Tokenism Revisited: When Organizational Culture Challenges Masculine Norms, the Experience of Token Is Transformed

Extant research on tokenism has documented the adverse consequences for employees in minority positions and how women’s possibility of action is constrained in male-dominated contexts. We present an in-depth qualitative case study of a male-dominated organization in a masculine industry in which, despite all expectations, the experience of tokenism for minority women is ambiguous. Furthermore, these women also display a strong agentic role in an organization in which culture favours gender equality. This case reveals an aspect previously overlooked in studies of tokenism: the importance of organizational culture. By exposing and challenging the implicit masculine norm through its organizational culture, this organization actively engages in the change of gendering processes and contributes to establishing an alternative norm. Theoretical contributions show the impact of normative control on the experience of tokens, and how it provides a frame for action toward gender equality.


Introduction
From law firms to corporate boards, studies of the dynamics of tokenism reveal how women's possibility of action in male-dominated contexts are constrained due to adverse consequences of their token position. According to Kanter (1977aKanter ( , 1977b, tokens are the minority members in a skewed group, that is a group where there is a large majority and a very small minority, perhaps up to a ratio of 85:15. Kanter's theory has been investigated and corroborated in a variety of maledominated occupations and contexts, in which women are in minority positions (e.g., Chambliss and Uggen, 2000;Simpson, 2000;Singh and Vinnicombe, 2004;Childs and Krook, 2008; and the review by Watkins et al., 2018).
In view of extant research, the case study presented here appears to be a paradox. Ada (pseudonym) is an international consulting company active in a masculine industry (IT); it operates in the Nordic countries in which societal contexts are characterized by a gender egalitarian ideology (Inglehart and Norris, 2003), yet in an industry where men are in majority numerically, where gender discrimination is common and masculine norms are well established (Holth et al., 2017;EIGE, 2018). Just as in the rest of the industry, Ada employs a large majority of men, especially among their core professional group, the consultants, making it a clear skewed group, where women account for 22 to 13% across the different positions. Women at Ada could therefore be seen as having a token position. However, unlike what would be expected from Kanter's (1977aKanter's ( , 1977b work and later studies of women in male-dominated contexts (see review by Watkins et al., 2018), including recent studies in Nordic countries (e.g., Alvinius et al., 2018;Baublyte et al., 2019;Nielsen and Madsen, 2019), we find that women's experiences of tokenism are more ambiguous and, remarkably, token women seem to even have a strong agentic role in defining Ada's corporate practices.
We show in this study that Ada's case holds potential for theory development as it highlights a new dimension: the contingent aspect of the organizational context in the experience of token, when current theoretical efforts for the development of token theory concentrate primarily upon individual experience or interpersonal and intergroup interactions (see Watkins et al., 2018). Ada is, indeed, an organization that actively nurtures an organizational culture that promotes gender equality. In other words, Ada is a male-dominated organisation that supports the norm of gender equality. Norms are ideas and behaviours that are expected, and that are seen as normal and suitable, within a given context. This particularity of Ada enables us to differentiate between two notions that have been entangled in previous studies: the majority group and its norm. The case shows how not only minority position and status hierarchies affect women in token positions; the norms promoted by the dominant centre in the company do so as well. Therefore, our research question investigates how the organizational culture influences the experience of token women.
This research is based upon a qualitative case study using in-depth investigation methods over a period of a year in which we were inspired by work on engaged scholarship (King and Learmonth, 2015) and tempered radicalism (Meyerson and Scully, 1995). Our case analysis first presents the various organizational practices that promote gender equality in the company; it then describes the unexpected ambiguous experiences among token women within this organization. Our analysis links these ambiguous experiences to the efforts of the dominant centre to expose and challenge the masculine norms in the industry and society, thus, spreading knowledge about the effect of tokenism and providing a normative framework for the application of this knowledge.
Contributions from this study are both empirical and theoretical. Our empirical contribution presents the case of a company active in a masculine industry that promotes gender equality and in which women in a token position do not appear to share the same experience as found in previous research. The presentation of the various practices in place provides a rich description of how organizations can undo gender (Martin, 2006;Kelan, 2018). By adding the dimension of normative control, the theoretical contributions further previous research that suggests numerical minority and status hierarchies are the principal explanations for the experience of tokens (Watkins et al., 2018).

Gendered organizations and tokenism
There exists a prevalence of pervasive norms in both organizations and society that favour men and masculinity (Acker, 1990;Connell, 1995).

