The Risks of Feeling Special: A qualitative study into the experiences of women in high leadership positions in the Dutch Armed Forces

THE RISKS OF FEELING SPECIAL

A qualitative study into the experiences of women in high leadership positions in the Dutch Armed Forces

Student: Fia van Heteren Student number: s1843745

Supervisors: J.E.T Schmidt, MSc & Prof. Dr. S. M. Groeneveld Second reader: Dr. N. J. Raaphorst

Date: 15-06-2018 Content: Master Thesis

1 The opinions and conclusions in this thesis are the analysis of the student in Public Management from the faculty of Governance and Global Affairs of Leiden University and do not necessarily represent the vision of the Ministry of Defense or other Dutch Public organizations.

2 Table of Contents

Foreword ... 4

1. Introduction ... 5

1.1 Introduction to the problem and research question ... 5

1.2 Justification ... 6

1.2.1 Scientific relevance ... 6

1.2.2 Societal relevance ... 8

1.3 Reading guide ... 9

2. Literature review ... 10

2.1 From the glass ceiling to the glass cliff metaphor ... 10

2.1.1 The glass ceiling ... 10

2.1.2 The glass cliff ... 11

2.2.1 Tokenism ... 13

2.2.2 Visibility ... 14

2.2.3 Assimilation ... 14

2.2.4 Contrast ... 15

2.2.5 Reactions to Kanter’s (1977) token framework ... 16

3. Research setting ... 19

3.1 Background of the DAF ... 19

3.2 Women in high military positions ... 21

4. Methodology... 23

4.1 Research design ... 23

4.2 Respondents ... 24

4.3 Data collection through semi-structured elite interviews ... 25

4.4 Qualitative data analysis ... 27

4.5 Validity and reliability ... 29

4.6 Ethical considerations ... 29

5. Glass cliff experiences of women in high leadership position in the DAF ... 31

5.1 Visibility and invisibility for women in high leadership positions in the DAF ... 31

5.1.1. Introduction ... 31

5.1.2 Experiences of heightened visibility ... 31

5.1.3 Experiences of lowered visibility ... 38

5.1.4 Conclusion: the experiences of heightened and lowered visibility ... 42

5.2 Assimilation of women in high leadership positions in the DAF ... 42

5.2.1 Trying to be yourself and stereotype threat ... 43 5.2.2 Misinterpretations of and opinions about women in high leadership positions in the DAF 46

3

5.2.3 Conclusion: assimilating in several ways ... 48

5.3. Contrasting experiences of women in high leadership positions in the DAF ... 48

5.3.1 Introduction ... 48

5.3.2 Inappropriate interruptions ... 49

5.3.3 Differentiating language ... 50

5.3.4 Conclusion: differentiation through interruptions and differentiating language ... 53

6. Conclusion ... 54

6.1 Discussion on the experiences of women in high leadership positions in the DAF ... 54

6.2 Limitations of this research project and recommendations for further research ... 56

6.3 Recommendations for practice ... 57

7. Bibliography... 59

Appendix A: Topic list ... 66

Appendix B: Consent form ... 67

Appendix C: Code table after open coding ... 68

4 Foreword

In this thesis project I explored the experiences of women in high leadership positions working in the Dutch Armed Forces. There are not many women working in the higher ranks of the Dutch Armed Forces. The women who work in those positions do have a great awareness share by for instance media and colleagues. This study calls the positions of women in high leadership roles in the DAF “glass cliff” positions, because their minority position, together with their gender status, could lead to experiences of risk and insecurity for their further careers. This thesis explores the experiences of this so-called glass cliff position by using Moss Kanter (1977) her token framework based on the mechanisms of visibility, assimilation and contrast. Moreover, the thesis further specifies the mechanisms of the token framework.

In order to answer the research question, I conducted empirical research through qualitative interviews during the months April and May of 2018. I gained access to these interviews through a five month internship at the main policy department of the Ministry of Defense. Here, I learned a lot about policymaking and about the Dutch Armed Forces. Both the internship and the interviews were very good experiences. Furthermore, the analysis of the interviews and the writing of the thesis were demanding, but very inspiring.

I want take advantage of the opportunity to thank everyone that helped me during this research project. Thanks to both my thesis supervisors, Sandra Groeneveld and Eduard Schmidt for giving me the opportunity to join their capstone project and for their critical eye on the progress of my thesis. Furthermore, the interviewees participating in this research project have been of great help and I very much enjoyed our conversations. In addition, I could not have been able to research this fascinating topic without all the help I got at my internship. I want to especially thank Mart-Jan Alkemade, John Wertenbroek and Corine Nieuwburg-Hofte for guiding me during my internship and during the first phase of my thesis project. Moreover, I want to thank my family for their mental support and my sister Lies van Heteren for her willingness to discuss my thesis progress.

I hope you will enjoy reading my thesis.

Fia van Heteren,

5 1. Introduction

1.1 Introduction to the problem and research question

This research project explores the experiences of women in high leadership positions in the Dutch Armed Forces (DAF), which is the military part of the Dutch Ministry of Defense (Moelker & Bosch, 2008). This research project is part of a Leiden University research program called “Women in senior civil service. Breaking the glass ceiling, but facing a glass cliff?” In this research program, Leiden University researchers explore the role of the glass cliff metaphor related to leadership positions that are difficult and risky in Dutch central government. In general, the share of women working in the Dutch government is increasing in line with the governmental goal of 30 percent. The Ministry of Home Affairs and Royal relations predicts that in 2025, women will outnumber men in the Dutch government (Hulzebosch et al., 2017). At the same time, the number of women in senior civil service leadership positions stays behind, but is slightly increasing (Hulzebosch et al., 2017; Van der Meer, 2010; Groeneveld, 2009). In contrast, the DAF did not reach the governmental goal of 30 percent women in their organizations. For years, the share of women in high leadership positions (Colonels and Generals) has been the same, at a level of five percent (Defensie, 2017). In the Dutch parliament, the small share of women in the DAF is a topic of ongoing discussion. Some politicians stress that a quota would help to make the organization more diverse and inclusive. State secretary Visser from the Ministry of Defense does not agree with this standpoint, because quota also did not work well within other governmental organizations, she explains in a general consultation on personnel on May 23, 2018.

