Casting for Change: Tracing Gender in Discussions of Casting through Feminist Media Ethnography

The moment of casting is a crucial one in any media production. Casting the ‘right’ person shapes the narrative as much as the way in which the final product might be received by critics and audiences. For this article, casting—as the moment in which gender is hypervisible in its complex intersectional entanglement with class, race and sexuality—will be our gateway to exploring the dynamics of discussion of gender conventions and how we, as feminist scholars, might manoeuvre. To do so, we will test and triangulate three different forms of ethnographically inspired inquiry: 1) ‘collaborative autoethnography,’ to discuss male-to-female gender-bending comedies from the 1980s and 1990s, 2) ‘netnography’ of online discussions about the (potential) recasting of gendered legacy roles from Doctor Who to Mary Poppins, and 3) textual media analysis of content focusing on the casting of cisgender actors for transgender roles. Exploring the affordances and challenges of these three methods underlines the duty of care that is essential to feminist audience research. Moving across personal and anonymous, ‘real’ and ‘virtual,’ popular and professional discussion highlights how gender has been used and continues to be instrumentalised in lived audience experience and in audience research.


Introduction: The Casting Call
The moment of casting is a crucial one in any media production. Casting the 'right' person shapes the narrative as much as how the final product might be received by critics and audiences. As recent casting decisionsfrom the all-female casts of Ghostbusters III (2016) and Ocean 's 8 (2018) to the casting of a woman for the titular role in Doctor Who (in 2018)-demonstrate, gender has become a more and a less important category for casting at the same time. Considering how the media have a pivotal role in representing specific gender definitions and in the chances offered to women and men to produce media content (Krijnen & Van Bauwel, 2015), the recent casting of women in roles formerly held by men and outside of heteronormative constraints could be cause for (feminist) celebration. Equally, discussion over the casting of cisgender actors for transgender roles is worrying. Given the scarcity of such roles, and the unequal opportunities for trans as opposed to cis actors, it brooks no argument that transgender actors should be favoured. One could be as uneasy, though, over insistence on gender correspondence between actor and role as about biological bodies grounding gender. This article addresses struggles over gender and gender distinction and pursues a post-structuralist feminist agenda for audience studies.
We understand gender as strictly a social convention that references but is in no way dictated by hormones or body parts.
While the media industry might be 'seeing the light', as audience researchers we see new struggles over gender and gender distinction coming up in interviews and in online discussion. The loosening of restrictive definitions of gender and gendered roles for some brings the link of gender to biology back with a vengeance. For this article, casting-as the moment in which gender is hypervisible in its complex intersectional entanglement with class, race and sexuality-will be our gateway to exploring the dynamics of discussion of gender conventions and how we, as feminist scholars, might manoeuvre. Tracing casting decisions from 1) (black and white) male actors in gender-bending comedies, through 2) female reinterpretations of male legacy roles, to 3) cisgender actors for transgender experiences, will allow us to approach discussion of gender(ed) performances from different angles. It invites us moreover to discuss the merits of the different methodologies we adopt. This article wants to use and develop feminist media methodologies to move beyond restrictive definitions of gender. We examine how different methods reflect (or challenge) our own position as feminists, researchers and fans and what specific insights different methodologies allow for.
'Casting' also describes how we approach research methods. Carefully testing and comparing the 'performances' of different methods reflects an ethic of care often applied to research subjects in feminist scholarship but rarely to methodology itself. Yet, as we will argue throughout this article, in casting methodological affordances and feminist considerations intersect. Feminist methodology grounds a politically engaged research practice that is sensitive to how gender is part of systems and mechanisms of exclusion-and decisions on who gets to play which role both on and off the screen. Expanding from the historical focus on the rights of women, we conceive feminist methodology today as more broadly the foundation for a research practice that is consciously and reflexively involved in an ongoing struggle for respect and equal opportunities without reducing any individual's intersecting identities to single categories (Crenshaw, 1991;hooks, 1981).
From the perspective of audience research, it is important to recognise that studying identity and representation involves the to-and-fro between publicly available media texts and their use in relatively private environments. Practices of use are informed by a plethora of texts, sources and contexts of conversation. Feminist media research is interested in everyday meaning making, and therefore needs to take into account what Fiske (1987) called primary and secondary texts (whether television series, films, books or trailers and reviews in newspapers) as well as tertiary texts which include audiences talking about all of these.