Implicit masculine norm in organizations
Gender and organization scholars have identified an important factor in the reproduction of masculine hegemony: the implicit nature of the norm favouring men and masculinity. For example, Acker (1990) argues the seemingly gender-neutral notion of 'a job' is implicitly gendered since it rests upon the assumption of a division of responsibilities: for example, jobs are organized as though the worker is always available and work-oriented. Men usually correspond to this ideal more easily. Thus, long-work hours and companies' expectations of employees' (total) availability favour those with no private commitment (Rutherford, 2001;Lindgren and Packendorff, 2006).
Moreover, the ideal worker not only corresponds to a man's body; it also relates to notions associated with masculinity: for example, this is apparent in management where there is a strong correspondence between perceptions of the ideal manager and perceptions of qualities often associated to masculinity such as, for example, assertiveness (e.g. Wajcman, 1998;Collinson and Hearn, 2014). In relation to this unquestioned seemingly neutral managerial norm, women managers are constructed as being deficient (Wahl, 2014).
The preference given to masculinity is embedded in cultural, societal, and professional norms of the context in which organizations operate and is unreflexively repeated through organizational culture (e.g., Gherardi, 1995;Rutherford, 2001;Aaltio and Mills, 2002). These cultures often reproduce societal gender symbolism attached to positions and type of jobs. For example, technical jobs and tasks calling for physical strength are constructed as masculine (Bruni and Gherardi, 2002;Zanoni, 2011) or, as Abrahamsson (2002) shows in her studies of gender and organizational change in industrial contexts, whatever is perceived as masculine is deemed to be important and when something becomes important, it is also perceived as masculine.
The preference for men and masculinity in organization is done through processes by which gender differences and hierarchies are unreflexively (re)produced; such processes are called 'gendering processes' (Acker, 1990). Gendering processes both reflect and reproduce, gender divisions (e.g., separations in space and in tasks) through policies and formal practices (e.g., defining criteria for promotion and recruitment); organizational culture that influences beliefs and behaviours (e.g., through gender typing of occupations or tasks); and through everyday interactions among employees and identities of organizational members (e.g., professional identities).
In sum, through gendering processes, an implicit masculine norm reproduces gender divisions and attributes higher status to masculinity in society and organizations alike. As long as the masculine norm remains implicit, it is difficult to challenge; therefore, the underlying gender hierarchy remains unchanged.

C. Holgersson and L. Romani
On tokenism Kanter's (1977aKanter's ( , 1977b studies are an early illustration of gendering processes that occur in organizations: constructing gender division. She presents the case of an organization in which the numerical advantage of men in management positions gives them greater power and influence: both materially and discursively. Kanter develops the concept of token position, that is, when a group represents a clear minority (roughly 15%). She then theorizes about the relationship between the token's experience related to three phenomena linked to this minority position: visibility, contrast and assimilation, and how they contribute to perpetuating the existing organizational culture and privileges of those in majority. According to Kanter (1977aKanter ( , 1977b, the effects associated with token positions emanate first from the heightened visibility token women experience in relation to the dominants: that is to say, the male majority group. This visibility entails performance pressures: being high performing in order to prove that women deserve their position, while trying to reduce their exposure by becoming socially invisible or not out-performing the dominants. The second effect is related to the stereotypical expectations token women are confronted and trapped into: for example, being limited to certain positions and types of work that, in turn, hampers women's career opportunities. The third effect is linked to the dominants feeling threatened or uncomfortable around tokens and, as a result, emphasizing their commonalities and the token's difference through informal exclusion and loyalty tests in which the token woman must show her loyalty to the dominant group: for example, by showing gratitude for being part of the group, or by excluding other women. In this way, the dominants are able to preserve the group and its culture. In their review of the vast amount of research investigating Kanter's tokenism theory, Watkins et al. (2018) stress that numerous studies confirm the effects of the token position, even for minority groups up to 20% (e.g., Flores, 2011). Yet the experience and consequences for token employees differ depending upon gender and status. For example, men's token positions in femaledominated context benefit from, or experience few, adverse effects from their minority position since they tend to enjoy a higher status linked to masculine hegemony (e.g., Williams, 1992;Simpson, 2004;Gustafson, 2008). Some scholars (e.g., Zimmer, 1988;Yoder, 1991;Williams, 1992) have already brought attention to the need to go beyond numbers alone since other factors that privilege the masculine and devalue the feminine intervene: for example, a masculine norm in society. Watkins et al. (2018) call for further token theory development by including contextual factors that moderate the significance of numbers. Simpson and Lewis (2012), who explore the impact of the invisibility or implicitness of masculine norm on the experience of token women in leadership positions, argue that Kanter's 'majority group' not only has a numerical dimension; it is also a dominant 'centre' that encapsulates a masculine norm. They also argue that norms in general are understudied in the research on tokenism. The case of Ada provides such an opportunity.