Furthermore, the attention from scientists and society for women in high leadership positions has been growing; making this topic highly relevant. Two paradigms are apparent in the scholarship on women in organizations. First, the dominant paradigm framing the challenges and barriers that women face in climbing the ladder to top leadership positions is termed the “glass ceiling” (Smith & Monaghan, 2013). Second, recently, scholars coined the “glass cliff” metaphor, which depicts the challenges women face when they manage to break through the glass ceiling into high leadership positions (Sabharwal, 2015). Until now, there has not been much attention for the career development of women and what opportunities and blockages they experience (Groeneveld, 2009). Therefore, Sabharwal (2015) stresses organizations should look further than achieving targets about the share of women in high leadership positions. Attention needs to focus on ensuring that women also remain in high leadership positions (Sabharwal, 2015). Literature on the glass cliff stresses that such positions could be precarious because of several reasons. For example, the organization is in crisis, they

6 have to undergo budget cuts, they do not get the right supplies or the women are excluded from informal networks and mentoring (Sabharwal, 2015). Taking into account that women are a very small minority in the DAF and according to research by Moelker and Bosch (2008), a masculine culture is dominant; women in high leadership positions in the DAF possibly face challenges. Therefore, the glass cliff metaphor is applicable to women in military high leadership positions, as these positions could be both risky and precarious (Ryan & Haslam, 2005; 2007). Therefore, this thesis focuses on the few women in glass cliff high leadership positions in the DAF. Chapter two offers a comprehensive description of what the glass cliff exactly entails for women in high leadership positions in the DAF.

Kanter (1977) describes the possible challenges and experiences for women, who are tokens, a small minority in a male-dominated context in her theory on tokenism. The tokenism framework consists of the mechanisms of assimilation, polarization and visibility. These mechanisms could for example mean that tokens face lacking networks, stereotyping, work pressure, and stress, making their positions precarious in themselves (Kanter, 1977). Even though Kanter formulated her framework with these mechanisms years ago, this thesis intends to show by researching the male-dominated glass cliff positions within the DAF that tokenism could still be relevant nowadays. The question that arises is what happens to women in the DAF after they promote into high leadership positions. In the setting with skewed gender proportions, tokenism is very prominent. Therefore, it is a concretization of the glass cliff and the perfect setting to expose the experiences of women in glass cliff high leadership positions. The research question that is at the center of this research project is:

What does it mean and imply to be a woman in a military high leadership position within the Dutch Armed Forces?

The answer to this research question will follow from the analysis of interview data from women working in high military positions in the DAF. The purpose of this study is to explore the experiences of women in high leadership positions in the DAF and to complement the current research on the glass cliff (Ryan & Haslam, 2007) and tokenism (Kanter, 1977).

1.2 Justification

1.2.1 Scientific relevance

This section will explain why more research is necessary on the topic of women in high leadership positions in the public sector and how this specific research project intends to fill this scientific gap.

7 First, so far, most research on women in organizations has examined what leads to their appointment to high leadership positions, while there is not much evidence of what happens to them after they secured such positions (Ryan et al., 2016). Thus, academics do not know much about how women in senior leadership positions experience the possible risks they encounter in glass cliff positions (Sabharwal, 2013). In addition, while researchers explored glass cliff experiences in American, British and Australian private organizations (Ryan & Haslam; 2007; Catalyst, 2014; Brady et al, 2011), research on women in high leadership positions in public organizations is scarce (Smith, 2014; Bowling et al., 2006). More research into women in high leadership positions in public sector organizations is necessary, as the organizational environments of the private and public sector organizations are not the same (Meier et al., 2006). Thus, glass cliff positions in the public sector could mean and imply something else than in private sector organizations.

Second, the current scholarship on women in the public sector is restricted to evidence that is not suitable for empirical examination of the underlying mechanisms of women that face a glass cliff (Wilson-Kovacs et al., 2006; Groeneveld & Schmidt, 2017; Smith & Monaghan, 2013; Groeneveld, 2009; Bowling et al., 2006; Portegrijs et al., 2006; Hulzebosch et al., 2017). Therefore, this research tries to get a better understanding of how women in high leadership positions in the DAF experience the glass cliff by using the tokenism framework. This research project is the first to take this qualitative and explorative approach to studying the glass cliff with the use of the tokenism mechanisms. Chapter two on theory and chapter four on methodology will offer a further explanation of the token framework and its use in this thesis.

Third, what makes this research project even more relevant is the setting of the DAF. On the one hand was the DAF not yet part of any glass cliff research. On the other hand, women in high leadership positions in the DAF are very clearly in a glass cliff position, because the women are a very small minority in a masculine organization. Other scholars conducted research on the glass cliff in contexts in which women are breaking through the glass ceiling in remarkable numbers (Ryan et al, 2016). Women are not breaking through the glass ceiling in great numbers in the DAF, making the setting of the DAF even more precarious and thereby relevant for an exploration of the glass cliff position. Chapter three will further outline the suitability of the research setting.

In short, the combination of these gaps in the public administration and sociology literature shows the necessity of exploratory research into the glass cliff. This study intends to fill these gaps by offering an in-depth exploration of what the glass cliff could mean and imply

8 and thereby providing a stepping-stone for future research into the glass cliff and tokenism in male-dominated settings of the public sector.

1.2.2 Societal relevance

This section discusses the social relevance of this research project. First, gender representation in high management positions in government is problematic and is therefore a hot topic in both academic research and in the public debate. Public organizations should behave like role models and to set standards for other organizations. Therefore, public organizations do need long-term strategic HRM perspectives (Becker et al., 1997 & Becker & Huselid, 2006), as they are competing on the same labor market as private organizations and their specialized managers have to know how to attract the right people. In the following years, the Dutch public service will face increasing shortage of personnel that is highly educated. At the same time, many employees will retire and flow out. Therefore, recruitment and retention of personnel is important in Human Resource Management policies. Already years ago, these challenges for the public sector led to diversity policies and strategies to increase the representation of women and ethnic minorities in the public sector (Groeneveld, 2009).

The context of the DAF is especially relevant, as the lack of diversity and shortage of personnel in the future will be even greater for the DAF. Civilian and military personnel together, the Dutch Ministry of Defense expects to have around 5600 vacancies this year (Defensie, 2017). At the same time, military personnel is very attractive for other organizations, because they know how to work hard, they have leadership experience and they know how to take responsibility (Oud-militair gewild als manager, 2010). The DAF tries to find out how to become a more attractive organization; otherwise, their personnel problem will become even bigger. Concerning this intention and urgency, in 2017, the Ministry of Defense asked the Social planning bureau (SCP) to make a report about the way staff members experience and appreciate diversity in the DAF. This report (Andriessen et al., 2017), showed that more diversity within the DAF supports the operation and that it has a positive influence on the culture of the organization. However, the SCP concluded that staff members are now working in a dominant, masculine and closed structure. Seen from a military point of view, this organizational structure is functional and it should contribute to the embedding of people in the organization. At the same time, this culture contributes to exclusion of people that deviate from the masculine norm (Andriessen et al., 2017). According to Moelker and Bosch (2008), the masculine culture is one of the persistent barriers women in the armed forces face. In addition, the SCP states that women and other minorities in the DAF experience more undesirable behavior than male, white and heterosexual personnel do (Andriessen et al., 2017).

