For the love of Sporty Spice: the interaction of gender, sexuality, and sport in /s/ production

1.1 Women’s football in the UK

Women’s football is currently breaking records in viewership in the UK. In 2019, the BBC (British Broadcasting Corporation) vowed to air more women’s sport on their channels than ever before through their “Change the Game” initiative (BBC 2019). As part of this line up, the highest level of professional women’s association football in England, the Women’s Super League (WSL), is regularly aired across the BBC channels during the season. The 2022 UEFA European Women’s Football Championship had the top level of viewing for any women’s Euros previously held, regularly doubling viewership from the previous 2017 tournament (UEFA 2022). The 2023 FIFA Women’s World Cup also broke several records, with a 29 % increase in attendance compared to the previous World Cup, as well breaking women’s football viewing figures in several countries around the world (FIFA 2023). However, this upsurge in female football is relatively new and still gaining its traction both globally and within the UK specifically.

In what may be surprising for a country that is so captivated by football, women’s football in England has had a difficult and ill-supported history (Williams 2003). In 1921, the Football Association, the governing body of association football in England, banned all women’s teams from any clubs that were affiliated with them. This had an unfortunate influence over the sport more globally and many countries followed suit and banned women’s football. The ban in England was not lifted until November 1971. These decades of banning women’s football not only made the game virtually inaccessible for women in those years, but left a devastating social association with women in football for decades after. Williams (2003: 113) writes, “[w]hat has never been satisfactorily explained is a peculiarly English expression of contempt for women who play association football”. The banning of women’s football for so long cemented the sport as being viewed as almost exclusively masculine in England. Caudwell (2007: 184) writes, “[a]nalyses demonstrate that football in England is gendered as masculine and racialised as white”.

Though women’s football was more widely accessible after 1971, it did not regain the popularity it once had before the ban. It also still faced societal stereotypes for those women who did play, since participation was intrinsically seen as a challenge to hegemonic femininity. As Sauntson and Morrish (2012: 153) argue in their study of a university women’s football team, by playing a sport almost exclusively associated with men due to the historic Football Association ban, the team members are inherently challenging gender norms. They write, “[p]articipants in women’s football are likely to be women who are less concerned than some others about observing the norms of hegemonic femininity in this age group, including heteronormativity”. This correlation between football and masculinity, therefore, means that women playing football are inherently challenging gender norms to some extent.

As noted above, there is also an assumption that female football players are more likely to be gay. In an interview with Sky Sports (2022), Beth Mead, a professional football player for the WSL club Arsenal and forward for the English national team, discusses how common being gay is in the sport. She says, “especially in the women’s game it [being gay or bisexual] is more the norm”. It was also reported that a record number of openly LGBTQ + players would be participating in the 2023 World Cup (Billson 2023). This high proportion of openly gay and bisexual players, as well as the concerted effort of fostering a more inclusive sport (Binley 2019), can be viewed as a further challenge to heteronormativity.

These understandings of gender norms and the incompatibility with football was not just a theoretical concept for the participants of this study, but something they were keenly aware of. The views of the participants around football will be discussed further in the discussion section below.

1.2 Theoretical frameworks

The current paper utilizes a third-wave (Eckert 2012) approach to variationist research and considers how the use of linguistic variation, in this case /s/ production, can construct identity based on the social meaning of a variant. When a specific language form is regularly linked to a certain social meaning, then sociolinguists would describe this process as indexicality (Hall-Lew et al. 2021: 5). The social meaning that a linguistic feature can index may be associated with a certain group of people who also use that feature. However, some linguistic features may also index a stance being taken, that may relate to a group of people. To give another example, a linguistic form such as the word “like” being used a lot as a discourse marker, can be associated with a group of people – maybe young women. But it can also be connected to an attitude or behavior, like being youthful, that has also been connected to that same group of people.

To illustrate this description of indexicality, we will consider research conducted by Moore and Podesva (2009) on the use of tag-questions by female students from a school in the northwest of England. In the study, four different groups of students were recorded through ethnographic work. These groups each use tag-questions differently, and in different manners to index unique aspects of themselves. For example, the Geek group used tags in a way to signal knowledge and that they had an understanding of the topic being discussed. However, the Populars frequently used tag-questions when talking about other groups of girls, particularly in evaluative ways. This shows us that specific use of tag-questions could be used to index either association with a particular group, Geeks or Populars, or could be used to indicate a stance, being knowledgeable or evaluative. This broad overview is to give an example of the framework used in the present paper. For more detailed discussions of indexicality consider Eckert (2008) or Hall-Lew et al. (2021).