Case presentation
Ada is an international consulting organization active in the IT industry. It has grown to over 1,000 consultants internationally and over 600 at the site of our investigation: the home office in a Nordic capital city. At the time of the study, a third of the group of partners were women; however, the top hierarchical levels (top 100 managers) were strongly male dominated. In Ada's home office, 25% of all employees are women. The core professional group of Ada (consultants) is also strongly male dominated, women ranging from 22% at entry level to 13-16% at senior levels (see Figure 1). Ada's gender distribution is rather similar to other IT consulting companies in the Nordic countries, offering the case of an organization in which women in senior positions are token.
The company's image is rather positive in the Nordic region. Ada has received awards and recognitions for employees' satisfaction, leadership, creativity, and contribution to the IT industry. Ada frequently appears in the national press and is active in its industry as a shaping actor; in particular, it is known as a gender equality champion. As of January 2018, about one hundred comments relating to Ada are available on Glassdoor, a site on which employees can anonymously rate their company. These comments are overwhelmingly positive, generally depicting the organization as an outstanding place to work. They also mention that employees need to invest considerable time and energy, which leads to great development and opportunity. Negative comments reproach this large number of unpaid working hours connected to employees' involvement in the strong organizational culture, which is sometimes called a 'sect'. In sum, Ada displays several characteristics that imply the likely unreflexive reproduction of an implicit masculine norm: namely, male-dominated management levels, long hours, and strong commitment requirements (Acker, 1990;Blair-Loy, 2009).

Tokenism Revisited 3
© 2020 European Academy of Management

A case study for theory development
Qualitative methods, and especially case studies, have an established record for theory development; the work by Kanter (1977aKanter ( , 1977b) is a forceful illustration of this. Single case studies can provide a unique setting in which a dimension previously overlooked becomes salient. In this matter, it is the uniqueness of a casenot its representativenessthat provides the potential for theory development (Lervik, 2011). In our study, we build upon a case that presents an intriguing feature: the organization is embedded in a masculine industry (IT and management consulting), is overwhelmingly employing men, yet female employees do not describe their experience as would be expected from both theory and extant research. Our case study contributes to theory development by providing a contextualized explanation of social phenomena (e.g., Ragin, 2009;Welch et al., 2011). A contextualized explanation 'often starts with a surprising contrast, triggered by the realization that an observed outcome is different from what had been anticipated, provoking the question, 'why not X?" (Welch et al., 2011, p. 748). We wondered: Why don't senior women consultants encounter what Kanter describes as the experience of tokens? Our case leads to re-consider the context of these women in token positions. It reveals the organizational processes that expose and challenge an implicit masculine norm and establish an alternative norm, and how these are decisive dimensions that define the experience of token employees. Therefore, this case illustrates how the study of organizational context helps the development of token theory.

Data collection
From the very start, our contact persons -Adam, Anna and, to some extent, Maria (pseudonyms, as for all other names)expressed their willingness for their views of the organization to be challenged and saw collaborative research as an opportunity. We initiated our collaboration in the Spring of 2016 and the core data collection took place in 2016 and 2017. Table 1 provides an overview of the range of data collected. Almost all interviews, observations, and meetings took place in the facilities of Ada.
During our study, we developed a particular relationship with Adam; he was trained in gender studies and understood and supported our search for empirical data on the topics of interest. When he deemed appropriate, he forwarded us emails that were sent to all Ada employees or copied a chain of emails or messages that touched upon gender equality. These messages are not reproduced here; rather, they are paraphrased to protect authors' anonymity (see Boellstorff, 2008 for similar practices).

Analysis
Our analysis adopted an abductive process (Dubois and Gadde, 2002) typical for qualitative case study method. In a first step, we independently went through all collected data and created 'experience-near' first order constructs (Geertz, 1975). Organizational narratives and practices related to the topics of gender equality and inclusion were strongly present in these first order constructs. Also, first C. Holgersson and L. Romani order constructs based upon the interviews of token women simultaneously revealed ambiguous experiences. In our next analytical step, we used Kanter's (1977aKanter's ( , 1977b three phenomena: visibility, contrast, and assimilation -and their ambiguous consequences on token women at Ada to elaborate upon second order constructs. We also used a framework by Rutherford (2001) for the analysis of the organizational culture at Ada. Organizational culture is here understood as shared and conceded symbols, beliefs and patterns of behaviour among organizational members. Rutherford's (2001) framework suggests that organizational culture is expressed through a number of aspects that can contribute to the exclusion of women and inclusion of men: the background of the organization (location, ownership, employee profile); physical layout and artefacts; gender awareness; management style; informal socialization (networking); sexuality (sexualized cultures, harassment); language and communication (metaphors, humour, meetings, talk); work ideology (private-public divide) and time management (long hours culture). Rutherford's framework builds on the work of scholars theorizing organizational culture and gender such as Gherardi (1995), and is adequate to identify organizational practices embedded in gendering processes (e.g. gender typing and gendered professional identities) that reproduce gender divisions and hierarchies (Acker, 1990). As we present in the findings, the analysis identified that the organizational culture of Ada and its organizational practices are means by which top management exposes and challenges gendering processes.
The following section presents these practices following Rutherford's dimensions. Thereafter, the ambiguous experience of token women is presented according to Kanter's (1977aKanter's ( , 1977b three phenomena. These presentations reveal how the exposure and challenge of gendering processes at Ada and the ambiguous experience of token women are intertwined.

