The Political Effect of Female Terrorists

Transcription

1 The Political Effect of Female Terrorists Do women become empowered when entering male dominated spheres? A quantitative analysis exploring the effect of prevalence of female rebels on the prevalence of female politicians. Bachelor Thesis Political Science C Department of Government Uppsala University Author: Moa Peldán Carlsson Supervisor: Elin Bjarnegård Word Count: 12445

2 Abstract This quantitative study aims to investigate mechanisms for female empowerment in the context of conflict by examining whether a female entry into armed rebel, insurgent and terrorist movements can lead to an increase of women represented in politics. Three different hypotheses for the relationship are set: that the effect is positive, that the effect is negative and that there is no effect at all. The study argues that one possible mechanism for the relationship is that female rebels become role models for other women when entering the male dominated domain of a rebel movement. These women disrupt and question existing gender hierarchies and ideas of what it means to be a woman. By expanding the idea of what women are capable of doing, these women could inspire other women to enter other male dominated spheres, such as decision-making domains in politics, and hence gain political power. The method for examining this possible mechanism is OLS linear regression between the dependent variable women in parliament and the independent variable women in rebel movements. The result found is that there is a significant positive effect, meaning that the prevalence of women in rebel movements do actually affect the prevalence of women in politics in post- or presentconflict countries around the world. The conclusion is therefore that, at least in militarized societies, women as a group can become empowered when some women enter male dominated spheres as this believably unties the traditional idea of the capabilities of women. Keywords: Female Empowerment, Rebel Movements, Representation, Insurgency, Empowerment Mimicry, Female Rebels, Gender Equality. 2

3 Table of content 1. Introduction Background Aim of the thesis Structure of the thesis 6 2. Previous research and theory Leadership Roles and Empowerment Mimicry The Conflict Context Grievances as the reason for insurgency and political activity Summary and hypotheses Research design and method Methodological considerations Data Dependent variable Independent variable Control variables Results and analysis Descriptive statistics Bivariate regression Multivariate regression Conclusion References Literature Data and coding Other sources 37 Appendix I: Dataset 39 Appendix II: Code Book 41 3

4 1. Introduction 1.1. Background Throughout history and still today, women are the subordinate sex relative to men. In society, this is visible in many ways, for example as women earn less than men and are systematically excluded from power. Further, in the context of conflict, a common interpretation of women has been that they are peaceful creatures and expected to be against war and violence (Sjöberg & Gentry 2007:3ff). The famous philosopher Friedrich Hegel launched the concept of women as Beautiful Souls (Hegel in Elshtain 1987), and means that these are fragile, removed from reality, and in need of protection in a way that the protector receives substantial honour for success (Sjöberg 2006; Elshtain 1992a, 1992b in Sjöberg & Gentry 2007:3f). Today, the mission of empowering women can be regarded as one of the most important tasks in the world, and the subordination of women in different domains must be eradicated. In 2017 women held only 24 percent of the world s national parliamentary seats (The World Bank 2018). Although the percentage of women in parliament is rising, it can still be regarded as a relatively low number. Female underrepresentation in political, economic and public life is today a widely recognized issue and the empowerment of women constitutes one of the United Nation s Sustainable Development Goals (UN Women 2018). One of the targets is specifically to ensure women s full and effective participation and equal opportunities for leadership at all levels of decisionmaking in political, economic and public life (ibid). An increase of female representativeness in politics could arguably serve as a force for gender equality. Partly since political decision-making power is something that brings about legitimacy, status and power for these women in politics, and partly because with more women in politics, there is possibly a greater focus on gender equality in society (Phillips 1994). The mechanisms for an increase of female representativeness in politics, that is, how we can understand what increases female political representation, has been studied widely. In the conflict-torn world that we live in today, the study of how conflict affects the trajectory of political change has received increased attention by scholars. Melanie M. Hughes (2009:190), for example, argues that wars serve as defining moments, turning points, and catalysts for transformation and changes, for example, opportunity structures and creating new space for women to enter politics. Further, Aili Mari Tripp (2015:194f) claims that the increase of women s political representation in Africa is most visible in the continent s post-conflict 4

5 countries, indicating that conflict plays an important role in the change of sex ratio in parliaments. For example, Algeria with its history of civil war between 1991 and 2002, has the Middle East s highest rate of female representation with its 32 percent women in parliament (ibid), and in Nepal the percentage of women in parliament increased from 5,9 percent at the beginning of the civil war to over 33 percent after the civil war (Rajivan & Senarathne 2011:9). In relation to this, Hughes (2009:190) poses a possible mechanism for how conflict increases women s power: by entering armed forces. When it comes to countries that have been, or still are, involved in armed conflict, i.e. militarized states, the rules, norms and conditions of civil society and political trajectories may be fundamentally different from other, peaceful, contexts. Therefore, the mechanism for increased female participation in political, economic and public domains may too differ from other settings. As Hughes suggests, perhaps in militarized states women gain power by joining conflicting domains. And as Tripp claims, when there has been conflict there are more women in politics. Could this be because the women fight? The historical record shows that women have, just like men, taken on roles as combatants in times of conflict. Generally, women make up percent of liberation movements, armies and guerrilla forces (Bouta, Frerks & Bannon 2005:11). Women did, for example, make up one third of the 15,000 fighter force in the Sri Lankan rebel group of the Liberation Tigers of Tamil Eelam (LTTE) (Goldstein 2001:83), one quarter of the combatants in the Farabundo Martí National Liberation Front (FMNL) in El Salvador, and 30 percent of the soldiers in the Nicaraguan Sandinista National Liberation Front (Karamé 1999). To test the theory whether women in armed forces actually do bring about greater power for women, this study will examine whether female participation in armed rebel combat can possibly lead to an increased female participation in one of the highest political power domain within a country the parliament. As women are most visible as combatants in insurgency and liberation movements rather than in state armies (Goldstein 2001:76ff), the focus of the study will be on female rebels. Therefore, the research question of the study is posed as: How does female participation in armed rebel, insurgent, and terrorist movements affect female participation in politics? 5