9 After publication of the SCP report (Andriessen et al., 2017), the Dutch parliament pushed the Ministry of Defense to focus on the topic of gender representation even more (Kamerstukken II, 34550, nr. 65, 2017). The parliament reminded the Minister of Defense several times of the suggestion that the Ministry should act on the recommendations in the SCP report. There are several projects on diversity and inclusion in the Ministry of Defense. However, a clear plan to show the parliament is still missing. The file on diversity and inclusion of personnel now belongs to the State Secretary of Defense, Barbara Visser. Recently, on May 23, 2018, State Secretary Visser answered questions about personnel in a general consultation in the Dutch parliament. State Secretary Visser made clear that she is against a quota. However, during this general consultation, she promised to explain more about the Ministry’s plans for diversity and inclusion in the next personnel report in 2018.

These above-described events do point to the urgency and relevance of studying women in high leadership positions in the DAF. The SCP report focused on personnel in all layers of the organization and used quantitative methods, while this thesis project offers an in-depth analysis of women in high leadership positions. This thesis captures the experiences of women in high leadership positions in the DAF and engages in discussions surrounding gender, diversity and leadership in the public debate (Ryan et al., 2016). At the same time, this study into the working of mechanism of in- and exclusion regarding women in high leadership positions in the DAF is relevant for policy recommendations (Moelker & Bosch, 2008) in the DAF.

1.3 Reading guide

This research project consists of seven chapters. After the introduction, the second chapter starts with explaining the theoretical framework and sensitizing concepts that this project makes use of. The third chapter offers an outline of the research setting. The fourth chapter will discuss the methodological approach that this thesis project makes uses. The fifth chapter goes into the analysis of interview data by discussing the mechanisms that the coding process of the interview identified. The sixth chapter offers a conclusion, including a discussion, limitations of the research and recommendations for further research. The bibliography in chapter seven gives an overview of the literature this study used.

10 2. Literature review

This chapter outlines the theoretical foundations of the study, by discussing scientific literature that are useful for studying the glass cliff positions in the DAF. As explained in the introduction of this thesis, the share of women in the higher ranks of the DAF is extremely small (Hulzebosch et al., 2017). Furthermore, DAF personnel do experience some situations of insecurity and risk: until recently there have been great budget cuts and there is great personnel shortage. To enlarge the understanding of the experiences of the few women that are working in high leadership positions in the DAF, this research project uses the token mechanisms as sensitizing concepts for studying the glass cliff position. Sensitizing concepts are the lens through which the analysis in this research project took place, which section 4.4 will further discuss. This literature review includes two sections. The first section assesses the origins of the current state of the glass cliff metaphor. The second section goes into the token mechanisms of the glass cliff, and it discusses the possible token responses found by earlier scholars.

2.1 From the glass ceiling to the glass cliff metaphor 2.1.1 The glass ceiling

Recently, several researchers focused on gender equity related to advancement into high leadership positions. Among others, scholars in public administration focused on barriers women face in their attempts to climb the career ladder. Ryan (et al., 2017) and O’Reilly (et al., 2015) demonstrated that women would more likely experience disadvantage moving to high leadership positions by receiving lower pay. Other scholars thought of multiple metaphors to illustrate obstacles to women’s advancement. For instance, the glass ceiling, the glass labyrinth and the sticky floor (Smith & Monaghan, 2013). The concept of the glass ceiling was dominant, which was coined in 1986 (Lockwood, 2004) and gathered support from evidence found by several researchers (Kanter, 1977; Morisson et al, 1987). The glass ceiling represents barriers for women who attempt to move to high leadership positions. The think-manager-think-male framework (Schein, 1975) theoretically explains these invisible barriers in a framework with stereotyping bias of men and women. It associates women with being communal and social (Meier et al., 2006). While it expects that men are more risk-taking, ambitious, confident and aggressive, the characteristics associated with a leader. Such gender role stereotyping can result in the idea that women are not as qualified as men are for leadership positions (Brenner, Tomkeiwicz & Schein, 1989), which subsequently could make it harder for women to get into those kind of positions. Other researchers found that women face a dilemma called role incongruity. This holds that apart from their role as a leader, society expects women to have a

11 social role (Eagly & Carli, 2007). Even though it seems hard for women to get into leadership positions, nowadays more and more women do (Dreher, 2003). The next section goes into what happens when women do get into leadership positions.

2.1.2 The glass cliff

Some empirical studies showed that when women reach high leadership positions, they often still encounter gender discrimination limiting them to become successful leaders (Smith & Monaghan, 2013; Sabharwal, 2015; Ryan & Haslam, 2005; 2007). The recently coined metaphor of the glass cliff intends to explain what happens when women advance into senior positions (Wilson-Kovacs et al., 2006). This metaphor argues that women who break through the glass ceiling are more likely to promote to positions that face a high risk of failure (Sabharwal, 2015). Thus, Sabharwal (2015) views the glass cliff like another kind of glass ceiling for women who get into high leadership positions (Sabharwal, 2015).

Moreover, the glass cliff holds that when one does break through the glass ceiling, the experiences of male and female leaders differ. Compared to men, women are still less likely to have positions with authority to accomplish strategic goals, to have opportunity for promotion, to get a reward and proper payment for their roles, to be part of networks and support systems and feel less empowered (Lyness & Thompson, 1997, 2000; Ryan et al., 2016; Glass & Cook, 2016; Sabharwal, 2015). In response to an article in the British newspaper The Times (Judge, 2003), which argued that women in boardrooms had weakened companies’ performance, Ryan and colleagues (2016) researched and found the opposite argumentation: poor company performance might trigger women’s appointment to leadership positions. According to them, the glass cliff metaphor holds the increased risk of negative consequences in times of crisis, regarding women’s leadership positions and the instability of their roles.

Ryan and Haslam (2017) argue that the following factors determine when glass cliffs are most likely to materialize; the nature of the organizational crisis and need for change, the availability of financial and social resources and the individual differences between decision makers leading to selection bias, gender, and leadership stereotypes. A crisis can include all sorts of incidents and accidents, for example scandals, changes in organizational dynamics, failing financial performance and cutbacks (Ryan & Haslam, 2005). In the DAF, the setting of this thesis, the glass cliff clearly materializes in the individual differences between the decisions makers leading to selection bias. The majority including the decision makers in the DAF are male, while the women on the glass cliff are female. The major source of the glass cliff is that women are a minority in the DAF. Chapter three on research setting discusses the contextual characteristics that make the context even more challenging and thereby sensitive for riskiness

12 and precariousness of high leadership positions for women (Ryan & Haslam, 2007) working in the DAF. Thus, a leadership position is in itself a risky position, which could be more challenging because of the context (Ryan & Haslam, 2017; Ryan et al., 2016). In times of crisis, it is likely that blame for leadership failure is highly visible and public. In addition, research demonstrated that non-prototypical leaders tend to get unproportioned punishment for their mistakes and to have their performance devalued (Ryan et al., 2016).