Our role as researchers
Our research approach is inspired by Meyerson and Scully's (1995) notion of 'tempered radicalism': insisting upon developing a critical friendship while exploring the practical concerns, which the participants raised (King and Learmonth, 2015). In our relationship with Ada, we found with trial and error the right balance of critique that our contacts were willing to hear. This allowed us to pose questions to kindly disrupt and challenge, and hence, potentially trigger a process of change (King and Learmonth, 2015). The first author became the go-to person for finding norm-critical publications for Ada's training. As we write, we are too early in the phase of providing feedback to Adam and Anna, as well as to the Tokenism Revisited 5 organization, in order to assess the potential impact of our critical friendship.

Organizational culture promoting gender equality
We offer here a succinct presentation of Ada's artefacts, narratives, and practices by using the framework that Rutherford (2001) developed in order to capture in what way organizational culture contributes to the reproduction and change of gender divisions and hierarchies, thus, gendering processes at Ada. We focus upon the official narrative that is perpetuated at Ada, in particular, dimensions touching upon gender equality.

Background, physical layout, and artefacts
Ada's modest beginnings in a rented unattractive apartment in the suburbs are not forgotten, despite its current fancy offices in the very centre of the capital business district. Ada's offices occupy several floors of a stunning office building and the facilities are organized to cultivate the image of an outstanding company. For example, passing Ada's doors feels like entering a gigantic contemporary piece of art; the stairway connecting the different floors is packed with recognitions and awards such as Best Place to Work or CEO of the Year.
The facilities are inviting with multiple meeting rooms and large tables at the dining area, yoga mats in the cloakroom, child-friendly changing tables and play room. We see men and women wearing a variety of clothing styles from business attire to informal sweatshirts: notably, in a way that is not particularly sexualized. The bathroom doors display a sign that reads: This is an All-Gender Restroom with a pictogram representing women, men, and transgender people. In brief, Ada provides an environment with outstanding artefacts, some of which explicitly challenge gendering processes.

Management style
The organization emphasizes equality, informality, and flexibility in interactions with colleagues, and in decision-making. Employees are encouraged to run their ideas with colleagues instead of asking for approval from a senior manager, as Peter, a consultant explains: 'In the culture we want everyone to think [independently] and then just involve people around them that they think can give good input, and [then] make the decisions themselves instead of having to go to some superior that makes the decision'. Moreover, salaries follow a set grid that is available to all, and which defines consultants' pay independently of their assignment. and taken into account in the yearly evaluation of consultants, adding 'but it is a challenge, especially since more and more are becoming parents'. Thus, there is an awareness of the tension that lies in the long hours culture, which relies upon the expected symbiosis between private and work spheres: on the one hand, it is a much-valued commitment to the development of Ada, on the other hand, it disadvantages Adaers that have other obligations, yet not for women specifically; but rather for those employees who are parents.

Gender awareness
Ada is firmly anchored in the IT industry and continues to mainly recruit highly educated persons from top national technical universities. Ada's employees in charge of recruitment are particularly attentive to identifying and recruiting women, as communicated in their PR material: 'since we know there are fewer women within the IT and technology sector, we search harder to find them'. Since 2013, gender equality is one of the company's top priorities. Magnus, who is involved with recruitment, explains: 'it's on the agenda every week; every week we have a follow up regarding how we are doing with gender equality'. This results in Ada annually recruiting about 30% of female consultants (currently the percentage of women graduating from the programmes, which Ada targets), which should eventually increase the total proportion of women in Ada's core operations.
Ambassadors are present on selected university campuses, and Ada has launched several initiatives to increase 6 C. Holgersson and L. Romani women's interest for the IT industry. Events on gender equality appear monthly on Ada's Facebook page: often internal or public workshops or training. In brief, Ada is an organization that displays a very clear engagement for gender equality: both internally and in its initiatives for the IT industry.