In exploring the different layers of social talk and interpretations across these spaces, our goal is not to find consensus but to strengthen the (possibility of) exchange and ongoing conversation. Feminist methodology, from our perspective, should insist on forms of analysis that are open to diversity-and therefore to controversy and paradox-as much as to common denominators and generalisations. In a political sense, we see change as most likely to happen as a result of making connections and bridging controversies. This entails understanding that 'polarisation' does not necessarily signify a binary stand-off. It can be a frame that hides how disagreement and tension can also be visualised as a 'spread' that includes middle positions as well as fierce disagreement. We will restrict our own 'casting call' for contemporary feminist media studies to the uses of ethnography-related approaches (Hammersley & Atkinson, 2007). Ethnography has become a shorthand term for qualitative research "concerned with studying people in their cultural context….While classical ethnography was characteristically concerned with describing 'other' cultures, contemporary ethnography has focused its concern to settings nearer to home" (Draper, 2015, p. 36). Even though, as media researchers, we tend to rely less on extended participant observation and long sojourns in 'the field,' we know ourselves informed by ethnography and indebted to 'the ethnographic turn' in cultural studies (Drotner, 1994;Hermes, 1997;Schrøder, 1994).
The current availability of audience discussion on social media has been a boost to hearing, researching and discussing everyday talk and interpretation of the widening visibility and discussion of nonheteronormative identities and representation. It is impossible to chart the history of these recent discussions of gender in such a short space. Suffice it to say that transgender studies, as they developed out of media and cultural studies and alongside queer studies in the same period, focused on the importance of understanding everyday meaning making, make clear that we need to understand gender as a contentious and over-asserted category that imprisons individuals (cf. Stryker, 2006). We are aware that there is no consensus among feminists about the dominance accorded to biology or the way in which gender is forcefully corseted into two distinct categories. In the interest of disclosure and reflexivity, we ourselves enjoy the current moment of fragmentation, diversification, choice and confirmation. We seek to support it in our academic work. We also understand the current conjuncture as an opportunity for methodological reflection and growth. Hence the casting of three different methodological approaches from the ethnographic toolbox.
Examples of gender(ed) casting will be offered in three 'miniatures'. These case studies allow for reflection and evaluation of methodology for feminist media studies practically, reflexively and politically as well as in terms of validity, generalisability and usefulness. They work best in tandem. The starting point for our discussion is collective auto-ethnography. Discussing our own memories of 1980s and 1990s comedies, white and non-white men dressed up to impersonate women for comic relief clearly touched a nerve for us. Auto-ethnography has the methodological merit of bringing one close to irritation and unease while offering unprompted and uncoerced access to memory. Mapping controversy, though, is not its strength. Our second mode of inquiry moves from the private sphere to semi-public forums for discussion which have the undisputed strength to foreground differences of opinion. That makes online ethnography (here called 'netnography') a promising route to locating the process by which gender loses some of its inscribed patriarchal logic in the early 2010s when women are cast in (legacy) male roles. This touched a nerve for others. Moving closer to the present and another dimension of gender roles, we turn to our third miniature which focuses on textual analysis of online and print media platforms to examine how professional journalists frame discussions about the casting of cisgender actors for transgender roles.
Together these three approaches allow for researcher self-reflexivity and for triangulating discussions of gender and media from a personal and academic, a general audience, and a journalistic perspective within the private, semi-public and public sphere. All three methods come from the ethnographic toolkit. All three offer the opportunity to not assume the meaningfulness of any social category up-front. And all three come with their own sets of advantages and disadvantages.

Autoethnography: Gender and Humour
The starting point for this article was a conversation about examples of gender-bending in film and television and how unfunny (some of) these comedies seem from the present. Rather than summarise the long history of playing with gender in the arts, we adopted our exchange as autobiographical method and drew on our personal recollections. This has the advantage of foregrounding how, as audience researchers, we are also audience members and affectively entangled with the examples we discuss (see Table 1). We both have vivid memories of the gender-bending comedies from the 1980s and 1990s. Some we saw, others we remember choosing not to. First then we will discuss two examples to 'set the stage.' Dustin Hoffman, an award-winning white man, plays an actor in Tootsie (1982) who needs to convince a producer that he is a great actress. This all goes wrong when he falls in love with his female co-star while his male co-star falls in love with him. Both of them think he is a woman. We remember key scenes as well as ironically distancing ourselves from the assumption that a straight guy could never fall in love with a man in drag or a younger woman for a slightly older one. Big Momma's House (2000) also tells a story of a man impersonating a woman to further his professional career. Here, a Black FBI agent impersonates an overweight grandmother and another case of complicated triangles of unrequited attractions unfolds. Neither of us ever watched it, the promise of racist stereotype and size-shaming in the trailer is such that we never chanced it. Whilst a boxoffice hit, critics similarly felt it was all a bit much.