6 1.2. Aim of the thesis By studying this, I will explore whether there is a relationship between women in rebel movements and women in politics and if so, I will further discuss how we can possibly understand this relationship. Can a female entry into male dominated domains (for example rebel movements) make these women role models for other women to enter other male dominated spheres (for example politics)? And can this further bring about greater female representation and thus greater gender equality in society s different spheres, domains and institutions? By generalizing the findings of this study to other fields regarding female empowerment, perhaps it can contribute to the discussion whether female empowerment is more successful when women enter male dominated spheres, or whether there is a need for a greater legitimization of female dominated domains in order to empower women. If women want to become powerful, is the best way to get there to mimic the powerful men, or to question and try to change the existing power structures of society as it is today, so that female dominated domains become equally legitimate and powerful? The previous research within the field of female insurgency has mostly been of case studies, or studies with a small N (Henshaw 2017:116). As this is a quantitative study with an ambition to get a cross-country perspective of effects of female insurgency, I believe this study fills a gap within the field of trajectories for female empowerment in conflicting contexts. By getting a clear answer to how female insurgency can create opportunities for female empowerment, I believe this study could serve as complementing the field with answers regarding the effects of female entry into male dominated spheres, and as a great starting point for future studies of the mechanism of female empowerment in militarized societies Structure of the thesis In order to answer the question regarding what effect female participation in rebel movement has on female participation in politics, I will first present the previous research within the field. From the discourse on female empowerment in times of conflict, I will formulate hypotheses regarding the possible relationship between women in rebel movements and women in politics. Second, I will present the quantitative method used for answering the research question and explain the data and variables used in order to reach an answer. Third, I will present the results of the study, examine these and analyse them. Finally, I will conclude the study with an answer to the question about the relationship between prevalence of women insurgents in a country and 6

7 women in the country s national parliament, and discuss possibilities for further research within the field. 2. Previous research and theory In order to understand the relationship between prevalence of female rebels and prevalence of female politicians, I will first examine the previous research and theory within the field of empowerment, representation, and female rebels. First, I will investigate theory of leadership roles in political spheres and empowerment mechanisms to see how the increase of female rebels could work as an empowering force. Second, I will go through what special conditions the conflict context brings to the situation and how the perception of women as armed fighters in conflict influences the trajectory for female empowerment. Third, I will study the motivations for why women join armed rebellion and examine its connection and similarity to why women engage in politics Leadership Roles and Empowerment Mimicry The scholar Anne Phillips (1994) is famous for her arguments regarding the importance of female participation in politics. She divides the arguments for increased female participation into four parts: the role-model argument, the justice argument, the women s interests-argument, and the argument for a revitalized democracy (Phillips 1994:228ff). Phillips theory of female representation concerns representation in the domain of politics and more specifically parliaments, yet I believe it is possible to generalize Phillips theory of female representation to other political domains as well. Even if rebel movements are violent and illegitimate domains, they could be argued to be political. The aim of a rebel movement is to influence the outcome of [a] stated incompatibility [that concerns government] (Themnér 2017:2f), i.e. to change existing political structure and behaviour. Therefore, Phillips theory of representation should be applicable to armed rebel, insurgent and terrorist movements as well as parliaments, especially in the context of militarized states, as rebel movements might be a central trajectory towards power in these societies. Further, as liberal feminists often argue, women should be equal to men on all political levels, thus including rebel movements. The scholar Joshua S. Goldstein (2001:39) argues that the exclusion of women in war is similar to the exclusion of women in other high positioned roles throughout history. The inclusion of women in rebel movements could therefore be regarded as feminist and empowering for women. 7