Communication, language, and sexuality
Another person working in recruitment tells us: 'Ada is a company that is value driven'. For example, Ada values, what it refers to as 'caring' and demonstrates its commitment to the well-being of its employees. Viktor, who holds a relative senior position, tells us that Ada terminated an assignment as soon as he, as with the previous consultants in that mission, started to experience too much stress due to tensions at the client's site. Anna also tells us about a female consultant who was sexually harassed at the client's site. Ada stopped the collaboration and demanded the client send the team of employees to training on diversity if they were to resume working with them.
We received several illustrations of the values that Ada officially promotes, including gender equality. For example, the CEO sent an email to all employees before a large company event and we paraphrase his core message as: it unfortunately happens that women (and men) are subject to unwanted physical contact and propositions during conferences and parties. There is nothing, nothing that can justify this form of behaviour. (…) Be aware of what is happening around you, take responsibility, and report any intolerable action. Similarly, one of the partners sent a collective email in the period prior to resuming the bi-annual round of Ada employee evaluations; we paraphrase that core message: from what I have previously seen, women tend to receive comments on their nice personality more often than men do; it's because our society is not used to acknowledging women leaders'talents and skills. We can start solving this problem today, in our evaluations, by paying attention to the terms we use. In other words, this partner names essentialist discourses of gender, and frames women as being equally competent as men. This practice performed by a representative of the dominant centre of Ada openly challenges a masculine norm.
Furthermore, new practices are introduced by management to support values promoted by Ada. For example, Suzanna describes how a new concept was communicated by the CEO during the annual meeting, which introduced new, more sensitive topics into the extracurricular activities and internal processes for the year to come: 'now we had one year practicing being vulnerable, talking about personal issues, talking about discrimination, about things we usually don't talk about.
(…) saying from the start [to a client] I am uncomfortable with this'.
In summary, the presentation of the multiple dimensions of the organizational culture at Ada highlights how a masculine norm is exposed in the gendering processes, which usually unreflexively reproduces gender discrimination. Gender division and the construction of masculine symbols are resisted at Ada; additionally, specific practices are introduced to promote gender equality.

Ambiguous experiences of tokens
Considering the male-dominated IT and consulting industry, we expected to hear about the pressures and dilemmas linked to visibility, contrast, and assimilation from female employees in a token position. We did hear about these; however, they were in more ambiguous ways than we had expected.

Visibility
Heightened visibility is the first phenomenon of the token position, which Kanter identifies. Visibility means that token women attract more attention and are, thus, under more scrutiny than their male colleagues. This could result in performance pressures, which lead to over achievement. These pressures come from tokens feeling the burden of being perceived as a representative of their group. This also adds the pressure of having to prove their competence despite being a female: their competence tends to be eclipsed by their gender. Lastly, performance pressure emanates from avoiding making dominants look bad when the token's (over) performance is visible.
In our empirical material, several senior women comment on being representative of 'all' women. A woman partner comments on her visibility, reflecting back upon her situation when she was the only woman among partners, and when gender equality was not as high on the agenda: 'it was classic. As a woman in a maledominated context, you are always the woman: you represent the female perspective; you represent all women'. Today, the company website displays images and videos in which women and men appear in almost perfectly balanced groups. On stage at the company event we attended, the gender distribution of speakers was even. This is a purposeful construction of Ada's image. However, even though women receive more exposure than their men colleagues, women did not touch much upon topics of heightened visibility.
Overachievement is a typical response to performance pressure due to visibility. Ada is a highly competitive company and everyone is subject to performance pressure. Adaers on all levels are aware of the risk of burnout, especially in relation to women. For example, evaluators

Tokenism Revisited 7
are aware of extra performance pressure on women. At one of the sessions of the annual performance review, women were discussed in view of how they could be helped to minimize stress. Despite the fact there were five times more men evaluated than women, there were no comments regarding burnout risk for menwith the exception of one. He was labelled a workaholic and the group mostly worried about what messages he was sending to others; thus, it seems that the evaluators are aware of women's tendency to overachieve and they strive to support their female employees and are attentive to controlling some men's behaviour that could signal a (masculine) culture of excessive work. Adaers interpret this gender awareness practised in evaluations as one of the reasons that support the current balance in length of steps in career progression: on average, men and women spend the exact same number of months in the same position before they are promoted. Another consequence of visibility is the pressure of avoiding actions that could make the dominant group look or feel bad. This pressure is connected to fear of retaliation. A female manager told us how women have been advisedby mennot to push the gender equality agenda as much: '[during evaluations] women see a large proportion of [anonymous] comments that we might do better if we piped down a little' […]. 'Don't be so loud or committed with this gender equality thing, it would be better". Although the interviewee did not express fear of retaliation, the threat seems to be present. Other typical responses to heightened visibility are limiting women's visibility by, for example, minimizing their feminine attributes and avoiding conflicts and controversial situations. We did not, however, see any particular signs of women minimizing feminine attributes, and women did not appear to be afraid of addressing sensitive topics, such as gender equality.

