Girls in Burundi : a study of primary-aged schoolgirls coping with violence

(1)

(2)

2

Girls in Burundi – A study of primary-aged schoolgirls

coping with violence.

Cover photo – Pupils at EP Nyakibingo. Source: author.

Master Programme: International Development Studies Faculty: Graduate School of Social Sciences

Institute: University of Amsterdam

Supervisor: Drs. Graciela Paillet Second reader: Dr. Jacobijn Olthoff

Author: Heleen Vis

Student number: 10003317 Email: heleenvis1992@gmail.com 17 August 2015

(3)

3

Dedication

There is no mountain too great. Hear these words and have faith. Have faith.

They live in you. They live in me. They’re watching over everything we see. In every creature. In every star. In your reflection. They live in you.

(The Lion King – They live in you)

I dedicate this thesis to my mother, who I love and miss so much. You always supported me with everything I did wherever I was. Although you may be far away now, I know that you will still travel the world with me.

(4)

4

Acknowledgements

Writing this thesis would not have been possible without the help of many people. First of all I would like to thank all the girls, boys and teachers from EP Mushasha 1, EP Mugutu and EP Nyakibingo. By greeting me with hugs and handshakes, inviting me to play football, and asking the most interesting questions about my life, you were the ones that made me feel most welcome in Burundi. My gratitude goes especially to the children and adults who participated in my focus groups and interviews. Thank you for listening to my questions and sharing your stories with me. I wish that all your dreams may come true.

A special thanks goes to my fellow Burundian travellers: Myrna, Juliette & Floor. I thoroughly enjoyed our moments together where we could share our ups and downs, give each other advice, laugh about the unfamiliar situations in which we found ourselves, and discover the country. Without you the Burundian experience would have been much more difficult.

I would also like to express my gratitude to Denise, my translator, and Léonidas, my colleague from the University of Burundi. Thank you for showing me around Gitega, introducing me to the schools, teachers and pupils, answer all my questions, and help me to bring about this research.

My big gratitude goes to my supervisor Graciela Paillet. Thank you very much for your enthusiasm, inspiration, patience, stimulating suggestions and encouraging feedback. I would also like to thank Jacobijn Olthoff for taking up the role as second reader.

Furthermore, I need to thank my lovely friends: Robin, Anke, Susan, Yukio, Hester, Nadine, and my boyfriend Stephan. Even though I may not always be around in the Netherlands, I know that you will be there for me whenever I need it. Thanks a lot for staying in touch during my travels, but also for the wine and dinners when we were together.

Last but definitely not least, my family deserves a special word of thanks. My dad Gerrit, my brother Gertjan and my sister Carolien. I am grateful for having such a wonderful and loving family who is always there for each other, even though we are often far away. Thank you also for your understanding and support throughout my academic career, travels and the writing process.

(5)

5

Abstract

Burundi is a country emerging from a long history of cyclical violence. Although the country’s political situation was relatively stable since 2005, tensions raised again in 2015 towards the national elections. Because there have been outbreaks of violence on a generational basis, it is important to work with children in order to ensure a peaceful future. This study focuses on girls, as it is said that they are more often impacted by violence than boys, while it is at the same time recognized that girls and women are highly valued in Burundian society. Based on the premises that education is a basic human right and that schools are important socializing institutions, this study focuses on school-related violence that primary-aged schoolgirls experience in Gitega, Burundi. Although girls are sometimes depicted as helpless victims, they can also be active agents who cope with violence in ways which can be both within and outside the limits of accepted feminine behaviour. This study also focuses on the ways how girls are able to cope with violence they experience, and how these coping strategies can challenge normative femininity.