Furthermore, a new process in the glass cliff metaphor, which was not apparent in the glass ceiling metaphor, is the organizational need for change. Women are associated with change, while men are associated with stability. Evidence suggests that a glass cliff occurs when organizations strategically move away from male or standard leadership and prefer a female leader in crises. Researchers refer to this phenomenon as think-crisis-think-female (Ryan et al., 2016), in contrast to earlier opted think-manager-think-male phenomenon (Schein, 1975). Moreover, women would be good at making change happen, with their social skills. Eagly & Carli (2003) argue that women may have an advantage in getting into leadership positions in times of poor company performance. However, women have to take the blame in case of failure, which increases risk and insecurity for them. Not surprisingly, researchers found that taking on such positions could reflect badly on their further careers (Eagly & Carly, 2003; Ryan & Haslam, 2005, 2007; Smith & Monaghan, 2013). In addition, women might be more willing to accept leadership positions in times of crisis, because they have few opportunities (Ryan et al., 2016). For instance, Smith (2014) explored the circumstances under which women got into leadership positions in public organizations. She argues that women were more likely to get into leadership positions in times when risk of failure was higher.

So far, scholars offered limited empirical support for the glass cliff phenomenon. However, the glass cliff is multiply determined as well as context dependent. Therefore, not in every research context, there will be much evidence of the glass cliff (Stichman et al., 2010). In this thesis, the glass cliff is a metaphor for women in high leadership positions that are set up for failure. In the context of the DAF, the extreme underrepresentation of women, the token situation, is as a glass cliff situation in itself. Therefore, this thesis applies the tokenism framework on the glass cliff situation. The next section goes into possible mechanisms of the glass cliff position described by Kanter (1977).

2.2 The token mechanisms

As described above, the situation of extreme underrepresentation in the masculine organization of the DAF is an example of being on a glass cliff. Therefore, in this thesis, the metaphor on

13 the glass cliff will be based on the mechanisms that are part of the token theory opted by Moss Kanter (1977), to understand how women experience being in a glass cliff position. This section first goes into the conceptualization of tokenism, after which it discusses the mechanisms of the glass cliff and tokenism: what they could mean in different contexts and what evidence earlier scholars found. Finally, it discusses what those concepts mean for this thesis project.

2.2.1 Tokenism

Kanter (1977) conceptualizes groups that consist of a majority of eighty-five percent and a minority of fifteen percent as a skewed group. Kanter (1977) calls people that are part of this fifteen percent minority the tokens and people that are part of the majority the dominants. This conceptualization of tokens fits well with women in high leadership positions in the DAF, which is a group of only five percent (Defensie, 2017). Tokens are in an exceptional position, because they are a very small minority. Tokens are not just different from the group along one dimension, but they reflect their status as a symbol. They carry characteristics with them as a set of assumptions about culture, status and behavior, striking for the majority of the group. This means that tokens represent their ascribed category (Kanter, 1977). Tokens do stand out from the group because they are a minority. At the same time, the dominant members treat the tokens as if they represent their sort, in this case women. This means that the majority will always see the rare tokens as different from them (Andolsun, 2004). The dominants do control the culture of the group. Therefore, “relative numbers interfered with a fair test of what men or women can “naturally” do” (Kanter, 1976b in Kanter, 1977, p. 970). Furthermore, feedback between structure and behavior can produce upward cycles of advantage, or downward cycles of disadvantage. It is hard for a token to break out of such upward or downward cycles, because of its self-perpetuating nature. According to Kanter, this cycle can only be broken from the outside (1977a, p. 249). In other words, people in a majority have high opportunity and behave in ways that create more opportunity and power. Those cycles of opportunity, power and numbers increases difficulty for tokens to enter certain contexts (Guy, 1993).

The token framework consists of three mechanisms. This research project uses the mechanisms of visibility, polarization or contrast, and assimilation introduced by Kanter (1977) as sensitizing concepts. These mechanisms are each associated with interaction dynamics that generate typical token responses and experiences. Several scholars, using numerous occupational groups (a.o. Stichman et al., 2010; Cassell & Walsh, 1997), did test Kanter’s tokenism theory. Scholars uncovered both support and criticism of Kanter’s findings, which the next sections will discuss, together with Kanter’s findings.

14 2.2.2 Visibility

In short, Kanter (1977) explains visibility as a mechanism in which tokens capture a disproportionate attention share, which could result in performance pressures and more stress. Kanter (1977) argues that two conditions heighten the visibility and effects of tokenism. First, the token is physically obvious. Second, the token’s social type is rare as well as new to the setting of dominants. In the research setting of the DAF, the women in high positions are rare and physically obvious because of their gender. As a result, visibility generates performance pressures through public performance, attention to a token’s discrepant characteristics and the tokens’ fear of retaliation (Kanter, 1977). According to Ryan (et al., 2016), also high scrutiny from stakeholders and media add to the pressure on performance. Paradoxically, the tokens’ achievements are noticed less (Kanter, 1977a, p. 248). While when tokens perform badly, people believe that all people of their social type must perform that way. Dominants start seeing the tokens as representative of their complete social group, thereby affecting not only the one person performing badly. This results in women acting as ‘women’, instead of themselves. This phenomenon could result in token women feeling more pressure to conform, to prevent making mistakes, to try to become socially invisible and not to stand out (Kanter, 1977). Several other scholars (a.o. Ryan et al., 2016; Boin et al., 2010) did research into the token mechanisms in glass cliff situations. In these cases, blame for leadership failure is both visible and public in times of crisis, because research and attention from media and stakeholders may be more (Ryan et al., 2016; Boin et al., 2010). The latter does sometimes result in more punishment for mistakes for non-standard leaders (Brescoll, Dawson, & Uhlmann, 2010 ).

An example of a typical token response to heightened visibility is overachievement, which can lead to resentment of the token. In addition, some tokens do respond through underachievement, so dominants do not see them, which can have negative consequences for job performance and promotions. The before mentioned responses were found by Kanter (1977), as well as Cassell and Walsh (1997). The latter call this a way for token women to handle or cope with the fact that they are treated differently in male-dominated jobs. To these strategies, Cassell and Welsh (1997) add flirtation, working behind the scenes and nurturing colleagues. Another possibility is for a token to limit their visibility by avoiding public visibility or by trying to fit into the dominants group (Kanter, 1977). Others try to benefit from their visibility and use it in their advantage.