Contrast
Due to the contrast in relation to the token, dominants become more aware of their commonalities and perceived the token as a threat to this commonality. The dominants respond to this by exaggerating both their commonality and the token's difference, resulting in what Kanter calls boundary heightening. Taking opportunities to remind the token of their difference and excluding the token from informal meetings and through loyalty tests are examples of this exaggeration. Loyalty tests consist of tokens allowing themselves and their category to be a source of humour for the dominants or tokens adopting the role as gatekeeper, thus, counteracting other tokens and working for change. By exposing the token to a loyalty test, dominants assure themselves the token will not challenge them.
There are no clear examples of boundary heightening in our empirical material; rather, it is the opposite. Ada's environment provides a space in which gender awareness is high among employees and in which men do not tend to openly emphasize their commonalities in contrast to women. Ada's organizational artefacts and practices tend to focus upon limiting gender differentiation instead, as illustrated in the preceding section.
Furthermore, loyalty tests, in the form of jokes and gatekeeping, were not mentioned among the interviewees as being part of the reality at Ada. A woman consultant explained how salacious jokes are frequent among men at client's and how this provides a strong contrast to Ada's environment in which this kind of behaviour is rare. Also, there were no stories of gatekeeping. Instead, senior women interviewed for this study were all highly committed to promoting women, senior partners as well as those on lower managerial levels. Furthermore, they are very actively and collaboratively involved in working for change. One of the female partners recalls: 'I think that these women [driving gender equality] all had key positions (…) [and] have very strong feminist values (…). We got the opportunity, but not for free. We have really had to fight, and have had each other [as support]'. Another female partner provided a story that illustrates how women on different levels collaborated in the initial stages of their work for change. They collected quotes from women at Ada that described how they were treated differently, for example: 'my judgement is often questioned' or 'I'm often interrupted by my male colleagues'. When these quotes were presented to the board along with statistics, the female partners said the quotes could just as well have been theirs. The chairman was shocked; he never imagined this discrimination could happen to such highly talented women. This was a decisive moment for Ada's corporate involvement with gender equality. Hence, token women at Ada do not seem to have to dissociate themselves from other women and gender equality issues but can rather join forces together with their male colleagues in work for change.

Assimilation
Assimilation is the third phenomenon linked to tokenism and involves tokens being interpreted in relation to preconceived ideas about their category (e.g. gender) and, therefore, face stereotypical assumptions and attributions. This results in what Kanter calls status levelling: when tokens are mistaken for someone on a lower level in the organizational hierarchy. This also leads to role entrapment by which tokens find themselves set aside in specific roles that perpetuate stereotypes and limit their career progression.
Assimilation of women with non-technical skills is a form of attribution, which female consultants at Ada 8 C. Holgersson and L. Romani experienced. They explain how male colleagues (especially at client locations) often assume they do not work with back-end, which has a male/technical connotation (Torgersson and Norén, 2016). A senior female consultant mused about her presentation strategy: she always finds a way to mention her degree in advanced maths and physics when meeting new clients. Despite recent increases in female consultants at Ada and its efforts to welcome more women into coding, women are more often found in roles linked to project management, design, user interface (front-end) than in more technical system development and coding (back-end). Also, the only session with two men on stage at the observed allemployee event was devoted to mathematical solutions to coding problems. Women are seen as competent in many ways, yet coding (and especially back-end) appears to continue to be gendered as masculine and is an area reserved for men. Women engaged in the promotion of gender equality could be seen as an expression of role entrapment: in the role of driver of gender equality issues. For example, one of the few female senior executives was given the responsibility for the large gender equality initiative adopted by the board in 2013. To a larger extent, women are involved in the gender equality activities. With time, however, the number of men engaged in these activities has also increased. For example, the industry-wide network initiative targeting other IT employers is led by a man; and during the last all-employee event in our fieldwork, a man was leading the session on gender equality. The (male) CEO is also a visible advocate for gender equality issues, both internally and externally. He announced a new initiative Focusing on Men's Recognition, Support, and Action for gender equality. This aims to counteract the tendency to have mostly women Adaers dealing with gender equality issue. He clearly explained in an all-employee email that this issue concerns everyone.
In summary, the experience of tokenism at Ada is more ambiguous than in other studies of women in token positions. The women at Ada are, indeed, visible; however, this does not appear to be a burden. Although performance pressures are present, efforts are made to counteract this risk. There are no clear examples of boundary heightening and loyalty tests resulting from the contrast effect. Instead, women support each other and collaborate with men to understand gender equality issues. Nevertheless, there are some indications of assimilation: when women face stereotypical perceptions and are overrepresented in driving the gender equality agenda.