8 Out of her four arguments for why female representation is vital, Phillips (1994) emphasizes the justice argument, the women s-interest argument and the legitimate democracy argument, and claims the role model argument to be the least interesting. However, in this study, I will focus on this specific argument: that role models play a fundamental role in female representation. Phillips role model argument is that when women enter politics, they raise a good example for other women to do the same, showing them it is possible and appropriate for both men and women to take part in political decision-making (1994:228). Further, the role model argument is something widely studied by other scholars. For example, David E. Campbell and Christina Wolbrecht (2006:324) agree that female political activity encourages other women to enter politics as well, and Latu et al. (2018) argue that this goes for all leadership positions, and not just the political. Female leaders can show other women how to act in certain situations and contexts, thus producing empowering effects. It is the actual visibility of women taking on leadership roles that account for other women to become inspired and encouraged to do the same. If the presence of women in male dominated spheres results in more women feeling encouraged to enter male dominated spheres, this could be referred to as empowering mimicry, where women mimic the powerful behaviour of the leaders (Latu et al. 2018:2). The rebelling women have greater access to power to disrupt and hence get a political voice, and could therefore possibly serve as encouraging for other women who want to challenge the existing gender roles within society, and consequently receive a political voice as well. If some women can change the conditions of society by rebelling, others might feel inspired to do the same in more peaceful (or violent) means. Thus, if women engage in rebel movements against the government to ensure women s rights and opportunities, the role model argument could be extended to women serving as insurgents and not only politicians, creating new spaces for women to become empowered. Empowerment is a concept with its base in power and indicates both the process of gaining more power and the outcome of having gained power (Wright 2008:8). The scholar Naila Kabeer (2005) claims one can talk about empowerment and power through three different dimensions: agency, resources, and achievements. Kabeer further points out that empowerment requires change, you cannot become empowered unless you have in some way been disempowered, or started off with no power in the first place. The process by which one can become empowered is represented by agency, which can be referred to in two different ways: a positive sense in which one has power to, meaning people receive the ability to make their 8

9 own choices, and a negative sense where one has power over, referring to the use of authority, violence and other forms of coercion (2005:13f). The medium in which agency is exercised is resources, and the result of agency through resources is achievements (ibid). Table Dimensions of empowerment Process: Agency Medium: Resources Result: Achievements (Kabeer 2005:13ff) The resources in which one can become empowered are often talked about as political and economic spheres of society, mediums highly recognized as legitimate. However, what happens when or if women gain agency in male dominated spheres seen as illegitimate? Thus, as this study aims to explore, what happens when women enter the male dominated sphere of a politically violent organization? The scholar Melanie M. Hughes (2009) actually do claim that women gain power in conflicting contexts in this manner; by entering into armed rebel force, because it disrupts conventional gender roles. (Hughes 2009:190). She states that Soldering may transform the identity of women (ibid 2009:178), and further, the scholar Codou Bop (2001:21) suggests that women serving in combat creates a sense of power leading to a change in the way which these women possibly perceive themselves, thus feasibly further creating a change in the collective identity of women. Can this be regarded as empowering mimicry and hence a trajectory for empowerment, even if it occurs in an illegitimate medium? One reason women feel more powerful by entering male dominated spheres could arguably be because women become more similar to men, as they take on roles traditionally possessed by men. And as men are the ones with more power in the world and society as we know it, men in these domains represent power. Therefore, by becoming more like a man one becomes more powerful. An illustrative example of this sense of worth and power when becoming more like a man is the case of Nancy Saungweme, a female rebel in the Zimbabwe liberation war, who stated that: After two months of training I stopped menstruating and so did the other women. We were just like men. We were pleased because we thought it was macho. We wanted to be identified as fighters, as men (Musengezi 2000:48 in Meintjes 2002:67). 9

10 A possible suggestion for the prevalence of women in rebel movements affecting the prevalence of women in national parliament in a positive manner has therefore risen here: the identity of women change by fighting in combat, in the way that they become more men-like. Women fighting in armed rebel movements become examples of the fact that women can do what men already do, as they now have the characteristics of men. Perhaps this could then apply to others spheres of life as well; as men are more represented in politics, the idea is that women therefore might feel they could enter these roles as well, now that they possibly feel inspired and encouraged to perform these male acts. Thus, the act of fighting in rebel movements can arguably open political opportunities for women, as women gain agency by challenging previously given power relations and resources by participating in the space of conflict, resulting in female rebels becoming leadership-role models for other women. In other words, when women enter male dominated spheres and gain characteristics of power typically possessed by men, they feel empowered as men represent power. If one believes the theory of role models and empowering mimicry, these women should then become role models for other women to also gain power in different political domains, which can be summed up in the following hypothesis: Hypothesis 1: the prevalence of women in armed rebel movements has a positive effect on the prevalence of women in politics The Conflict Context As this study aims to examine whether women become empowered when entering male dominated spheres, it is of significant importance to discuss the special context of conflict and militarized domains in this case. A rebel movement is just not any male dominated sphere, it is a highly violent and dangerous domain. Can women really become empowered or serve as role models by entering these illegitimate domains? Further, even if women do become empowered by entering armed forces, Hughes (2009:180) among many other scholars in this field of research points out that these mechanisms for women becoming empowered during conflict the disruption of gender roles and new opportunities might not last into the post-conflict society. Wright (2008:38) claims that the empowered roles of women during conflict become trivialized in the post-conflict society, since women s contributions during war are seen as 10