Another important aspect of third-wave linguistic research to consider is construction of identity, and in turn how variationist research uses categories for participants. In previous sociolinguistic work, participants were placed into categories with the assumption that the differences between these categories was stable. If, for example, a speaker was part of the working-class category, then their productions of a certain feature would consistently pattern in an expected way. However, third-wave research argues language users are more active, though not always consciously, creating their identity through different manners, including language. When discussing queer theory, Motschenbacher (2011: 153) summarizes this elegantly by stating, “The relationship between language and identity is here rather considered as constructive, i.e. identities are seen as constructed in the very moment of language use”.

This view of fluid identity construction has helped push researchers away from a strict binary comparison within language and sexuality work. In his research on a speaker referred to as Heath, Podesva (2007) demonstrates that the use of falsetto in different contexts and with different audiences is employed to construct a diva persona. This diva persona is actively used by the speaker when in a social setting with friends, but is not used when talking to patients at work or when speaking to his father because this particular identity is not relevant in those contexts. This work demonstrates that the same person can perform different identities that are constructed in the moment based on the specific context, and we can use linguistic features as one way to perform these identities.

It is not just research on intraspeaker variation that can demonstrate identity construction, but also looking at in-group variation. Work by Calder and King (2020); Hazenberg (2016); Podesva and Van Hofwegen (2014), Willis and Youssef (2023); Zimman (2017) all demonstrate that intersecting identities play an important role in /s/ productions. In these studies aspects like gender identity, ethnicity, and orientation to the country or city show variation within speakers who may all fall under similar categories in previous first-wave variationist research. These studies will be considered in more detail in the literature review of /s/ below.

What the studies listed above do, and what the present study does, is attempt to find locally significant categories that develop our understanding of participant identity. Instead of presupposed, large-scale categories, third-wave research attempts to focus on locally significant categories based on speaker identity. Eckert and Podesva (2011: 8) write, “Our point here is not that categories should be avoided, since the quantitative variationist enterprise centers on the identification and explanation of correlations between social categories and linguistic practice. Rather, we suggest first that categories be ethnographically significant and second that they be viewed as products of, rather than explanations for, variation patterns”. By considering ethnographically motivated categories, researchers can access a more accurate sense of the social meaning behind the variation being researched.

The present research will consider how different groups of female speakers in England use /s/ realizations to index specific social meaning, which contributes to the creation of a distinct identity. In order to understand why /s/ has been chosen for the current research, the next section will highlight the key literature on /s/ , gender, and sexuality.

1.3 /s/, gender, and sexuality

Having considered how football is placed within the UK and some of the overarching themes of third-wave sexuality research, the next step is to understand the linguistic feature that this paper focuses on. As will be evidenced below, /s/ has been a thoroughly researched feature when considering both gender and sexuality. This long history provides ample data for comparison with the current study.

One of the first studies regularly cited when researching /s/ is Schwartz (1968), where it was determined that the gender of an English speaker could be accurately distinguished based on isolated voiceless alveolar fricatives (i.e., tokens of /s/) and that women tend to have higher peak frequency in /s/ than men. With the consistency in production differences based on gender, as well as consistent gendered perceptions, one could believe that gendered /s/ productions can partially be predicted by physiological factors. One of the main physiological features credited with this difference is oral cavity size (Fuchs and Toda 2010).

To produce /s/, there is a constriction near the front of the mouth, between the tongue and the alveolar ridge, though the exact placement of the tongue may be slightly more forward or back (Ladefoged and Johnson 2011), which will be discussed further below. Due to the nature of where the constriction is, and the lack of voicing in /s/, the front of the oral cavity is the main influence on the frequency of /s/ production. It is therefore important to consider if female and male oral cavities differ significantly in such a way that would consistently influence /s/ production. Fuchs and Toda (2010) attempted to test this physiological variability by measuring the palates of 24 participants (12 male and 12 female speakers) and comparing their /s/ realizations. They found a trend of male participants generally having longer palates than the female English participants. However, they conclude that these physiological differences are not great enough to explain the difference in /s/ realizations alone, and it is the combination of physical and social factors that produce these differences. The social side that this study demonstrates is significant in gender identity and we can see that /s/ realizations can index gender (among other identities discussed further below). Many studies have demonstrated the importance of socially motivated productions, in particular, gender.