Tokenism and norms
The experiences of token women at Ada neither fit perfectly the expectation from theory, nor the multiple illustrations provided by previous research. We suggest that one explanation for this situation is that the organizational culture at Ada exposes and challenges the implicit masculine norm.

Norms and the experience of token women
Several studies have pointed out that Kanter's theory on tokenism overlooks the impact of societal norms. They argue that token women have negative experiences resulting from visibility, contrast, and assimilation due to the masculine norm in society and organizations (e.g., Ridgeway, 1991). This implicit masculine norm also explains that token men experience the glass escalator (e.g., Williams, 1992). Lewis and Simpson (2012) argue that the implicitness of the masculine norm is precisely what contributes to its perpetuation. This norm can be reproduced because it is kept implicit and, consequently, protected from being openly challenged.
In the case of Ada, the implicit masculine norm within the industry is exposed and challenged. Kanter (1977aKanter ( , 1977b, as well as Lewis and Simpson (2012), do tell about 'exposure' of the norm, yet this exposure is done by tokens only, at the periphery, which contributes to reinforce the dominant centre. Indeed, Lewis and Simpson (2012) explain that despite tokens exposing the norm, they also conceal it in their efforts to deal with tokenism effects: for example, masking or distancing themselves from feminine attributes, accepting assimilation, and accepting loyalty test. In Ada's case however, tokens do not conceal the norm, instead they expose the norm in their work for change. Moreover, they do not distance themselves from feminine attributes, nor assimilate entirely to a masculine norm. Furthermore, the dominant centre, the top management team of Ada and the partners, also challenge the masculine norm.
This exposure by the dominant group is a distinctive feature of the Ada case. We see how top management engages in processes of normalizing gender equality: it is expected from Adaers that they promote gender equality. Women are not framed as being 'different' employees; they are showcased to provide gender equality in the company's internal and external communication. Gendering processes are exposed (e.g., in emails on biased assessments before evaluation rounds; at specific initiatives with men), and certain discourses are challenged: masculinity constructed in an oppressive way, such as attached to potent sexuality (e.g., emails on sexual harassment prior to events) or excessive long hours (e.g., concerns over a male consultant's possible workaholism as being detrimental for the organizational culture).
Ada offers the case of an organization that not only exposes and challenges the implicit masculine norm; it Tokenism Revisited 9 also aims to replace it with a gender egalitarian one. We argue that women, despite their token position, are not limited to working for change at the margins (as discussed by Kanter, 1977a, 1977b, or Lewis and Simpson, 2012 since a gender egalitarian norm is promoted by the dominant group. Building upon this gender egalitarian norm, they can use it to foster and support change. We advance that this gender egalitarian norm contributes to explaining the ambiguous experience of token women and their change agency which, in turn, further supports the normalization of gender equality. Top management thus actively develops symbols for collective actions, and, through multiple initiatives, exposes and challenges the masculine norm in society and the IT industry alike, and conveys a gender egalitarian norm. Employees may or may not adhere to this discourse; however, the official narrative of what characterizes the organizational culture at Ada is made explicit to all. This is a form of control, called normative control, since the explicit narrative orchestrated by the management team constitutes a clear point of reference for the behaviour of all employees. Normative control is closely linked to the management of organizational cultures (Kunda, 1992;Willmott, 1993) and requires employees to 'develop self-images and work orientations that are deemed congruent with managerially defined objectives' (Alvesson and Willmott, 2002, p. 619). This form of control is not a physical or economic form of coercion, but rather, control based on the expectations that employees identify with company values, norms and goals. By emphasizing that Ada is value-driven and carefully managing the organizational culture, Ada's management requires employees to internalize a gender egalitarian norm, by which the underlying beliefs that guide their behaviour is influenced (see Müller, 2017).
A contribution to tokenism research brought by the Ada case is the differentiation between the alleged effect of a majority group and a dominant norm. When previous research pointed to this differentiation, they did not study an organizational culture that challenges the masculine norm and promotes a gender egalitarian norm. In Ada's case, the numerical advantage of men does not lead to clear tokenism effects since it is not (fully) supported by a masculine norm. The Ada case specifies the relationship between the organizational culture, through its normative control, and experiences of tokens. In an organization dominated by men, yet where the organizational culture explicitly controls its norms in favour of gender equality, the experience of tokens is transformed and their room for manoeuvre expanded, allowing them to become agents for change. The effects of tokenism are thus not dissolved but instead take on a different form opening up for potential change.