Based on observations, in-depth interviews and focus groups with girls, boys, teachers and parents in Gitega, this study found that there are many different forms of violence that girls experience which can be both caused in school (e.g. corporal punishment) and caused by factors around schools (e.g. poverty). Moreover, it is found that violence is often gender-specific as girls experience these violations in a different way than boys, which is related to normative femininity. Although girls opt diverse coping strategies – such as asking help from others, praying to God, or just bearing the situation – this study found that how girls cope with violence is mostly within the limits of normative femininity. Therefore, it is argued that coping strategies do not challenge normative femininity very much, because girls do first of all not fail to live up to expectations of normative femininity, and because normative femininity is embedded in broader political, economic and social structures which make it difficult to become aware of and adopt other – rebellious – forms of femininity.

Key words: violence, femininity, coping strategies, girls, Burundi.

(6)

6

List of figures

Maps

Map 1 – Map of Burundi 20

Map 2 – Map of the city of Gitega, including the research locations 30 Figures

Figure 1 – Conceptual scheme 19

Photos

Photo 1 – A school desk at EP Mushasha 1 9

Photo 2 – The road towards EP Nyakibingo 12

Photo 3 – Cultural heritage: les Tambours du Gishora 23

Photo 4 – Road towards the central market in Gitega 24

Photo 5 – A school class at EP Mushasha 1 35

Photo 6 – Impression of a focus group 28

Photo 7 – View on the big sports- and playfield at EP Mushasha 1 31

Photo 8 – EP Mugutu 33

Photo 9 – A school class at EP Nyakibingo 33

Photo 10 & 11 – Impressions of the central market in Gitega 36

Photo 12 – An unfinished classroom at EP Nyakibingo 41

Photo 13 – Women selling oil on the side of the road 51

Photo 14 – “Mind the gap” 57

Photo 15 – Pupils at EP Mushasha 1 59

Photo 16 – Cathedral Christ the King in Mushasha 63

Photo 17 – Pupils at EP Mugutu 72

Tables

Table 1 – List of themes, number of participants and location of focus groups 27

Table 2 – Absence of pupils from grade 6A in 2014 48

Boxes

Box 1 – Going shopping 36

Box 2 – A happy teacher 39

Box 3 – An angry teacher 39

Box 4 – Conversation with the headmaster 40

Box 5 – Diane 43

(9)

9

List of Abbreviations

ADDF l’Association de Défense des Droits de la Femme (Association for the Defense of

Women’s Rights)

CE Civic Education

EAC East African Community EP École Primaire (primary school)

GDP Gross Domestic Product

NGO Non-governmental Organization

PBEA Peacebuilding, Education and Advocacy Programme

SOJPAE Solidarité de la Jeunesse Chrétienne pour la Paix et l’Enfance (Christian Youth

Solidarity for Peace and Childhood)

UN United Nations

UNDP United Nations Development Programme

UNESCO United Nations Educational, Scientific and Cultural Organization UNICEF United Nations Children’s Fund

(10)

10

1 Introduction

1.1 Framing the issue – Girls’ behaviour in a violent environment

Burundi is a small country in East Africa and mainly known for its conflictual past. The country has experienced several outbreaks of violence since its independence in 1962, ending with a 12-year civil war from 1993 until 2005. After the war the country had to deal with challenges such as rebuilding infrastructure and institutions, but the conflict also left marks on the lives of Burundians: people had lost jobs, years of education, possessions, and – most importantly – they had lost family members. It is estimated that more than 300,000 people died and over 50 percent of the population had been displaced during the civil war (Falch, 2008; Verwimp & van Bavel, 2014: 390-391).

Berckmoes (2013) states that the different factors that have led to conflict in the past – such as poverty, the limited availability of resources, the high population density, and the weak government – are still pervasive. Uvin (2009) argues that Burundians face massive difficulties, and Sommers (2013) even states that since Uvin’s research “a lot of worrying trends appear have become much worse” and the hope and optimism that the population felt after the 2005 elections is diminishing ever since (Sommers, 2013: 16). Just after I left Burundi in April 2015 the situation became worse due to political unrest preceding disputed national elections, leading dozens of people dead and wounded and hundreds of thousands of Burundians to flee the country (Falisse, 2015; Nkundwa & Rosen, 2015; Essa, 2015). Nordstrom argues that violence and war should never be seen as a passing phenomenon, but are determining factors in shaping the future – even that of future generations (Nordstrom, 2004: 59-60). Because there have been outbreaks of violence in Burundi on a generational basis and the country has a large youth population1_{, it is therefore important to work with children on the prevention of} violence in order to ensure a peaceful future.