2.2.3 Assimilation

Assimilation often happens when in diverse environments, individuals differ with respect to status. Assimilation is a mechanism in which tokens distort their attributes to fit pre-existing

15 generalizations or stereotypical roles about their social type, which can result in role encapsulation in sex appropriate roles in the organization (Kanter, 1977). This role entrapment can lead to missed opportunities of promotion. Advancement does not fit with the way the dominants see the tokens. Furthermore, misinterpretations of a tokens’ identity and role in the organization can limit their effectiveness (Kanter, 1977; Stichman et al., 2010). Kanter argues that due to the tokens minority status, other people make assumptions about their roles. Kanter calls this statistical discrimination, when “an unusual woman may be treated as though she resembles women on the average” (Kanter, 1977, p. 980). Kanter identified the roles of mother, seductress, pet and iron maiden. Tokens often respond to role entrapment, by playing into the roles, rather than proving they are their own individual (Kanter, 1977). Yoder (1991, 1994) adds that women do experience assimilation in gender-inappropriate occupations, because they have a subordinated gender status.

According to Kanter (1977), a possible reaction of a token to assimilation is to develop a preference for already-established relationships compared to new ones. A qualitative study by Archbold and Schulz (2008) found that female police officers reacted through the mechanism of assimilation by refusing to participate in promotional processes, because they thought they would get a promotion just because they are women. In this and several other studies, women perceived themselves as singled out in organizational life because of their sex (Hassell & Brandl, 2009; Martin, 1979; Morash & Haarr, 1995). According to Shore and colleagues (2011), tokens assimilate to the dominant culture and downplay their uniqueness, so others could see them as insiders. Ely (1995) also found this; female attorneys adopt behaviors that are masculine so others view them as a good attorney. In Hewlin (2009), tokens suppress their personal values and pretend to embrace organizational values. Cohen & Swim (1995) concluded that tokens expect majority members to evaluate them by the use of stereotypes, which could in turn cause increased stress for tokens.

2.2.4 Contrast

A third and last token mechanism of a glass cliff situation is that the differences between the tokens and the dominant group are exaggerated and polarized. Contrast takes place as result of high visibility and pressures to assimilate by the dominants (Kanter, 1977). Boundary heightening of the dominants could result in social isolation of the tokens. Tokens are too few in numbers to prevent the application of familiar generalizations or stereotypes by dominants. “It is thus easier for the commonalities of dominants to be defined in contrast to the token than it would be in a more numerically equal situation.” (Kanter, 1977, p. 971). In addition, the presence of a token makes dominants more aware of what they have in common, while the

16 tokens form a threat for their commonality. Therefore, dominants tend to exaggerate culture of which before they might not even have been aware, thereby keeping tokens outside their boundaries (Kanter, 1977). Moreover, Kanter (1977) implies that contrast could show in dominants apologizing beforehand for an action that the tokens might not agree with. When the token objects, she would be interrupting the dominant. Kanter calls this kind of events loyalty tests in which dominants look for a way to pressure tokens to turn against their subgroup or social category’s members. Especially relevant for this thesis is when Kanter (1977) shortly writes about high-status tokens. They could enter formal discussions, but they are still not part of informal networks. Thus, if the token’s status is higher than that of the dominants, the interaction might change, while the contrasting dynamics remain the same.

A possible token response is for instance to act like an exceptional member of their subgroup in the hope of becoming an insider (Yoder & Sinett, 1985). Others try to accept the isolated position they already have. The latter option could mean that token women are likely to experience exclusion from informal peer networks, which could result in limitations in this source of power-through-alliances. They possibly have fewer opportunities to be ‘sponsored’, because there are not many people like them in high positions as well (Kanter, 1977a, p. 249). More researchers studied tokenism in male-dominated fields. For example, Stichman et al. (2010) found that female police officers are more likely than male officers to believe that they were underestimated and to believe there was gender bias in their organization. In addition, among others, Archbold & Schuldz (2008) their study supported Kanter (1977) by arguing that female officers felt isolated from male officers based on their sex. Furthermore, Gustafson (2008) found that sex significantly relates to the amount of gender-related jokes in someone’s presence. The next section further evaluates scholars’ use of and reactions to Kanter’s (1977) mechanisms.

2.2.5 Reactions to Kanter’s (1977) token framework

As has become clear in this section so far, Kanter's (1977) tokenism theory did get a lot of attention from social scientists. Many scholars supported Kanter’s (1977) claim that female tokens experience negative consequences like performance pressure, role encapsulation and heightened stress. Moreover, many used the framework for research on male-dominated jobs, for example police officers (Archbold & Schuldz, 2008). Almost all studies on this topic found evidence that female tokens have had negative experiences in the work place due to them being a token in a masculine organization (Archbold & Schulz, 2008; Gustafson, 2008; Krimmel & Gormley, 2003; Martin, 1979; Wertsch, 1998).

17 At the same time, there is some criticism on the tokenism framework offered by Kanter (1977). Some scholars believe that Kanter’s theory is too simplistic, because of its strong focus on proportions. The numerical representation of tokens could reduce any struggles that might be associated with individual characteristics of group members. For example, different treatment related to social inferiority due to sex, race or age could be relevant for struggles of tokens as well, instead of just skewed numbers (Stichman et al., 2010). Evidence from research in female-dominated settings supported this proposition. Williams (1992) first argued that male tokens did not experience the same negative effects as female tokens in male-dominated settings. Male tokens advanced quickly with the assumption that they were better suited for leadership positions and experience what Williams (1992) introduced as a “glass escalator”, compared to the glass ceiling and the glass cliff women could experience in male dominated settings. Men resisted women entering and elevating in male occupations, while women did not do so for female professions like nursing, teaching and social work (Ott, 1989; Williams, 2013). Twenty years later, this argument appears outdated, because it fails to recognize race, sexuality and class; and Williams based her argument on traditional organizations (Williams, 2013). Still, Williams’ (1992 & 2013) and other scholars’ (Yoder, 1991, 1994; Ott, 1989) arguments give body to the useful idea that tokenism is not gender-neutral as the numerical representation of tokens alone was not enough to produce tokenism. Yoder (1994) examines the proportions of tokens in occupations as a whole, in addition to tokens in a specific work group or organization to find what she calls the factor of “occupational inappropriateness”, which relates to gender stereotyping of the occupation. She found that disadvantaged gender status contributed to tokenism, with no regard of the gender appropriateness or image of the job (Yoder, 1994 & Ott, 1989). More research projects add to the argument of the importance of social status in the experiences of tokenism. For example, a study into the relationships between sex proportions and performance evaluation among high-ranking officers in the Israeli Defense Forces (Pazy & Oron, 2001). The results support Kanter’s framework on tokenism with the finding that when women were token members, their bosses rated them lower on performance. Their bosses rated the women higher on performance when they constituted a higher proportion. Men’s overall performance did not vary with proportion. Thus, this study is a clear example in which gender status and numbers do interact.