Holmes-Elliott and Turner (2019) demonstrate that the change of center of gravity measures over time for adolescent speakers also cannot be based on physiology alone. The young people in the study, and particularly the young women, show variability within their average center of gravity measures as they are recorded across four years. Some participants increased in center of gravity measures, associated with a more fronted /s/. But other female participants had lower center of gravity measures over the four-year span. We would also expect the oral cavity to change in these speakers as they age from 9 to 11 years old to 13–15 years old and presumably grow in that time. Despite the likely potential growth, which would predict a lower center of gravity measure for all participants regardless of gender, several female participants had a rise in center of gravity. If participants have higher center of gravity measures despite growing, then we can clearly say that /s/ realizations are doing social work. And this understanding of social significance in /s/ can be seen in speakers as young as 13 years old. The variability within the gender groups also demonstrates that center of gravity is not based on physiological differences alone and further confirms that social factors play a part in the production of /s/.

This complexity within gender identity and /s/ can be seen in Stuart-Smith et al. (2003). They discovered that of the variables they considered, only /s/ showed a clear and consistent gender distinction, with women having overall higher mean and peak measurements. The mean refers to the central point where all the energy is equal and peak measures were taken as “the frequency at which the highest amplitude peak occurred in the spectrum” (Stuart-Smith et al. 2003: 1852). However, there was also significant variation between groups of female speakers and their /s/ productions. They found that social class played a significant role in particular for the younger working-class women. These findings show that not only is /s/ being used to index gender, as the previously discussed studies have also shown, but how /s/ is realized can also distinguish groups of female speakers. While the group of younger working-class women do identify as female, they are also using /s/ realizations to distance themselves from other groups of women, in this case middle-class women or older women. We can see that /s/ does not only index gender, but through this study we can see that /s/ can index a certain gender identity.

Zimman (2017) also shows the importance not just of gender distinctions, but of gender identity distinctions of /s/ realizations among transgender speakers. In this data, there is significant variation within gender groups, and they point to other aspects of identity, that interact with gender, which may be influencing /s/ realizations. Zimman (2017: 1,016) writes, “this dynamic interplay between self and other demands a consideration not just of biology or gender assignment, but also the gender identities that individuals’ claim for themselves and the semiotic expression of those identities in everyday life”. The present study furthers this understanding that /s/ realizations are not simply defined by gender or biology alone, but through specific gender identities that are relevant to a given context.

The study of /s/ has also been a cornerstone of sexuality research in linguistics due in large part to the stereotype of the ‘gay lisp’. Munson and Babel (2007) discuss the prevalence of the gay lisp in popular culture, from television shows to books and stand-up comedy. This stereotype has in itself led to a considerable amount of linguistic research investigating /s/, in both production and perception, in connection with sexual orientation. The stereotype of the gay lisp is associated specifically with gay men, and it is partly for this reason that much of the previous linguistic work on /s/ and sexual orientation has focused on male speech. While these same stereotypes do not exist for gay women, the prevalence of this early male sexuality research, as well as the gendered productions of /s/, has led to further investigations of how female speakers may index their sexuality through /s/ realizations.

In a study based in Northern California, Podesva and Van Hofwegen (2014) analyzed the center of gravity of /s/ in relation to several factors: gender, sexual orientation, and whether the participant was more country- or town-oriented. They found that straight country men had the lowest average center of gravity and straight town men had the second lowest average. Straight town women had the highest average center of gravity, with straight country women having the second highest. In Podesva and Van Hofwegen’s study, straight speakers set the poles of the continuum for center of gravity in productions of /s/ and LGBTQ speakers made up the middle. A similar continuum is found by Hazenberg (2016) with participants from Ottawa, Canada. In the study, straight women and straight men create the poles of the continuum, with straight women having the highest average center of gravity measurements and straight men having the lowest. Queer women, trans women, queer men, and trans men fall within these poles, respectively from highest measurements to lowest. Importantly, Hazenberg (2016: 285) writes, this data “also reinforces the idea that the distribution is socially orientated, rather than biologically determined”. These studies demonstrate that it is possible that the LGBTQ speakers in the community were indexing their sexual orientation or gender identity through their pronunciation of /s/. As with Stuart-Smith et al. (2003) and Zimman (2017), these studies show that /s/ does not simply index gender, but can be used to demonstrate specific gender identities that interact with other factors, such as sexuality.

