepresenting gay men for high-status representation roles. Among heterosexual men this preference was predicted by more anti-gay attitudes and among both groups it was predicted by greater sexism. Moreover, by using audio-visual stimuli of masculine versus feminine-presenting gay men and a concrete status outcome, the present study improved on the ecological validity of past literature on this topic. These findings inform the need for advocacy and training to counter apparent bias against feminine-presenting men in a range of contexts and populations.

Author's Contributions
All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by Benjamin Gerrard. First draft of the manuscript was written by Benjamin Gerrard and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.
Funding Open Access funding enabled and organized by CAUL and its Member Institutions.
Data Availability Deidentified data are available upon request from the corresponding author Benjamin Gerrard, bger5861@uni.sydney.edu. au. + 61,406,620,614.
Code Availability NA.

Declarations
Ethics Approval Ethics approval for this research was granted by HREC of the University of Sydney (project number: 2018/391) and was conducted according to the National Statement of Ethical Conduct in Human Research (2007). Informed consent was obtained from all participants.

Conflicts of Interest
The authors did not receive support from any organization for the submitted work and have no relevant financial or non-financial interests to disclose.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.