11 extraordinary and a kind of crisis management, thus having no place in conventional society. Further, Christine Chinkin claims that this return to traditional gender roles is inherent in the post-conflict terminology of for example reconstruction, referring to a way of going back to the former situation of society (Chinkin in Wright 2008:46). In other words, the priorities during war can be stated to be different than those in a context of peace, and that the inclusion of women in new spheres is just a matter of survival and success. This would mean that the entry into male dominated domains during times of conflict has nothing to with the sex ratio in national parliaments, as the mechanism is broken when going back to conventional societal structures in the post-conflict society. This leads to the following hypothesis: Hypothesis 2: the prevalence of women in armed rebel movements has no effect on the prevalence of women in politics. In addition to this, one should consider the conventional perception of female fighters. In the case of women turning to armed rebel movements, they are not just women acting like men, they are women committing crime and violent actions. The image of women as violent actors has been said to disrupt feminist views of women as capable and equal, resulting in a systematic exclusion of women who turn to combat roles in rebel movements (Sjöberg & Gentry 2007). In the post-war period in Eritrea, for example, the National Union of Eritrean Women (NUEW) were strong advocates for the political empowerment of women, but their campaign strategies did not include female ex-combatants, because these were not fit for politics with the argument that female ex-combatants were uneducated and their only skill was fighting (Meintjes 2002:73). Female rebels are often seen as less than women and as less than human, and as a flaw in women s femininity (Sjöberg & Gentry 2017:10f). In relation to what Chinkin states regarding reconstruction, this has in many cases also meant that the needs of female excombatants have been overlooked and that female ex-combatants have become even more marginalized than the regular women in a pre-conflict context. Thus, the possible agency gained by fighting in armed conflict is reduced to a minus in the aftermath. As not fitting into the conventional identity of what it means to be a woman, female rebels are claimed to be ignored in feminist scholarship because the reality of feminism, according to some, is highlighting female superiority and not advocating for equality between the sexes (Morrissey 2003:156 in Sjöberg & Gentry 2007:18). Thus, when women commit crimes, they are denied legitimacy. If this is the general case, the hypothesis of female empowerment in 11

12 wartime spilling over to female empowerment in politics could be considered false, because it does not lead to empowerment, but stigmatization for some women. Further, this stigmatization could in the post-conflict society possibly lead to an emphasis that women should go back to the conventional idea of what it means to be a woman and take on traditionally female roles (Sjöberg & Gentry 2007). Women wanting to gain power in any way could hence possibly be seen as threatening and rebellious, even if it is getting power through peaceful means, because it is stepping out of the traditional female role. One could then arguably even imagine a negative relationship between women in rebel movements and women in politics. Thus, this leads to the following hypothesis: Hypothesis 3: the prevalence of women in armed rebel movements has a negative effect on the prevalence of women in politics Grievances as the reason for insurgency and political activity In her study, Alexis Leanna Henshaw (2017:36) argues for the reason women turn to rebel forces and how this disrupts gender roles. She claims that one can understand the reason why some women turn to rebellion in two ways. First, women go into rebellion as a manifestation of protest of existing structures in society. These structures could, for example, be about gender norms, thus some women rebel to disrupt or deconstruct these, aiming towards creating greater space for women to operate. Second, women go into rebellion as an opening for personal gains. The two reasons could nevertheless of course be seen as interlinked; when (or if) women in rebel become empowered they have succeeded in both disrupting gender roles and benefitted from this personally, through the fact that they belong to the group of women. What is important to note here is that the reason women rebel is similar to the reason why women become politically active in decision-making spheres: to get access to the system and to change it. One can summarize the reasons why women turn into rebel (and to political activity) into categories of political grievance, human security grievance, economic grievance and ethnoreligious grievance (Henshaw 2017:78). The political grievance concerns the force to ensure greater rights for women in the political system, the human security reason is to ensure that women are free from threat and oppression, and the economic reason is to ensure an economic redistribution favouring women. Women joining rebel movements for the three first reasons often lead to women having combatant roles in the movements, whereas women joining for greater rights for their ethnic or religious group often serve in the movements as non- 12

13 combatants (ibid). Therefore, I will focus on the political, human security and economic grievance in this study, as my subject of study is female combatants, i.e. women bearing arms. Further, the political, the human security and the economic reason women enter rebel movements are as noted possible reasons to why women enter politics as well. It is about the quest for change and empowerment of women in both cases the main difference is that armed rebel movements as opposed to national parliaments, are seen as illegitimate as they use violence in order to achieve their goals. If the reason why women enter rebel movements are the same as why women engage in politics, then that would mean the possible relationship between the prevalence of women in rebel movements and the prevalence of women in politics is false. In this perspective, there is perhaps something else affecting both situations, i.e. if one is affected because of greater grievance, the other should be affected as well. In other words, this would mean there is no relationship between the prevalence of women in rebel movements and the prevalence of women in politics, which supports hypothesis 2, stated as: Hypothesis 2: the prevalence of women in armed rebel movements has no effect on the prevalence of women in politics. One disclaimer is important to make here. The reason women become insurgents are difficult to generalize, because the reason to join a rebel movement is highly individual. Henshaw (2017:125) claims that access to the system is not always the initial motivation for women to fight and women are in some cases forced to join, directly, or indirectly because they have no other choice. It is also important to note that women, both serving as combatants and not, often are victims of the war as well especially in new war strategies such as rape and forced marriages (Hughes 2009:178). Further, even if women turn to rebel movements and becomes combatants, they can still suffer from sexual violence and marginalization. For example, South African black women who joined the liberation movement reported rape and sexual abuse by their male comrades and other state security forces (Sideris 2002:48). Therefore, it is difficult to generalize the reason why women turn to rebel forces. Although, Henshaw argues that even if there is no political, human security or economic motivation at first, sometimes the actual engagement in insurgent movements results in the wake for a desire for fighting for women s rights in these individuals after they have joined (Henshaw 2017:92). 13