A final study of /s/ to consider is work done by Saigusa (2016). This study examines the effect the addressee has on the speaker and the center of gravity of the speaker’s /s/ production. In the study, two interviews with Jane Lynch, an openly gay actress, were analyzed. The first interview was by an openly gay news reporter and the second was with two female TV hosts considered not to be gay. The other hosts may not identify as straight, but as they are married to men it is unlikely they identify as gay. Saigusa found that Lynch had a lower center of gravity in /s/ when speaking with the openly gay news reporter than when speaking to the non-gay TV hosts. It was also found that when Lynch was discussing LGBTQ issues, she had a lower center of gravity than when she was discussing other topics, such as beginning her acting career. As with the previous research, gender is not the only factor that can influence /s/, but it is the specific interaction between gender and sexuality that can impact /s/ production. In this example, this variability between relevant identities is seen to fluctuate within the same speaker, but it provides evidence that /s/ is a fruitful area to consider the importance of gender identity, particularly how it intersects with sexuality.

By considering a group of participants that vary both in their sexuality, but also in their participation in sport, we can further see that /s/ is doing complex, social work that is more than simply making a binary gender distinction. Previous research on Northern California nerd girls by Bucholtz (1999), Chicana gang girls by Fought (1999); Mendoza-Denton (2011), and young, Glaswegian working-class women by Stuart-Smith et al. (2003) have shown that speakers, and in these cases particularly female speakers, make great use of the symbolic potential of linguistic variation to distinguish themselves from other groups of their same gender. In the present study, where gender, and in particular gender typicality, is highly salient and relevant, we can closely examine how /s/ is used in a unique way to distinguish a specific gender identity that interacts with both sexuality and sport.

The consideration of women’s football in the UK, as well as previous work on sexuality, demonstrate that a women’s football team presents a unique area to consider how sexuality may be indexed through language. There has been previous work that considered female sexuality using conversational analysis (see Jones 2018 for a thorough discussion of the previous conversational analysis), but less sexuality research focused on sociophonetic features with female speakers. This data therefore allows for a novel consideration of not only a particular group of speakers, but a less considered sociophonetic variationist approach. Not only will this work explore a broad idea of language and sexuality, but it also allows for a more nuanced view of how gay women may differ from each other, in this case through participation in sport.

2.1 Participant profile

In the present study, there are 12 participants who identified as gay and 10 who identified as straight, based on a questionnaire completed during the time of interview. The majority of the participants that identified as gay were also part of the Yorkshire Town Ladies. The age span was from 21 to 46, though most participants were aged between 27 and 35 at the time of recording. All the participants identified as White-British and they were all born and raised in Yorkshire, England. Each participant was given a pseudonym to protect her anonymity. Table 1 provides a summary of the participants, their age, sexuality, and team membership.

2.2 Data collection

Data were collected using traditional sociolinguistic interviews (Tagliamonte 2006), lasting on average 30–45 min, and carried out by the author. They were held either in the participant’s home or another neutral, quiet meeting space, such as a university classroom. These sessions included discussions around their childhood, personal interests, memories of the football team if they were members, and more general discussions of gender and sexuality.

The recordings were captured on a Zoom H4n solid-state recorder (sampling rate 44.1 kHz; bit depth 16 bit). Time-aligned, orthographic transcriptions were completed using ELAN (The Language Archive 2018) and forced-aligned using the software package FAVE (Rosenfelder et al. 2014) resulting in a Praat (Boersma and Weenink 2023) TextGrid time aligned at a phonemic level.

All files were then manually checked to ensure the /s/ tokens were correctly identified and the boundaries precise. Observations of /s/ were collected for the entire duration of the recording. All /s/ tokens that were adjacent to another sibilant were removed, as distinct boundaries could not be drawn. Tokens were also removed if there was interference from ambient noise or overlapping speech. Finally, /s/ tokens in which voicing was present were removed. Using a Praat script adapted from Fecher (2011), /s/ tokens were measured for center of gravity. Each token was first bandpass filtered to a 1,000–22,050 Hz bandwidth, following Podesva and Van Hofwegen (2016). This filter was used in order to decrease the influence of ambient sounds coinciding with the fricative by deleting any acoustic data with a frequency of less than 1,000 Hz. A spectrum was created from the full duration of each token and the mid-point was measured to determine the center of gravity.