Knowledge about tokenism and normative control
The Ada case provides an additional insight to our understanding of tokenism: the importance of knowledge about tokenism. Contrary to expectations, token women have not been forced through loyalty tests to dissociate themselves from other women and gender equality issues. Incidentally, Kanter (1977a, p. 238) suggests that token women who are highly identified with 'the feminist cause or with their own social group', can develop a strong alliance and resist the pressures of the token position. In the case of Ada, we see senior women highly committed to gender equality. We argue, however, that it cannot be assumed that commitment will automatically provide individuals with specific knowledge. In the interviews, it appears that it was rather their knowledge about gendering processes, in particular about tokenism dynamics, that wrought the key to their alliance. Therefore, to Kanter's claim that identification with other womenor with the 'feminist cause'enables token women to form alliances and share the burden of tokenism, we add that having knowledge of gender dynamicsparticularly about tokenismalso helps mitigate tokenism's negative consequences.
However, knowledge does not necessarily result in changed behaviour (see Noon, 2018). Kanter (1977a) suggests educating the dominants about tokenism as a measure for change. This knowledge, however, could be used to reinforce tokenism effects in a company with a masculine norm: as an act of resistance towards gender equality. In addition, this knowledge does not necessarily translate into gender-aware behaviour, since much of the 'doing of gender' is, indeed, unreflexive (Martin, 2006). However, at Ada, through normative control, the norm of gender equality and the level of awareness of and resistance against (unreflexive) gendering processes seems to affect what can be said and done. For example, evaluators were aware of possible extra performance pressure put upon women and used this knowledge to tailor their monitoring of women's performances. We, therefore, add that the conditions for applying this knowledge should be considered. The organizational culture at Ada, that conveys a gender egalitarian norm, provides direction for how knowledge on tokenism should be applied and guides men (and women) when putting their knowledge into practice.
In brief, Ada's work to share knowledge on the negative consequences of tokenism (e.g., through emails and social media, training, events and other projects specifically supporting gender equality) first renders gendering processes explicit: it exposes and challenges the implicit masculine norm and diffuses knowledge on the negative impact of tokenism. Second, Ada's management, through normative control, establishes an organizational culture that favours gender equality and endorses behaviour that mitigates the consequences of 10 C. Holgersson and L. Romani tokenism. We argue that these two phenomena affect the experience of token women.

Conclusion
The case of Ada provides a description of a company in which token women in senior positions convey strong agency for change instead of previously documented limited room for manoeuvre. The aim of our analysis has been to understand these ambiguous consequences of tokenism. We show that the purposeful management of Ada's organizational culture by which a masculine norm within industry and society is exposed and challenged, and an alternative gender egalitarian norm is introduced, affects the experience of token women employees.
We also find that knowledge of gendering processes, including an awareness of tokenism dynamics, among token women and powerful men, is important in order to understand the ambiguous experiences of token women. We, nevertheless, argue that it is the normative control at Ada that directs in what way this knowledge is put into practice in order to challenge gendering processes. This case exposes how normative control promoting gender equality is performed through multiple organizational practices. In addition, it shows collaborative actions between token women and the (male) dominant centre: token women are not the only agents of change. This provides an illustration of how women and men can undo gender: that is, reduce gender difference and gender inequality (Kelan, 2015(Kelan, , 2018. While previous theory development has centred on individual experiences, numbers and status hierarchies (Watkins et al., 2018), our theoretical contribution primarily identifies the importance of organizational culture and norm(s). This case shows how the organizational culture, by resisting (unreflexive) gendering processes, transforms the experience of token women. Ada's case, therefore, insists upon the necessity to add contextual dimensions on an organizational level. An additional theoretical contribution touches on the role of norms. Ada's case illustrates how a normative control sanctions the use of knowledge on tokenism in favour of gender equality. Therefore, the case of Ada also emphasizes the strength of norms and normative control for the experience of token women.
It is, however, important to acknowledge the impact of the societal context in Ada's case. The gender egalitarian norm conveyed by Ada's management is indeed in line with the overall gender egalitarian ideology present in the Nordic countries. The normative control exercised by Ada's management would probably have been different in a societal context where discourses on gender equality were not as prevalent, and where arrangements in society, such as legislation and the welfare system, did not promote equality between women and men. Yet, the Nordic countries are also the place of the gender paradox: the co-existence of a strong gender egalitarian ideology alongside gender inequalities in the labour market (Teigen and Skjeie, 2017). The Nordic paradox highlights that a societal gender egalitarian ideology does not guarantee gender equality in organizations. The case of Ada, nevertheless, illustrates how this ideology can be drawn upon, and through normative control, be used in order to promote gender equality at work and transform the experience of token women.

Statement of conflict of interest
The authors have no conflict of interest.