14 2.4. Summary and hypotheses The previous research within the field of representativeness and empowerment in conflicting contexts states that female role models and empowering mimicry can be a central mechanism for female empowerment in militarized societies. Yet, this is not a given fact, especially since the mechanisms of empowerment in times of conflict may not last in the post-conflict society, and further, female rebels may become stigmatized as opposed to becoming role models. Additionally, perhaps one cannot even claim there is a connection between women in rebel movements and women in politics because of other factors affecting both situations equally. With these being the theoretical starting points, three main hypotheses regarding the relationship between prevalence of women in rebel movements and prevalence of women in politics have been formulated, these being: 1. the prevalence of women in armed rebel movements has a positive effect on the prevalence of women in politics. 2. the prevalence of women in armed rebel movements has no effect on the prevalence of women in politics. 3. the prevalence of women in armed rebel movements has a negative effect on the prevalence of women in politics. Accordingly, the hypotheses suggest that the effect of prevalence of women in rebel movements on the prevalence of women in politics can occur in three different ways: a positive effect (H1), a negative effect (H3), or no effect at all (H2). The aim of this study is to determine which one of these hypotheses is correct. 3. Research design and method 3.1. Methodological considerations In order to answer the question if, and how, female participation in rebel movements affect female participation in politics, I have used the quantitative method of OLS linear regression analysis. My aim is to explore and statistically test the effect of prevalence of female rebels on the prevalence of female politicians, and further control this relationship for other underlying 14

15 or intervening variables, such as political, human security, and economic grievances affecting the relationship. As this study tries to explain an effect, one can talk about the four criteria for such an explanation. First, the explanation should illustrate a contra-factual difference. This means it must prove that if x doesn t occur, neither does y, ceteris paribus. In this case that would mean the study needs to prove that if an increase in female insurgents doesn t occur, the percentage of women in politics won t increase either. Second, there needs to be a clear order of time, that x is prior y in time, i.e. that prevalence of female rebels is prior to prevalence of female politicians in time. Third, one needs to isolate the explanation, meaning ruling out all other possible explanations for increased percentage of women in national parliament. Fourth, one should try to explain the mechanism of how x affects y, i.e. how the prevalence of women in rebel movements affect the prevalence of women in politics, in what way does that happen? (Teorell & Svensson 2007:64). The quantitative method used in this study has an advantage in the first criterion, proving contra-factual difference and the third criterion, to isolate the explanation. When it comes to time order, OLS regression can possibly put one in problematic situations. The regression analysis can state a correlation, but it cannot provide the information of which one of the variables that stand before the other in time. I will get back to this issue further down in the study. The research question of this study is to plot in what way the occurrence of female rebels affect the occurrence of female politicians; if there is an effect and whether it s a positive or negative effect, thus the focus is not to closely study what happens in-between x and y. This means the fourth criterion of examining the mechanism in between x and y turns out to be less relevant for this study. It would, however, be something (highly interesting) for future research, but initially there needs to be clarity in the relationship between female rebels and female politicians, which this study aims to investigate Data As discussed, the previous research within the field of female insurgents has mostly been qualitative studies of single case studies, or studies on a small-n level (Henshaw 2017:116). The main reason for this is the lack of data on women in war and in rebel movements. Perhaps this shortage could be argued to be in line with what I have already discussed, that contemporary feminism denies women that do not support the quest for claiming the female superiority 15

16 because these women are characterized as bad. However, in 2017, the scholars Reed M. Wood and Jakana L. Thomas launched the Women in Armed Rebellion Dataset (WARD), providing information of the prevalence of female combatants in 211 rebel organizations involved in armed conflict between 1979 and 2009 (Wood & Thomas 2017). With this dataset, I received the opportunity to study the effect of female rebels in a more extensive way. Further, as I find it fascinating and important to study the concept of terrorism in order to question the existing power structures the idea of terrorism holds, this dataset provides the possibility to widen the discourse on terrorism and tangle out possible mechanisms of empowerment within the field of militarized domains. For data on the other variables included in the study, I have used The World Bank (2018), Freedom House (2018) and The United Nations Development Programme s (UNDP) Human Development Reports (2018), as these are acknowledged sources providing data on levels of democracy, development and peace globally. Important to note about the data is that there is a shortage of data even in this study. Some countries lack data in certain categories because of issues of conflict, new states rising, old states falling etc., thus unfortunately making the study less reliable as it includes unsystematic errors. This also means that in the regressions, the N also varies, possibly making the results of the study less clear Dependent variable The dependent variable of the study is prevalence of women in politics, operationalized as percentage of seats held by women in national parliament. One could possibly argue whether women in politics automatically translate into women in parliament, since there are many different forms of political activity. Yet, as a national parliament is one of the highest decisionmaking political domain within a society, one could definitely state that parliamentarians are politicians, though it is not the only politicians in a society. Though, as this study s operationalization of politics is national parliaments (which is a conventional operationalization by other scholars as well), it makes the theoretical definition of women in politics and the operational indicator of women in parliament somewhat closely linked which in turn provides validity the absence of systematic measurement errors to the study. Yet, this is of course arguable and some may claim that the concept of politicians includes much more than the people represented in parliament. 16