Tootsie works as situation comedy and handles its stereotypes deftly, allowing us to enjoy its humorous rendering of the main character's efforts to land a job. Big Momma's House careens out of control and invites laughter about its characters rather than the situation they are in. It invites its viewers to share in a form of scorn that feels very wrong to us. Discussing these two films, we remembered a much older favourite which we cherish for staying well away from realism and caricaturing all of its characters almost equally. In Some Like It Hot (1959), two musicians are on the run from the mob after witnessing a shooting, infiltrate an all-female orchestra and, of course, both fall in love with Marilyn Monroe. It is a friendly satire of wealth and class difference and offers silly but likable male characters against a strong woman.
In all three cases men dress up as women for professional reasons. As they fall in love with female characters and are pursued by men, they all come up against what the films suggest are the limits of their disguise. When dressed as women, their love interests will not fall in love with them and they are deeply uncomfortable by being pursued by other men. Heteronormative restrictions necessitate the end of the charade. As the wigs and women's clothes come off, the narratives end with a happily ever after.
As method, autoethnography employs the lens of the self. It allows for approaching gender-bending films at the intersection of 'innocent' entertainment and unease. We laughed at Tootsie with Oscar-winner Dustin Hoffman and avoided Big Momma's House. Other methods of data collection are less likely to bring to the surface how casting decisions are part of how well generic codes work; and how gender-bending within a genre can feel very different: A sign of (possibly) changing times or a vehicle for sexist and racist jokes. Our unease and dismay also signal how a quarter century ago belittling and demeaning representation of women and gay sexuality were deemed acceptable to an extent we would balk at today.
Methodologically, sharing media memories is a form of 'collaborative auto-ethnography' as proposed by Chang, Ngunjiri, and Hernandez (2013, p. 17): "A group of researchers pooling their stories to find commonalities and differences and then wrestling with these stories to discover the meanings of the stories in relation to their sociocultural contexts." Bochner and Ellis' urge for ethnographic stories "to be used as well as analysed, to be revised and retold rather than settled and theorised, and to promise the companionship of concrete, intimate detail as a substitute for the loneliness of abstracted facts" (Bochner & Ellis, 1996, p. 4). When used with long quotes, the validity of collaborative auto-ethnography is high. Its generalisability depends on triangulation with other methods that offer a broader either theoretical or a select form of sampling of discussions or discussants. Looking back, watching the gender-bender comedies of the 1980s to early 2000s and enjoying some of them, would seem to be in poor taste. Yet, the mistaken assumptions and short-sightedness that characterises so much daily interaction were sometimes portrayed really well and absolutely funny. Talking about them again-and challenging our own recollectionsmakes us recognise the situatedness of our perspective. It also helps see how humour affords a double take on gender. The gender comedies both challenge and reconfirm gender conventions. Collaborative autoethnography allowed for self-reflection and for happy, 'serendipitous' discovery and suggests sensitising concepts (Merton & Barber, 1958/2011. Here those concepts would be humour and ambiguity. The ambiguity relates to nostalgia as well as to the triple position of the researcher: as audience member, as fan and as feminist intellectual. Humour will return as a key mechanism in ongoing social gender negotiation. A host of gender-bending comedies with crossdressing male protagonists quickly followed our two 1990s examples: Matt LeBlanc in All the Queen's Men  (2011). This "craze for cross-dressing in film and popular culture" may well have revealed "a desire to put identity into question, [that was] not limited to a small coterie of feminist and queer theorists in the academy" (Modleski, 1997, p. 523). On the other hand, as much as comedy can offer hope and social criticism, it also almost always returns to the status quo (Marc, 1997). The genre inscriptions of gender reversal in (romantic) comedies stages and undermines the transgressive potential of playing with identities. At its worst, "cross-dressing in film represents the needs of comedy and society to have a subject to ridicule" (Miller, 2015, p. 127). As with drag culture, men impersonating women can strike too close to the bone and become derogatory and insulting (González & Cavazos, 2016;Taylor & Rupp, 2006).