The studies discussed here show that the token framework (Kanter, 1977), its three mechanisms together with the acknowledgement of the impact of societal gender-based discrimination (Yoder, 1994 & Ott, 1989) will be of use for this project on the glass cliff position of women in high military positions in the DAF. However, there is no record of recent

18 qualitative studies using the token theory in a context of women in high leadership positions in the DAF. Therefore, it is likely that this exploratory study will find further specifications of the three token mechanisms in this specific context. Section 4.4 on qualitative data analysis will outline how exactly this thesis makes use of the above-discussed sensitizing concepts of the glass cliff position.

19 3. Research setting

The setting in which can be learned most about women in a glass cliff positions is the setting of women in high military positions in the Dutch Armed Forces. The first section of this chapter goes into the background of the setting that could add to the precariousness of the glass cliff position for women in high leadership positions in the DAF. This background includes the historical role of women in the DAF and recent political and organizational developments related to the DAF. The second section discusses what the concept of women in high leadership positions in this study implies.

3.1 Background of the DAF

The DAF consists of four service branches. The Royal Netherlands army, or the land forces, the Royal Netherlands Navy, the Royal Netherlands Air Force and the Royal Military Constabulary (Krijgsmachtdelen: Samen werken aan vrede en veiligheid, n.d.). Women have always been underrepresented, in all four branches (Moelker & Bosch, 2008, p.7). During the 17th and 18th century, the DAF consisted of just men. Two centuries later, women became acceptable supporters of the armed forces. However, women were restricted to conditions of employment with limited education and career potential and they got dismissed in case or marriage of pregnancy. Still today, the DAF are a masculine institution in authority structure and formal values, in majority of personnel and organizational cultures (Bosch & Verweij, 2002).

Recently, a lot has changed in the DAF. The most traditional task of the DAF is protecting the Dutch country, sea and the Dutch economic interests. After the fall of the Berlin Wall, the acute threat from the East disappeared. The Dutch Defense organization increasingly contributed to promoting the international legal order and peacekeeping missions became important. Then, the Netherlands stopped obligatory attendance to military service and the Defense organization transformed into a professional organization. This transformation influenced the composition of the organization (Moelker et al, 2005; 2017). From 1989 to 1999, the workforce decreased from 130.000 to 75.000. When the budget for Defense in the Netherlands decreased, the composition of the DAF came under pressure and shrunk even more, unto 56.600 staff members in 2016 (Defensie, 2017). In addition to huge personnel shortages in the DAF (Personeelstekort bij Defensie: Wie wil dit in godsnaam nog?, 2018), the overall realized influx of new military personnel was smaller than planned (Defensie, 2017, p. 5). In light of those problems, in 2016, Minister Hennis- Plasschaert from the Ministry of Defense stated that the DAF was not able to defend the Netherlands anymore (Defensiebond, 2016). The HR-policy of the DAF for the next few years needs a different focus, because of the context of

20 appointments in the coalition agreement, increasing tightness on the labor market, aging of the labor market, increasing demands for educated staff, DAF officials in need of career development possibilities, importance of a diverse staff and the question for new public leadership. Several measures from the last few years have not yet solved the personnel issues for the Ministry of Defense, which will have many vacancies in 2018. The Ministry of Defense has to make extra efforts to become a more attractive employer for a diversity of people (Defensienota, 2018). One way the Ministry of Defense intends to react to changing demands, is by using adaptive armed forces. This should increase the ability of the DAF to be more flexible but still robust in changing circumstances. The core value of adaptive armed forces is that the DAF does not always need to have to own material or personnel, but it is much more important to be able to have access to it when they need to (Defensienota, 2018).

Furthermore, since 1991, downsizing and reorganization combined with activating personnel policies promote the presence of women in the organization (Moelker & Bosch, 2008). The DAF works for inclusive leadership and an inclusive work culture in order to increase and make use of the diversity of staff. One reason is that during missions, diversity has proved to be very important. For instance, women are said to be better at communicating with local female civilians (Diversiteitsbeleid Rijk, 2016). With the exception of the submarine service, all departments, positions and training facilities are open to women nowadays and they formally have equal opportunities as men (Bosch & Verweij, 2002). Furthermore, on February six of 2018, the parliament voted about the proposal to change the framework law on obligatory military service. The bill arranges that women that become 17 now also receive a letter for obligatory military service (Kamerstukken II, 34764, nr. 36, 2017). What this means in practice, is that all provisions in the law will be formulated in a gender neutral way. This change symbolizes equality between men and women. However, as PhD student Bootsma (2018) states, this focus on equals tasks is not the same as equal access to power and influence in the political and military top. In addition, the Ministry of Defense did focus on intensifying the recruitment and tried to find ways to keep women in the organization, in order to increase the number of women in the organization and in the higher positions specifically (Diversiteitsbeleid Rijk, 2016). Despite many efforts, an official change in policy and an international shift in perceptions of women’s human rights, discriminatory gender representations persists in the DAF and it appears difficult to keep women on long-term basis (Bosch & Verweij, 2002). Masculine culture, norms and values that seem resistant to change, form barriers to higher participation of women in the Armed Forces (Moelker & Bosch, 2008).

21 Even though this study focuses on the underrepresentation of women in high leadership positions in the DAF, the above-described developments: constant underrepresentation of women in all ranks, budget cuts, not enough personnel, and a persistent masculine culture could all make the positions of women in high leadership positions in the DAF even more precarious. The next section will move to a more detailed description of what the concept of women in high leadership positions means in this study.

3.2 Women in high military positions

This thesis explores the experiences of women in high military positions in the DAF. Therefore, it is important to explain what high military positions are in this thesis, without entering in too much detail. This thesis views the ranks of Colonel and General as high leadership ranks and the associated jobs as high leadership positions. The interviewed women all have or had the rank of Colonel or General. The ranks are similar to salary scale 15 and higher from the Dutch government organizations. To indicate, this holds a salary from 5438.16 euro to 10667.77 euro per month (Defensie, 2018; Defensie, 2017). To become and to be a Colonel and a General, one usually needs to follow certain routes and have specific capacities to be able to function in this role. There are three routes possible, which were taken by some of the interviewees. First, the standard route is to follow the complete training path that the DAF offers. Many people that followed this path have worked in the DAF since they were around or under 20 years old. Second, some people did a specialized academic education, often doctors or psychologists. For specialists it is possible to follow a shorter path after they graduate and they often flow in in the rank of Lieutenant Colonel or Colonel. A third option is to flow in horizontally at the rank of Lieutenant Colonel, Colonel or General, because one has knowledge and capacities from another work field.