Each /s/ token was coded for linguistic factors to control for them in the analysis. These factors include: syllable position, phrasal position, prosodic prominence, preceding sound, and following sound. The final corpus, at 18 h, 10 min length, yielded 12,368 /s/ tokens from 22 participants.

Data were statistically analyzed in R (R Core Team 2008) using Linear Mixed Effects Regressions (lmer) using the lme4 package (Bates et al. 2015). Fully saturated models were developed using Sexuality and Team Membership as fixed social factors, with Duration, Previous Phonetic Context, Following Phonetic Context, Syllable Position, Phrase Position, and Stress as fixed linguistic factors, and Speaker and Word as random factors. Therefore, reported values from social factors are based on model estimates which control for phonetic and random factors. Fully saturated models were stepped using “LmerTest” (Kuznetsova et al. 2017). This stepping process would indicate if any of the fixed factors should be removed if they were not significantly impacting /s/ realizations.

4.1 Sporty identity

As established, football is stereotypically associated with masculinity within the UK and there is an assumption that the majority of female players are gay. The participants in the present study, and in particular the participants that were part of the Yorkshire Town Ladies, were very aware of the societal expectations around women’s football. During the interviews I collected, we discussed how their gender identity and sexuality were part of how they viewed themselves as footballers, and how they believed others viewed them.

One of the founding team members, Michelle, discusses how she may be considered butch based on her activities, which all revolved around the sports she participants in. But in aspects like fashion and personal presentation, she believes she is quite feminine. She says in her interview:

Again I I think I can I can go both ways, um. In one sense you’d probably say I’m butch in the sense I do boxin’ and I play football and I can wear guys’ clothes, so that is more butch than my sister. So, she would never do anythin’ like that. In a sense I’m hittin’ like some male kind of things. But in another sense I wear make-up. I like my hair [laughing], yeah I like my hair done nice. I like perfumes. I like my clothes to look good like when I’m wearin’ the girls’ clothes. I like, if I’m wearin’ a dress, I like the handbag and shoes and nail varnish to go to go with it. I can go go either way.

With these words Michelle encapsulates the complexity of her identity and clearly demonstrates that she falls within the boundaries of different stereotypes. Many of the gay members of the football team discussed this idea of having qualities that they see as feminine, while also having qualities that they see as masculine, and therefore cannot just be butch or femme. The particularly masculine qualities appear to be based more on activities they participate in, and this can focus on sport. For these reasons, I argue the gay football players are demonstrating a ‘sporty’ persona, as they are performing a specific female identity that connects not only their gender identity, but their sexuality and participation in sport, particularly football. As discussed above, the participation in football opposed to many other sports, has certain social expectations and this is folded in with the rest of their social identity. With the current group of participants, I use the ‘sporty’ label to acknowledge a specifically lesbian identity. However, with further researcher, this label may need to be expanded to other sexualities.

The connection between gender and participation in football is something that many have been aware of from a young age. Three members of the team, the three that established the Yorkshire Town Ladies originally, all discussed playing football when they were younger, when the teams were mixed gender. However, once they reached a certain age, the teams were divided, and boys and girls could not play together. For one participant, this meant her parents telling her she was no longer allowed to play. Elizabeth says:

When I was a kid, um, I used to play left of defense, um, up until I sorta twelve years old, uh, and then at twelve, girls didn’t play football when I was kid. So I hung, my dad said, right time to hang your boots up lady. So I did and then just me being me found other sports to play and but I’ve always, uh, regretted not continuin’ or being able to continue playin’ football through my teens and into my early twenties.

Two other members also discuss the frustration they felt when they were in school and football was not a sport girls were allowed to play. Each shared that they pressured their teachers to let them play until they were allowed to practice with a football on their own while the other girls participated in different sports. This gatekeeping around who could play football at a young age intrinsically linked football and gender with these participants. To play football as a girl would therefore be going against gender norms, and this precedent was established by an imposing institution in most young people’s lives: school.

In line with previous research by Sauntson and Morrish (2012: 153), those who are part of the football team appear not to fulfill hegemonic gender norms. This is then taken a step further when it is considered that their sexuality also does not fit within hegemonic femininity. This shift away from hegemonic femininity, at least in certain contexts, creates a unique intersection around gender identity, sexuality, and sport.