Reconstructing the uses and limits of gender-bending in film comedy, it is clear that masculinity does not suffer from temporarily masquerading as femininity. In Tootsie, Dorothy Michaels, the female alter-ego of Michael Dorsey removes her wig at the end of the film in front of the camera, to end a near-endless array of sexual mix-ups and questioned sexualities and reveals their true (read: male) identity as a struggling actor. Mrs. Doubtfire (1993), another classic gender-bender played by Robin Williams, does not even need to disappear for the character to return to being a guy in good standing. The female alter ego is restricted to the realm of children's television, while the 'real' Daniel re-embraces his masculinity as a father. White Chicks (2004) ends with all gender, class and race 'confusions' cleared and the soldiers-turned-spies-turned-Marlene-Dietrich-impersonators in All the Queen's Men (2001) return as (male) war heroes. In all of these films, femininity is a temporary state-a reversal that by the end credits will be turned right again. With the characters returning to their 'own' gender as closure of the narrative enigma, these examples-rather than 'bending' the restrictive definitions of gender as the genre name suggests-reaffirm an essentialised gender dichotomy.
Adding insult to injury, mainstream gender-bending movies (re)produce spectacular masculinity in its embodied comedic performance. Underneath the more or less convincing feminine makeup and costume are increasingly attractive guys. With masculinity shining through 'feminine' performance even before their final reveal as fake, the transgressive potential of this play with identity is lost. There is no loss of virility for actors or characters, on the contrary: If anything, their short embodiment as 'women' only made these protagonists more assured and appealing in their sexuality.
Collaborative autoethnography is useful as it "preserves the unique strengths of self-reflexivity associated with autobiography, cultural interpretation associated with ethnography, and multi-subjectivity associated with collaboration" (Chang et al., 2013, p. 17). In relation to researching gender, it makes clear how mainstream media's 'playing with gender' has accompanied us longer (and more subtly) than we realised. Rather than focus on texts that became part of the feminist canon, we find ourselves taking a broader look and realising that gender-bending comedies in hindsight mostly highlight the existence of restrictive gender definitions rather than blur these. Whilst 'something was in the air' in the 1990s, resistance against gender conservatism was absorbed "through commodification…and incorporated into consumer capitalism. In this movement the subversive potential of gender was muted" (Van Bauwel, 2003).
This still begs the question how we were okay with the message of Tootsie, which, according to Showalter (1990, p. 371), seems to be "it's time for men to step in and show the girls how to do it?" Laughter, it seems, can be as much a response of release as of unease. We see too how we avoided quite a number of these films, rather than question or challenge them. It emphasises the connection between humour and gender as performance and the delicate line between entertainment and insult. It shows how the times have changed and how our tolerance for gender intolerance has greatly lessened. Lastly, it highlights how media production in general-and casting decisions more specifically-allow the idea of change to both materialise and evaporate. In these examples, cultural conventions were magnified into absurdities and yet remained difficult to challenge.

Netnography: Gender and Legacy
Where collaborative autoethnography works well to gain access to how times and social codes change, it also privileges one's own social and professional circles, and a personal perspective. Tracking struggles over gender as an audience researcher requires balancing and quite possibly recalibrating personal memories by engaging with the perspectives of others. In doing so, our own unease with gender-bending Hollywood film warns us to be careful. In taking care, in caring, feminist methodology (McRobbie, 1982) and the affective dimensions of audience experience meet. Nostalgia and irony make discussion, especially among friends, easy, whether texts are provocative, insulting or silly. The closer we get to the present and the further we extend our interest in what others have to say, the more difficult audience research can become. Online discussion of the (relatively recent) casting of female actresses for male legacy roles-from James Bond to Doctor Who-includes a fair amount of outright sexism, for instance. Netnography therefore comes with its very own challenges as well as affordances (see Table 2).
Elsewhere, one of us discusses how the storm of reactions to BBC's 2017 reveal of the new (female) Doctor Who is instructive. It demonstrates how changing the Doctor's gender perhaps was not the main problem irate fans and viewers had when complaining bitterly about the BBC's decision (Eeken & Hermes, 2019). Rather, gender becoming an unstable, undependable and unpredictable category was. Remarkably, other online discussions of the 'recasting' of legacy figures appear to follow a similar logic of resistance against the overturning of female-male difference. Inspired by the Doctor Who example, we put our methodological considerations of netnography to a 'screen test'-and share subsequent insights about this experiment below. While the multiple layers of online discussions cannot be done justice here, drawing on a small number of examples will clarify the methodological process and highlight exactly this layeredness as an advantage of netnography.