17 To gather information regarding this I have used The World Bank s dataset on proportion of seats held by women in national parliaments, as the percentage of parliamentary seats in a single or lower chamber held by women (The World Bank 2018). I have chosen to measure percentage of seats held by women in national parliament in the most recent year with data available: The reason for this is to create the greatest possibility for societies and female identities to transform and consolidate as this may take some time. Although, the choice of data in 2017 provides two possible issues. First, regression analysis cannot determine time order of the dependent variable and the independent variable, as discussed earlier. This means that the order of the two could theoretically be the other way around: perhaps the prevalence of women in parliaments affects the prevalence of women in rebel movements, i.e. the opposite of what this study aims to investigate. This one cannot be sure of. When choosing to study the year 2017, it makes the analysis less valid since women in parliament in the future cannot affect women in rebel movements in the past, even though the statistical conclusion from the regression analysis might be just this. Yet, having in mind that the identities, norms and hierarchies can require time to transform, the data from 2017 can perhaps provide some information about the effect of women in rebel movements on women in parliament today, making it a relevant way to measure the dependent variable after all. Though, this further accounts for the second issue. Choosing a year this long after the measurement of women in rebel movements (as women in rebel movements are measured between 1979 and 2009), there can be a lot of other factors having had effected the prevalence of women in parliament today, and therefore make it a lot more difficult to measure the sole effect of prevalence of female rebels on this. I will get back to this issue later Independent variable The independent variable of the study is prevalence of women in armed rebel, insurgent and terrorist movements, sometimes shortened in this study as women in rebel movements, female rebels and female insurgents. How is this measured? As the WARD is based Uppsala Conflict Data Program s (UCDP) Dyadic Dataset (Allansson, Melander & Themnér 2017), the operationalization is taken from the UCDP s dyadic dataset s definition of armed conflict, which is: a contested incompatibility that concerns government and/or territory [with] the use of armed force between two parties, of which at least one is the government of a state (Themnér 2017:2f). 17

18 One of these parties, the rebel movements, or oppositional organizations, is defined as: any non-governmental group of people having announced a name for their group and using armed force to influence the outcome of the stated incompatibility [ ] The focus is on armed conflict involving consciously conducted and planned political campaigns rather than spontaneous violence. (ibid 2017:3). Moreover, it is in these groups that the information regarding female combatants are collected, where female combatants are defined as: women and girls who participated in armed conflicts as active combatants using arms... This category includes women employed in frontline combat, female suicide bombers or assassins, female members of armed auxiliary and civil defense forces, and other female members of the organization who receive military training, carry combat weapons, and would be expected to participate in combat operations if called upon. It excludes women who served in noncombat support roles such as fundraisers, couriers, cooks, and porters but did not directly engage directly in combat (UN Women 2012:22f in Data Appendix: The Social Origins of Female Combatants Wood & Jakana 2017). Accordingly, the independent variable prevalence of women in armed rebel, insurgent and terrorist movements is operationalized as prevalence of actively armed female fighters fighting in non-spontaneous non-governmental groups with announced names, doing planned political campaigns to influence the outcome of a contested incompatibility against the government. The close interrelation between the theoretical definition and the operational indicator in this case, arguably gives the study solid validity. In order to measure the prevalence of female rebels within a country, I have translated the WARD s variable of women in rebel organizations involved in armed conflict, into the variable of countries in which female rebels have occurred. This in order to be able to study its effect on women in national parliament, since national parliaments are studied country-wise 1. 1 See the appendix for further information about coding and measurement of the variables. 18

19 Further, the WARD includes two different measurements of occurrence of female rebels: one prevalence variable and one binary variable. The prevalence variable is measured on a categorical scale where the countries are divided into the three categories of low (<5%) prevalence, moderate (5 20%) prevalence, and high (>20%) prevalence of female rebels in the organizations 2. It is not always clear exactly the percentage of female fighters within a rebel movement, and therefore even less in the countries, which of course makes the result less clear. Yet, since there exists no data today on the exact percentage of female rebels within organizations, and thus not on female rebels within countries either, this is the data available to answer the research question. This might not be able to reveal the absolute truth, but at least it provides a sense of the relationship between the prevalence of female rebels and female politicians. The binary variable constructed for the occurrence of female rebels in the organizations is divided into whether female rebels have occurred in an organization or not. I have further translated this variable into a binary variable of occurrence of female rebels in the countries, as with the prevalence variable. This provides a clearer measurement of the occurrence of female rebels in the countries, but tells us nothing about how many percent of the rebels were women. Thus, since the focus of this study is on the frequency of female rebels and its effect on the frequency of female politicians, the prevalence variable can possibly provide a more interesting result than the binary since one can somewhat measure levels of prevalence and thus learn about the effect on levels of prevalence of female politicians Control variables In order to clarify the relationship, control variables are added to the regression. This is to make sure it is not something else affecting the relationship between the independent variable and the dependent variable, as some previous scholars have claimed. I.e. one must control for the effect of other variables. In my multivariate analysis, I want to control the relationship between prevalence of women in rebel movements and prevalence of women in politics today for political factors, human security factors and ideological factors, i.e. political, human security and economic grievance. 2 ibid 19