To further exemplify the complexities around their intersectional identities, it is interesting to consider Taylor, the only participant from the Yorkshire Town Ladies who identified as straight. When discussing playing football previously at university, Taylor says:

Uni, when I was at uni, it was very like my group of friends at uni was very gay, like all gay nearly, to be fair, uh, there was only a couple a straight ones during my time at uni, but. Now it’s more straight people I’d say [laughing].

(Interviewer: Was it because of all the sports that you played?)

Yeah, just ’cause the football side of things, and it was like, well people on my course was they were straight and, um, some of the people that I lived with, they were straight as well, um, but. As to go out it was always with the football lot, which obviously it was about two straight people, that was about it.

This study provides initial indications that participants may be indexing a sporty identity through their /s/ realizations. From the discussions highlighted above, it is clear that members of the Yorkshire Town Ladies see themselves as diverging from at least some norms of hegemonic femininity, though not all of them. It is possible that part of this diversion is indexed through their use of language, and in the case of this study, through their realizations of /s/.

I believe this /s/ realization may be a resource for gay female football players who feel they do not fall into stereotypical femininity and who are of a certain age. There is previous evidence that gay female speakers have lower center of gravity measures, but there is also evidence that more rural female speakers have lower center of gravity measures than their urban counterparts (Podesva and Van Hofwegen 2016). What these studies tell us is that /s/ retraction may be a way for female speakers to demonstrate their divergences from hegemonic femininity.

As women’s football continues to develop in mainstream society, it is possible that many younger players, both gay and straight, no longer feel that they are challenging gender norms simply by participating. As reported by the English Football Association (FA 2023), 75 % of schools in England are delivering equal access to football in PE (physical education) lessons. Therefore, the resource described above may prove to be in the unique space of not only sexuality and sport participation, but also age. However, any shifts in stereotypes associated with women’s football will take time to manifest.

4.2 Sporty as flexible

Based on the data presented, I argue that the gay team members of Yorkshire Town Ladies are active in a sporty identity that the other gay participants are not necessarily part of. It is important to note that this does not mean that every woman who plays football is inherently sporty, or that every gay woman that plays football is sporty. It is possible that gay women who do not actively participate in a sport, but have an interest in them, could also be sporty. Since the core of the YTL was created through word of mouth and friend-of-friend soliciting, it is likely that this group of women have more in common than a more randomly created team. Their friendships, particularly off the pitch, could be connected to this shared sporty identity and common interests. They would form a loose community of practice, as defined by Eckert and McConnell-Ginet (1992). At the time of recording, this community was starting to disperse and may not be as directly comparable to communities of practice that others have studied.

As the eventual disbandment of the team demonstrates, this concept of sporty can be fluid and changing. The interviews with the YTLs were always in the context of football. I got to know them through attending some of their matches and they knew I was interested in interviewing them specifically because they were part of the team. During the interviews we discussed football at length, including their history in playing, as well as their enjoyment in viewing the sport as fans. Football was at the forefront of our relationship and therefore the forefront of the recordings this data captures. The patterns seen within the group of speakers is unique to this time and context and could shift with other speakers from a different aspect of their life. When sport, or even sexuality, are not relevant, it may be likely that /s/ realizations may alter, as that indexical field may not be recognized.

However, the fact that this data only gives a small sample of who these women are in no way downgrades its importance. This adds to the body of research that demonstrates the fluidity of people and the ever-shifting nature of our identities. While Ruby may sound and look quite different off the pitch, which she discusses in her interview, it does not mean that who she is on the pitch is any less her. Though these recordings only showcase one specific context, the content of the discussion shows that similar findings may be present in a more ethnographic study with a similar group of speakers.

4.3 Future work

It is clear through this study that there is significant scope for future research. Work that includes more straight football players would allow for closer analysis of the interaction between sexuality and team membership, to understand if a sporty identity may be unique to gay players.

To widen our understanding of a sporty identity, it would be worth researching female athletes that participate in sports that are frequently associated with femininity (e.g., gymnastics, netball). If participants in these sports do not feel that they are challenging femininity norms in the same way as the football players described above, how they index their gender identity may significantly differ from the participants in this current study.

This research may also benefit from more explicit questionnaires on gender identity. Kachel et al. (2017) utilized two different questionnaires to gain a deeper understanding of participants’ gender identity: the Traditional Masculinity-Femininity Scale and the Extended Personal Attributes Questionnaire. Questionnaires like these, along with ethnographic data, may provide valuable insight into complex gender identities interacting with sport and sexuality.