The possible routes bring people with different backgrounds into the organization. However, the interviewees argue that paradoxically, their bosses review them based on capacities obtained through the first and standard route. The team for management development (MD) of the Ministry of Defense summarized the capacities that a Colonel should have as follows: to have empathy, to inspire, to delegate, to network, to negotiate and to develop employees. Furthermore, the capacities one needs to have when growing from Colonel to General are to communicate, to dare, to be people oriented, to be environment-oriented, to have a vision and to be result-oriented (Management Development competenties, 2018). In addition, for anyone working in a military position in the DAF, the work post is the Netherlands, which means that it is possible that you have to move every few years. These conditions are often even

22 stronger for people in higher positions. From one day to another, the organization can need and place you elsewhere. A description of the tasks that are inherent for a high leader in the Dutch military is difficult, as no position in any rank is the same. For example, one could be a commander of an operational organization like an airbase. This job would bring responsibilities such as being the representative of the airbase towards governments and stakeholders, business planning and participating in and leading national and international operations. Moreover, one could be a policy officer or a director of an internal academic institute, which understandably brings very different responsibilities (Management Development taakbeschrijving, 2018). The women participating in this thesis project want to stay anonymous and their tasks are very diverse. Therefore, this section will not describe anything else related to job descriptions of women in high leadership positions in the DAF.

As described above, women have had a minority position within the DAF for years. Recruitment of women for military positions stagnated from the '90s onwards and does, despite several policy intentions and measures, not get higher than nine percent within the armed forces. The percentage of women is smallest in the high leadership positions of military personnel, about five percent. The percentage of women participating in career trainings fluctuates (Moelker et al, 2017). Within the training called the Middle Defense Academy (MDV, Middelbare Defensie Vorming), the percentage of women fluctuates around nine percent, similar to the occupation of female soldiers. Within the Higher Defense Academy (HDV, Hogere Defensie Vorming), the percentage of women is lower, around five percent. This percentage corresponds with the share of female soldiers in higher ranks combined with older age categories (Defensie, 2017). Military personnel sees attending the HDV as a step that a military needs to take to possibly promote into the rank of Colonel. However, it is not a guarantee that one can become a Colonel. Moreover, some Colonels did not do the HDV. The secretary general and the MD team select people who they argue to be fit for becoming a Colonel or a General. Military personnel cannot send in their job application in these high ranks, members of the MD team approach you and do decide for Colonels and Generals which job they will get next.

Thus, this research project will study the case of female military personnel in high ranks, from Colonel onwards, and their position and career experiences. The next chapter on methodology will outline how this research project took place.

23 4. Methodology

This chapter will outline and give arguments for how the research process took place by explaining the research design, the respondents, the data collection, the qualitative data analysis, the validity, reliability and ethical considerations.

4.1 Research design

This research project used a single case study design. This design was appropriate because the study intention was to explore the single case of women in high military leadership positions in the masculine organization of the DAF intensively and in-depth. The special context of the DAF described in chapter three made this case very suitable for an in-depth explorative analysis. The benefit of this special context was that it is an ideal context to capture the experiences of women in leadership positions, as they deviate from the norm of the organization. In addition, the explorative nature of the study fitted well, because this was the first research project studying the topic of women in high leadership positions in the DAF.

In this thesis, the units of observation were women in high leadership positions in the DAF. The unit of analysis was the group of women in high leadership positions. The case was the Ministry of Defense. Furthermore, selection of respondents from the research population took place through purposeful sampling, as is common in qualitative research (Boeije, 2010). The sample was representative of the population, as almost all women in the small population cooperated in the study. In the end, two women did not cooperate and I could not get into contact with two other women, leaving interviews with eleven women. The high level of cooperation of eleven out of fifteen does imply that it is hard to ensure anonymity of the participating women. In order to pursue anonymity, this thesis did not go into details about the respondents’ statements, their personal details and backgrounds, which section 4.2 will further discuss. If researchers would want to replicate this study, they could analyze the anonymized transcripts of the interviews again. Moreover, researchers could use a sample of eleven women in high military positions like those described in section 4.2 and one could use the topic list shown in appendix A.

The glass cliff metaphor and the token theory informed this study a priori and they formed a basis for the research question into the experiences of women in high leadership positions. As women in high leadership positions are under-studied in the Dutch public sector, this project was a great opportunity to explore the mechanisms underlying that kind of positions. The case of women in high military leadership positions in the DAF could teach us a lot about the issues surrounding women in high leadership positions in male-dominated public

24 organizations. It was thus a great opportunity to offer a specification of the token mechanisms. The discussion section of this thesis will further discuss this project’s possibilities for analytic generalizability.

4.2 Respondents

As shortly discussed in the section above, the participating research group consisted of eleven women. The non-response was two women. I could not get hold of the contact details of two of the women in high leadership positions in the DAF. Thus, at the time the research project took place, the population of women in high leadership positions in the DAF, consisted of fifteen women. According to data from 2017, there were 300 Dutch Colonels and Generals. Of those 300, five percent was female (Defensie, 2017). Formally, in the DAF only Generals are top officials. This rank is respectively salary scale sixteen. DAF personnel sees the rank of Colonel, salary scale fifteen, as the breeding pond for Generals. Like in section 3.2, there are differences between the ranks of Colonel and General, but they both embody an important strategic, visible position, with a lot of responsibility and authority. Therefore, it served the purpose of this project to view both Colonels and Generals as top officials or high military leaders. Furthermore, the number of Dutch female Generals I interviewed was only two, which is too few to base the research results on. The latter fact also says something about the high urgency and importance of this study.