Methodologically, online communities offer a unique opportunity to study how gender is discussed spontaneously while a double performance is in play: People talk about gender and in doing so perform their own. This holds particularly true for anonymous online spaces. In the absence of other, clear identity markers, usernames, avatars and self-declared gender designations give a fair idea of the gender an online contributor wants to present. As a form "in which moderators downplay…or even eliminate…the participatory element of the technique" (Kozinets, 2010, p. 96), discourse and affect analysis of comments posted online fit this description of netnography well. For the questions of gender, identity and transgression that we are interested in here, the 'culture of anonymity' afforded by platforms like YouTube and Reddit (Kilgo, Ng, Riedl, & Lacasa-Mas, 2018), in combination with unrestricted access to posted material for non-participant researchers provide ideal conditions for observational ethnography. Perhaps superfluously, it should be noted that this is but one form of 'netnography.' It can also be combined with other types of online data gathering, with interviews or (open) questionnaires (Baym, 2000;Hine, 2000).
For the purposes of this article, we conducted a search for the keyword combinations 'gender + casting,' 'gender + recasting,' 'female recasting' and 'gender swap + film' across Reddit and found a vast amount of material. To limit the yield of our search, we reduced the data to threads with at least 10 comments. While this restriction will have excluded potentially interesting perspectives, the potential of online ethnography to give insight in interaction (and counteraction) among users is safeguarded. Cross-reading entries, we established that here, too, the discussion revolves around the apparent danger that legacy roles are in. Surprisingly, though, the 'famous' recasting examples mentioned at the beginning of this article appear as only the starting point of discussion. Viewers happily and angrily discuss all sorts of examples, including fully hypothetical ones ranging from The Lord of the Rings (2001) trilogy to Mary Poppins (1964).
As the second method in our experimental triangulation, netnography allows us to trace discussions of gender in anonymous, semi-public spaces. Whereas the absence of sociological indicators (such as gender, class, and ethnicity) is a loss for many forms of audience research, we see this as one of the strongest arguments in favour of netnography in anonymised online spheres. As elsewhere identity needs to be performed. On Reddit, however, identity needs to be performed in the general absence of profile pictures and personal information. Commenters therefore need to discursively present the authority of their opinion. While it is not necessary in discussing gender and gendered roles to return to the gender identity they are used to performing socially, we assume it to be the obvious move. Going by the material, there is an interesting difference between the women and men who lament the loss of legacy roles for the men who have always performed them and those who do not. Those chiding others for their narrow-mindedness disclose their gender identity less often.
For Doctor Who, James Bond, Mary Poppins, and other examples, conspiracies are a recurring theme across both the hypothetical and the 'actual' recasting decisions. Commentators clearly presume that gender swapping legacy roles is motivated by the need to cover up bad writing or to attract viewers to cheaper products. Hiring female actors, after all, incurs lower 'costs' as they (still) earn less. The comments read as both conspiracy theory and as a form of 'savvy viewing': Gender swapping a well-known character in a film, series or video game is often a stunt to distract people from not addressing fundamental issues or flaws with a work.
It's an easy publicity stunt.
They aren't doing it to make the show better but rather to seize on the hope some people with certain ideological leanings will support the show no matter how bad the stories actually are. If you aren't going to stay true enough to a source material to get the race or sex right, you most likely don't give enough of a shit to get any of the rest of the adapted work right too.
Layered underneath both implicit and explicit sexism, we recognise the tone and voice of the comments as not necessarily angry but unsettled. As the re-imagining of beloved characters produces stress, the online sphere becomes a 'safe' space to work through change together with others experiencing the same. Unexpectedly, and far more than in the reactions to the Doctor Who reveal, humour crops up: I'd love to see 'Michael Poppins' about a male nanny.
Omg, we need steven Carell do it.
Right?! You could even add it to the ethos. Like the Poppins idea is a gender fluid alien or creature that comes as a nanny when there are kids in need and gender matches the need. Some kids need a male nanny some need a woman nanny.
Humour and irony are as important in this online discussion as in our collaborative autoethnography. Here though, jokes are used to ridicule what is seen as overthe-top feminism. In interviews with female feminist interviewers or in a questionnaire format such material would not have surfaced. It needs to be taken seriously in order to deconstruct sexist commentary and resistance against the gender redefinition and fluidity that is high on our agenda. In addition, comments on the business models of the entertainment industry are not without merit or insight. Although the Reddit commenters would be astonished to find themselves grouped with Van Bauwel, they agree that commercial motives can acquire a sheen of feminist change. Then again, once processes of change have been set in motion, they might be difficult to contain. When doing ethnography, it is important and sometimes not easy to respect that gender, race, class or sexuality may not matter to others in the way they do to a feminist researcher. As Butler writes in the introduction to Gender Trouble: Feminism and the Subversion of Identity: "The breakdown of gender binaries is considered by many to be so monstrous, so frightening, that it must be held to be definitionally impossible and heuristically precluded from any effort to think gender" (Butler, 2014, p. viii). As feminist ethnographers we cannot discount the ideas, opinions and truths of those we encounter in research, regardless of what they are. Being open to definitions, intuitions and experiences that are diametrically opposed to one's own is imperative and an enormous challenge at the same time.