20 As Henshaw (2017:38) argues, women sometimes turn to rebel movements for political grievance. In some cases, the goal of insurgency is to achieve access to and rights from the state. In the same manner, this is sometimes the reason people become politically active by ruling for parliament. As I have discussed, one can argue that both national parliaments and insurgent movements are political domains. Therefore, the reason for why women rebel and why women rule for parliament might be the same, thus a variable z affecting both the independent variable x and the dependent variable y, making it relevant to control for. The political factor possibly influencing the relationship between the independent variable and the dependent variable is operationalized as the mean freedom rating of the country during the same years as women in rebel movements are measured in the WARD. The data on freedom rating is collected from Freedom House (2018) and is based on subcategories from the Declaration of Human Rights representing individual s political rights and civil liberties 3. Another reason why women enter rebel movements, according to Henshaw, is human security grievance. Henshaw argues that when people are exposed to threats to their security by the oppression of state, a desire for human security develops (2017:45), which in some cases leads to women entering rebel movements to secure their safety, rights and opportunities. Just as in the political case, these are reasons people turn to peaceful political activities as well, since it is another way of accessing the system and changing existing politics. Therefore, this becomes relevant to control for, since it is another variable z possibly affecting the relationship between x and y. How is human security measured? Since it is difficult to discuss in general terms what human security means it becomes challenging to measure. UNDP claims that human security is when people can exercise their choices and opportunities safely and that these are protected (UNDP 2003). There is also a great emphasis on threats and the struggle for freedom from these. Therefore, human security is embedded in the first control variable of freedom rating, however I have added two measurements to complete the measuring and to try to capture as many aspects as possible. The first variable added consists of mean Human Development Index (HDI), an aggregate measure of decent standard of living, a long and healthy life and knowledge collected from UNDP s Human Development Reports (2018). Human security and human development are highly interlinked, as they both focus on: 3 See the appendix for further information about coding and measurement of the variables. 20

21 the right to live in freedom and dignity, free from poverty and despair [ ] with an equal opportunity to enjoy all their rights and fully develop their human potential. (UN General Assembly 2012). The second variable added to the measurement of human security is mean percentage of women in the work force for the 64 countries included in the study, as ability to work and the effects of work are not the focus of HDI. The data on this is collected from The World Bank s (2018) dataset over female participation rate in labour force between 1990 and 2017, as these are the years with existing data. Enjoyment of having a job as a woman does not automatically indicate human security, but can contribute to it. Hence, by adding HDI and percentage of women in the work force as a measurement for human security, in combination with freedom rating, one can get a clearer view of how much freedom of choice and opportunities people enjoy. A higher human development and more women in the work force in many cases equal higher human security as these are as discussed interlinked, although one should stress the fact that that high human development does not automatically translate into human security. The last control variable in this analysis is the ideological one. Henshaw (2017:41) argues that one grievance related to engagement in insurgency is the economic, and that women enter rebel movements to fight for economic redistribution. This most often takes the form of women entering leftist oriented rebel groups. Groups with a leftist orientation often propose an essential challenge of existing social hierarchies in their quest for liberation. This naturally involves the patriarchy as well as other existing social hierarchies. These leftist movements, often drawing inspiration from Marxism and Maoism, rely on principles of the fact that women s equality and liberation is a crucial part of revolution, as the communist leader Mao Zedong stated (Henshaw 2017:41). Further, Mao Zedong also emphasized the importance of recruiting women into the liberation movement, and evidence (although somewhat unreliable) indicates that when a rebel movement has a leftist ideology, they tend to give women greater opportunity to decisionmaking positions within the organization (ibid 2017:34f). Therefore, there is great reason the believe that if a country has hosted a leftist oriented armed rebel group, there is a higher prevalence of rebel women in the country as well. And just as in the political and the human security case, the same could be considered for peaceful political activity: if there are leftist currents in society, this might call for greater representation of women in parliament as well, as 21

22 these aim to change the structures of society oppressing women. This ideological control variable measures whether leftist movements have occurred in the country or not. To sum up, the variable measuring political grievance is freedom rating, the variables measuring human security grievance is freedom rating, HDI and women in work force, and the variable measuring economic grievance is leftist rebel movement 4. These control variables have been added to the analysis as they represent different reasons why women engage in insurgency and in peaceful political activity, i.e. running for parliament. If there is something affecting both the dependent and the independent variable, then one can claim the relationship between prevalence of female rebels and prevalence of female politicians to be false, as hypothesis 2 claims: the prevalence of women in armed rebel movements has no effect on the prevalence of women in politics. 4. Results and analysis 4.1. Descriptive statistics Before exploring the relationship between the prevalence of female participation in rebel movements and the prevalence of female participation in politics, one can begin with studying the descriptive statistics of the variables. This might provide further insight into the case and help determine which one of the variables are best to use when measuring the prevalence of women in rebel movements, since it can be measured in two different ways. Table Descriptive statistics of the dependent variable and the independent variable Descriptive statistics: Min 1Q Median Mean 3Q Max Dependent variable % of women in parliament ,74 30,5 61 Independent variable Women in rebel movements (binary) , Women in rebel movements (prevalence) 0 0 0,5 0,82 1, See the appendix for further information about coding and measurement of the variables. 22