Furthermore, an internship position in the Ministry of Defense, granted me access to the contact details of women in high leadership positions from all four armed forces in the Netherlands. Most of the times, my colleague and supervisor at the ministry sent the possible participants an email beforehand to introduce me formally, after which I contacted all participants using my email address at the Ministry of Defense. This email consisted of an invitation with an explanation of the interviews and a consent form. Appendix B contains the consent form. Ten of the interviewed women worked in the DAF at the time of the interviews and one woman was already on age dismissal. The aim of this study was to represent a wide range of perspectives and experiences; it included women working in the four different armed forces in the Netherlands into the sample. The sample contains only one woman working in the land forces. One other woman that was in a high leadership position in the land forces did not respond to my request. The other two women that did not respond to my request respectively work with the Military Police and the Air Forces. The biggest share of six interviewees, works with the Air Forces. From the interviewed women, two work with the Military Police and two work with the Navy. All interviewed women are between the ages of 42 and 52 and they flowed

25 into their current position in differing ways. Most women did go through the ranks in the standard way. Some participating women started their career in another field, are specialists in their field or they promoted faster than average into the higher ranks. Their positions vary very much. Unfortunately, I could go into the positions and tasks of the women in high leadership positions in detail, because of confidentiality. Although, when this supports the argument, I did make careful use of descriptions of responsibilities, careers, tasks and capacities of the interviewees. However, my main concern has always been to guarantee anonymity. In the text, this thesis made use of respondent numbers, to indicate that statements or descriptions belong to certain interviewees. This thesis did not include an overview of the respondent numbers with their ranks or the armed forces they work in, because unfortunately even these data could obstruct anonymity for the participants.

4.3 Data collection through semi-structured elite interviews

The goal of this thesis was to see a part of the social world from the participating women’s perspectives. Semi-structured interviewing suited well with the purpose of the interviews. As section 4.1 discussed, this topic is of sensitive nature and the sample is small. Therefore, the perfect way to grasp the experiences of women in high military leadership positions was through conversations and examples or stories that came up in the interviews. This section outlines how the data collection went.

The main method of data collection for this thesis was qualitative interviewing. I conducted eleven semi-structured interviews during March and April 2018. This means that I prepared the formulation of the questions posed and the sequence in which I posed them. The strategies for conducting elite interviews from Harvey (2011), prepared me for what kind of reactions I could expect during the interviews with elite women in high leadership positions in the DAF. Harvey for example explained that elite interviewees would sometimes overthink and analyze their answers very much. Therefore, I did sometimes diverge from the prepared sequence and posed the questions differently when the situation allowed this to happen.

The interviews all took a minimum of one hour and a maximum of one and hour and a half. Most interviews took place in a small conference room in the Ministry of Defense in The Hague. Some interviews took place in the interviewees’ workplaces or the canteen of their workplaces and one interview took place in the home of an interviewee. The respondents had the option to choose a location where they would feel at ease. I wanted them to feel at ease. I started the interviews just with knowledge of the glass cliff metaphor, key publications on the token framework and some basic knowledge about the DAF. In the process of interviewing, I

26 started reading more about possible reactions and of the token framework by other scholars that seemed important at that moment. However, the topics on the topic list, based on the sensitizing concepts, remained the same over the course of the data collection. I will further discuss the use of the sensitizing concepts of visibility, polarization and assimilation in section 4.4. Appendix A contains the topic list for the interviews. The first part of the interview consisted of introducing one another, discussing the goal and implications of the research project. Explaining the procedure, the topics of the interview, discussing, and signing the consent form took a few minutes. I asked the questions that I formulated based on the sensitizing concepts and the research question. These first interview questions are introductory and some questions do not directly relate to the sensitizing concepts but ask about the past, current and future career of the women. The idea behind this strategy was to enable the women to talk about their positive and negative experiences in their current positions. In addition, information about the women’s careers provided background knowledge for me to get a better understanding of the context they work in. Questions further down on the topic list focused more specifically on the sensitizing concepts. These questions often brought up topics that are more sensitive than the earlier topics. Therefore, the order of the question mostly worked very well.

Furthermore, the open-ended interview questions created the opportunity to answers that were not restricted by the literature. Moreover, the interviewees could bring up examples of experiences that are meaningful to them. The interviews were audio recorded and transcribed verbatim. Audio recording of the interviews allowed me to focus on the interview without having to worry about taking notes all the time. I used two audio-recorders, in order to be sure the interview would at least have one good recording. However, I did take notes when the interviewee said something I wanted to remember right away, if there was important body language or if I wanted to ask a follow-up question later on. In addition, the audio recordings allowed me to use quotes in this thesis. The interviews and the transcripts are in Dutch. I translated the quotes a literal as possible, with the intention of obtaining the meaning of what the interviewee said. The use of examples from the transcribed interviews made the data more alive and gave an in-depth, detailed exploration into the experiences of women in high leadership positions. In addition, the transcribed examples provided evidence of the collection of empirical data and showed what language and behavior was apparent in the research field (Boeije, 2010, p. 201).

In addition, from the start of this research project, several documents were used as data. Important were recent policy letters on emancipation, diversity and inclusivity from the Ministry of Defense and the Dutch government, earlier research on female leadership and

27 research from researchers from the NLDA, the Dutch Defense Academy. Those documents were useful for the chapter on research setting, the background and characteristics of the case. However, interviews were my main source of data for answering the research question. I stored all the data: the audio files, the transcripts, the documents and analyses in sub files in the file thesis on my personal computer and several other disks at home. While storing the data, the main considerations were the safety of the data, preserving the anonymity of the respondents and the management and chronological order of the data, so it would be easy to retrieve during the thesis analysis- and writing process, which the next section will discuss.

4.4 Qualitative data analysis

Already during but mainly after the collection of the data, the process of analysis started. Before reading the transcripts, I already had the sensitizing concepts of visibility, assimilation and contrast, based on the literature review. These sensitizing concepts are broad and general (Boeije, 2010). The specification of the concepts took place in the stage of open coding (Boeije, 2010).

This phase helped me to reduce the amount of data and to store is systematically. Open coding started with analyzing the data in detail. Judging what is relevant was difficult at the start. The research question and the sensitizing concepts provided a guideline in this process (Boeije, 2010). The data describe experiences and in order to name them appropriately, I looked for patterns on a lower level of abstraction than the sensitizing concepts. This phase consisted of reading through the transcripts twice. First, with an open mind. Second, I read the transcripts by asking myself what labels I could ascribe to the data fragments until I could not find any more groups in the data. The number of groups that emerged inductively from the data first increased into 19 labels. The result of open coding is the code table in Appendix C.

The next phase is an example of axial coding (Boeije, 2010, p. 108). I regrouped those 19 groups described in open coding and put them together when they overlapped. The seven remaining labels and their brief descriptions are visible in the code table in appendix D. This table also shows the result of the selective coding phase, in which these more detailed codes and their descriptions were related to the sensitizing concepts (Boeije, 2010, p. 118, 119). I identified patterns within- and connections between the groups in order to organize the data further and to find explanations for what I heard (Hoonaard, 1997; Boeije, 2010). In this way, I tried to make sense of what the women in high positions describe is happening in the field of the DAF. I found that all identified groups fitted within the sensitizing concepts. Some groups were not one-sided; those had aspects that fitted within several of the abstract concepts. Other