It entails that we find ourselves taking outright sexism seriously.

Textual Media Analysis: Gender and Solidarity
For our third and final example, we draw on the use of textual media analysis as a methodological approach (see Table 3). In doing so, we move from the private sphere of collaborative autoethnography and the semipublic sphere of netnography to the public sphere and news media. Methodologically and maybe somewhat surprisingly, we follow Hammersley (2006) in understanding textual media analysis as important to informing the first-hand experience and meaning making that ethnography charts-particularly in the context of the examples to be discussed here. Using available online print articles about the casting of cisgender actors in transgender roles, we will show how media discourse provides an implicit agenda for how we can discuss noncoercive and unsubjugated definitions of gender.
The examples drawn on here were published in the late 2010s and concern the casting of Matt Bomer as the transgender sex worker Freda Von Rhenburg in Anything (2018) and Scarlett Johansson as Dante Gill in Rub & Tug in the same year. The timing of these news items is significant. They come after Felicity Huffman, a cis woman, was complimented with her portrayal of the transitioning woman Bree in Transamericana (2005). Jared Leto, a cis man, had played Rayon, a trans woman living with AIDS in Dallas Buyers Club (2013) and Eddie Redmayne, another cis man, had been nominated for an Oscar for his performance as Lili Elbe, a transgender pioneer in The Danish Girl (2015). While these earlier performances were mostly discussed for (either or not) winning Oscars, the casting of Matt Bomer and Scarlett Johansson (both cis persons) became the starting point for a more fraught discussion. Johansson eventually withdrew from the project, Bomer persevered but proved an unconvincing trans character. Here, we are especially interested in how news articles represent the negative reception of Johansson and Bomer's casting and how online discussion is used as 'proof' without providing overview or contextualisation. Where newspapers simply report on casting decisions, not much agenda making appears to occur.
Practically, textual media analysis begins with and depends on sound data gathering. Its strength is ultimately in offering generalising conclusions. For this miniature, we conducted a double search query on the news archiving site Nexis for articles in newspapers, magazines and journals featuring the keywords 'anything + casting,' 'anything + casting + Matt Bomer' and 'Rub & Tug + casting + Scarlett Johansson' to collect discussion in the news media on how gender and casting decisions are discussed. Considering the international focus of this discussion, the chosen search keywords leave language consciously open, as the term 'casting' has been adopted in Dutch, German and in other European languages. In addition to an unlimited location setting, the and their loss of transgressive potential unlimited timeframe of our search query underlined the importance of these two casting decisions as they are recurrently referenced when transgender roles are mentioned. Nexis found only 38 media articles for our search query. If this had been a full-sized research project, we would have continued our search using other terms to check whether other terms are used in news media (as we expect that more was written about these films and their casting). Here we forego the possibility of generalisation and will instead reconstruct the logic of representation across the entire data set. Attention to gender in professional broadcast media is important. Here, as in online discussion among peers, agendas are set, and conventions reaffirmed. Mapping authoritative agendas emphasises "how users' agency hovers between the bipolar categories of producer versus consumer, and of professional versus consumer" (Van Dijck, 2009, p. 42). In addition, news media tend to underplay the 'savviness' of non-academic and nonprofessional viewers; Teurlings (2010Teurlings ( , 2018, amongst others, documents this phenomenon. Rather, the image of audiences, particularly the trans-activist one, is painted as highly emotional, unpredictable and volatile in international magazines, newspapers and blogs collected via Nexis, undercutting the political legitimacy of their claims. Regardless of the fact that the news media apparently did not pay much attention to Anything and Rub & Tug, recognising this mechanism is important. News media's highlighting of anger, outrage and activism undermines the potential for productive dialogue as-in however limited a manner-afforded by humour and ambiguity in the two earlier sections, exactly on the most public of forums for civic information and exchange. Textual analysis of these 38 articles shows how the opening up of restrictive, binary definitions of gender (which is at the heart of transgender activism), is hindered not by a concerted counter discussion but by news sources whose agenda becomes invisible behind rote professional moves, such as using incidental and unconfirmable '(wo)man in the street quotes' found online. From the perspective of feminist ethnography, the reconstruction of such backgrounds to ongoing discussion on other platforms fleshes out forms of opposition and resistance against open gender conventions that are not otherwise easy to identity.