23 Table presents the mean prevalence of women in parliament in 2017 to be 21,74% in the 64 post- or present-conflict countries included in the study, which somewhat below the world average of 24%, as discussed previously. Further, the table provides us the information that the mean prevalence of women in rebel movements (in the binary variable) is 0,66 and the median is 1, indicating that there are more countries in the dataset where female rebels have occurred than not. When studying the prevalence variable of women in rebel movements, the mean of the 64 countries is 0,82 on the categorical scale of category 1 (<5% prevalence), category 2 (5 20% prevalence) and category 3 (>20% prevalence). This implies that the mean prevalence of women in rebel movements in the countries are somewhere below 5%, as 0,82 doesn t even reach category 1. Studying the distribution of prevalence of women in parliaments, the fact that the value of the third quartile is 30,5 means that in 2017, 75% of the countries in the study had 30,5% or fewer women in parliament. When it comes to prevalence of female rebels, the variable value of the third quartile (of the prevalence variable) is 1,16 meaning that 75% of the countries have around 5% or fewer women in their rebel movements. The fact that the value of the first quartile is 0 and the median (or the second quartile) is 0,5 tells us that 25% of the countries have no women in rebel movements and that 50% of the countries don t even reach the first category of less than 5% women in rebel movements. Thus, only 25% of the countries in the dataset host 5% or more women in the existing rebel movements. The conclusion is therefore that in about half of the countries female rebels have occurred (because of the median being 1 in the binary variable) but that the prevalence of female insurgents in these countries are still low. With the information provided by the descriptive statistics one can conclude that the 64 postor present-conflict countries included in this study have a relatively low number both when it comes to prevalence of women in parliament and prevalence of women in rebel movements, which leads to the research question whether this is correlated or not. 23

24 4.2. Bivariate regression In the initial bivariate linear regression, measuring the relationship between women in rebel movements and women in national parliaments, I measure both the prevalence variable and the binary variable. The reason for measuring both variables is to get the most accurate result, since both variables aim to measure the occurrence of female rebels, but in different ways. The dependent variable in the bivariate regression is percentage of women in parliament Table Bivariate regression with percentage of women in parliament 2017 as dependent variable (Standard errors in parenthesis. *p=10, **p=05, ***P=01.) Women in rebel movements (binary) 6,607 (3,365) Model 1 Model 2 Women in rebel movements (prevalence) 3,924* (1,718) Intercept 17,333*** (2,747) 18,468*** (2,124) Adj. R 2 0,044 0,064 N Table presents the relationship between frequency of women in rebel movements and frequency of women in national parliaments to be slightly positive, in both the prevalence variable and the binary variable. I Model 1 we see the result of the binary variable. The b- coefficient is 6,607 which means that if we move one step up in the scale of occurrence of women in armed rebel, insurgent or terrorist movements (i.e. from 0 to 1), we move 6,607 steps up in the scale of prevalence of women in national parliament. Substantially, this means that countries that have had women in rebel movements between have around 6,6% more women in parliament in However, the result of the binary variable is not significant, and the result can therefore not be accounted for. If the result is not significant, we cannot rule out the possibility that the result is not just random. Therefore, it becomes irrational to generalize the result to a greater population (Teorell & Svensson 2007:216). One possible reason for this insignificance is that the binary variable measures the prevalence of female rebels within a country in a very stiff way. To say that countries where a single occurrence of a 24

25 female rebel has 6,6% more women in parliament seem quite radical and unbelievable. This can further be supported by the fact that the adjusted coefficients of determination (the adjusted R 2 ) is only 0,044 on a scale between 0 and 1 where 1 equals all variation explained and 0 equals no variation explained. That the adjusted R 2 of the binary variable is 0,044 means that this variable can only explain 4,4% of the variation of women in parliament, thus a not very strong measurement. Model 2 illustrates the result of the prevalence variable, where there is a positive effect on the percentage of women in parliament as well. If we move one step up in the scale of occurrence of women in armed rebel, insurgent or terrorist movements, that means we move 3,924 steps up in the scale of prevalence of women in national parliament, i.e. around 4% more women in parliament. The result is significant at a 90% percent security level, which means we can possibly state that there is an actual effect of the prevalence of women in rebel movements between 1979 and 2009 on the percentage of women in parliament in In the prevalence variable, the adjusted R 2 is 0,064 meaning this variable can explain 6,4% of the variation not a very strong measurement either, however stronger than in the case of the binary variable. The results of the bivariate regression are illustrated in the figures below. 25

26 Figure Illustrated relationship between women in rebel movements between 1979 and 2009 (binary variable) and women in national parliaments 2017 Women in parliament Women in rebel movements (binary) Figure Illustrated relationship between women in rebel movements between 1979 and 2009 (prevalence variable) and women in national parliaments 2017 Women in parliament Women in rebel movements (prevalence) 26