As with the previous miniatures, a small selection of illustrative quotes from the media texts are presented here. We have chosen to highlight the international reach of these casting decisions. Checking on the validity of our search terms, we found that discussion of casting for transgender roles when part of reviews is mostly low-key and does not reference protest or activism. The selection below offers quotes from well-recognised news media and from news items that use terms such as 'outrage,' 'widespread backlash' and 'expected uproar' and hint at an angry and scary collectivity that threatens the social order: Ruffalo's comments came after members of the LGBT community lashed out about the casting of Bomer, saying that a real trans woman should have been cast in the film. ("Mark Ruffalo backs Matt Bomer,'' 2016) Scarlett Johansson has dropped out of her role in fact-based drama Rub & Tug after backlash from the trans community. (Lee, 2018) As expected, uproar ensued. This controversy comes around every time a cisgender male actor is cast to portray a transgender woman onscreen. (Mahavongtrakul, 2018) There was a sizeable backlash against the new drama on social media, due to Matt Bomer being cast as a transgender sex worker. Many suggested that the part should have been given to an actual trans actress, while actress Savannah Burton (herself trans) claimed that 'casting men to play trans women leads to violence against trans women.' (Hooton, 2016) The transgender community has repeatedly expressed outrage at cisgender actors playing the roles of transgender characters, like Eddie Redmayne did in The Danish Girl. When Twitter users called for Bomer's part to be recast, Ruffalo said the movie had already been filmed. ("Mark Ruffalo defends Matt Bomer," 2016) Media impressions of debate, comments and critique become a reference point for audience members discussing how gender matters from more or less involved points of view. As online discussions are taken up and reframed in other news media, a layered intertextuality is produced that suggests more unified and polarised groups (Fiske, 1987;Hiramoto & Park, 2012, p. 1) than we found in our methodological experimentation with netnography. When audiences are presented as a collective in news media, individual differences are obliterated, and issues are needlessly (further) politicised.
Interestingly, in discussion of who can and should play transgender parts, the transgressive potential of playing with gender is completely lost: "Anyone should be allowed to play anything" (Stolworthy, 2019), as Johansson phrased it in her response, reduces the political importance of offering embodied transgender identities as a reality beyond the fiction in entertainment media. Confirming the legitimacy of the bodies of transgender actors matters. While the gender-bending comedies of the 1980s, 1990s and 2000s turned (male to female) gender change into a joke, safely contained in the realm of fiction, casting cis actors for trans roles affects the world outside of the text. When cisgender actors play transgender roles, they appropriate what are only a limited number of possibilities for trans actors to 'play themselves.' However critical reactions might be for Bomer and Johansson, Leto and Redmayne added to their star text and benefited from playing trans characters. While neither has (yet) been voted 'sexiest man alive'arguably the ultimate benchmark for (American) male attractiveness-both are regarded as male sex symbols. Leto was featured on People Magazine's list of 'Sexy at Every Age: These Men Exude Sex Appeal that Spans Four Decades' in 2014, one year after his performance as transgender woman in Dallas Buyers Club. Redmayne tops Vogue's '10 Unconventional Alternatives to 2016's Sexiest Man Alive' list thanks to his "unique handsomeness and seductive voice" (Okwodu, 2016) in the year after his Oscar-nominated performance in The Danish Girl. Lists like these function as reassurance and as reinscription not just of the attractiveness but specifically of the spectacular heteronormative maleness of Leto, Redmayne and others that can withstand showing itself in feminine poses, clothes and availability. Featuring Jared Leto and Eddie Redmayne in addition boosts the image of the media that do so and their apparent openness to 'unconventional' choices.
Textual media analysis adds to collaborative autoethnography and netnography by allowing insight in how dominant gender power relations are maintained.
As the examples quoted above underline, discussion in major news media is not neutral. Cisgender actors are depicted as 'courageous' individuals for taking on trans gender roles while critical voices are characterised as 'outraged.' Potential support for transgenderism is undercut. Media texts such as the ones found by our open data query reflect and inform the complex negotiation of gender in everyday life-while at the same time re-inscribing gender in terms of social norms. The result is a lose-lose situation. Those who fear change and loss of identity will do so more, those who hope to gain a legitimate, public presence see their hopes squashed. Discussion of fiction affects real lives.
Approaching qualitative research of news textsand these particular casting decisions-after our exploration of collective autoethnography and netnography made us realise we missed the double and even triple layers we found in discussions of gender elsewhere. Instead of interpreting online discussion in terms of emancipation or solidarity, the news items reduced audience engagement to polarised positions. (Online) discussions of gendered casting decisions lose their complexity when looked at through the lens of professional media. When researching news media texts, their inbred logic of seeking oppositions and newsworthiness needs critical deconstructing in order for researchers not to follow the media's exaggeration of oppositions. When coding for recurring themes, we stay within the media logic and politicise and appropriate in the same way news media do.
Reconstruction of discussions of gender in the news affords a view of implicit public agendas that inform debate elsewhere. These agendas however are the result of a particular type of professional practice. Likewise, netnography and collaborative autoethnography are shaped by social and academic practice and thus have their limits. Together, though, they offer a start to taking feminist gender discussion further. Unsurprisingly, this case study-in-miniatures that follows discussion of casting ends with a plea to triangulate and work with rather than against the drawbacks of different methods. Rather than cast for the best possible method, we want to cast for change.

Conclusion: The Final Cast
Using discussion of casting decisions, this article addressed struggles over gender and gender distinction while pursuing a post-structuralist feminist agenda for audience studies. That agenda needs its two distinct parts to mesh: strong academic research methods need developing and testing and politically these methods need to serve the goal of loosening gender restrictions. We chose discussion of casting in order to step beyond binary definitions of gender. A spate of recent media productions that portray transgender and non-binary characters allowed us to do so. That is not to say that a multifold open definition of gender has become a common good. As of now it is still the prerogative of a minority to think of gender in such a way.
We compared autoethnography, netnography and textual media analysis for their merits and downsides from the perspective of intersectional feminism. Specific media texts were taken as a point of departure. Mostly, that was a choice of convenience: In real life, media texts are far less significant than in media and cultural studies. Tootsie, Doctor Who and Anything provided shortcuts to wide-ranging discussion, a great deal of which is sexist commentary and misogyny (which comes with the territory of being an audience researcher). In addition, feminist scholarship for us not only addresses a political agenda but also the need to recognise and reflect on one's own situatedness and respect how others are equally tied to specific histories and horizons of expectation (Harding, 2009;Warnke, 1987). With Bochner and Ellis (1996, p. 4), we believe that "interactive ethnography that refuses to close off further discussion or quiet the voices of the other" is what is needed.
Surveying three consecutive periods through the lenses of different methods and texts, we find that humour is exceedingly important in gender discussion outside of academia. Of course, humour and ridicule are, as Billig has argued, prime socialising mechanisms that warn us that we are transgressing important boundaries. Likewise, from a personal perspective, an ironic undertone allows for distance, for engaging and disengaging with gender transgression and fluidity at the same time (Ang, 1985;Billig, 2005). Ultimately, gender and gender distinction are not the bone of contention. The laughter stops when masculinity is threatened, whether directly because of women taking on male legacy roles, or indi-rectly by gender as a category becoming more fluid when transgender bodies are recognised as having rights. The more masculinity as a category is threatened, the more humour is replaced by anger. In the popular genderbender comedies of the 1990s, it is men impersonating women who return to being regular guys before the film ends. No one objects to this. Well, we might have but avoided the more sexist and racist of these films instead. When women take on male legacy roles, there is some joking but predominantly a fair amount of anger and conspiracy theorising comes to the surface. The anger and sexism we expected, the jokes were a surprise. How cis actors playing trans roles is discussed in everyday life, our third question, we do not really know. We used analysis of news sources to reconstruct discussion about this but found what we think is a caricature instead. Seeking urgency or newsworthiness, the small number of news items that were collected in Nexis all reduced controversy over casting decisions to activist 'outrage.' All methods have their strengths and drawbacks. From a feminist perspective, collaborative autoethnography is confrontational for the researcher herself, netnography demands respectful treatment of sexism, while qualitative media analysis has the unexpected drawback of reinforcing polarisation rather than querying implied definitions of gender as a closed binary system. Together though, they triangulate well and offer insight into the layered logic of gender definitions and the resistance against changing this. Rather than move to more anger and polarisation, our results suggest that humour might be explored as a methodological tool. Humour, like the magic circle in playing games, allows for entertaining extraordinary ideas (of which thinking gender in a more open manner for many certainly is one). Moving across personal and anonymous, 'real' and 'virtual,' popular and professional discussion, we explored how gender has been and continues to be instrumentalized in lived audience experiences. It is high time to reflexively and consciously bridge and connect positions by allowing for the idea that quite a number of our ongoing gender arrangements are actually a bit silly.