This study is part of the UNICEF Peacebuilding, Education & Advocacy (PBEA) Programme, which aims to build resilience, social cohesion and human security through strengthened policies and practices in primary education (UNICEF, 2013b). Education is considered essential for the development of a country, and the school environment is one of the most important places where children are being socialized. However, it is also stated that “where the social and physical environment of the community is hostile, the school environment is unlikely to be spared” (Pinheiro, in: Antonowicz, 2010: 13) and it is a place where “the power relationships, domination and discrimination practices of the community and wider society are reflected” (Antonowicz, 2010: 4). This study therefore focuses on violence in and around schools in Burundi.

(11)

11

In the 1990s a growing body of literature emerged on school-related violence against girls in developing countries and, as a result, gender equality became an important focus in many agreements and policies.2_{It is argued that in Burundi violence against women and girls is still widespread} (Berckmoes, 2013; Sommers, 2013; UN CEDAW, 2007; UNICEF, 2013a; MPCD & UNDP, 2011). This is due to the male-dominated, patriarchal structures in which women and girls often have an inferior status in society and men have more power and responsibility (Berckmoes, 2013: 22). Concerning education in Burundi, Verwimp and van Bavel (2005: 385) state that although an equal number of girls and boys are enrolled in schools, girls suffer a general schooling disadvantage. This disadvantage is not only caused by factors that are created in schools, but also due to factors outside of school such as poverty. Uvin (2009) nevertheless, states that gender roles are changing slowly and that girls’ education in Burundi is nowadays valued just as much as boys’ education (Uvin, 2009: 137-138).

In the literature on violence against girls in Sub-Saharan Africa, girls and women are often viewed as passive, obedient, and therefore victimized by violence, while boys are depicted as aggressive, fearsome, and perpetrators of violence (Greig, 2009). However, Cobbett (2014) states that this view of girls as either ‘victims’ – highlighting their passivity and vulnerability – or ‘heroines’ – highlighting the potential of girls to contribute to society –, has negative implications for development practice (Cobbett, 2014: 310-311). Several academics argue for a more nuanced view of how girls cope with violence, like Leach and Humphreys (2007), who state that it is critical to acknowledge female agency and complicity in the gendered landscape (Leach & Humphreys, 2007). However, the ways of coping can be limited due to norms and values of accepted, normative, feminine behaviour. On the other hand, though, can girls exceed these norms and values while coping with violence.

In the dynamic Burundian context, where violence is ingrained in the everyday lives of girls while femininity can be dynamic, it is therefore important to gain a better understanding of the diverse experiences, perceptions and choices of girls concerning violence, femininity and coping strategies. The outcomes of this thesis will contribute to the limited academic work on children and violence (Korbin, 2003) and femininities in sub-Saharan Africa, and especially in Burundi. Furthermore, I hope that the outcomes will contribute to development practice, which can help organizations like UNICEF in shaping and implementing policies.

2_{For example, the Millennium Development Goals (MDG) incorporated gender equality in education in their second}

goal (‘achieve universal and primary education’) and third goal (‘promote gender equality and empower women’) (United Nations, n.d.). The Education for All Goals emphasize in three out of six goals the importance to especially focus on girls and women (UNESCO, 2000.). Many organizations take these global goals into account in their own goals.

(12)

12

1.2 Research questions

This study aims to foster more understanding of forms of school-related violence that Burundian girls encounter in primary schools, and how girls’ ways of coping with violence are constructed. The main question leading this study is:

What coping strategies do primary-aged schoolgirls in Gitega, Burundi, choose when facing violence in and around schools, and does this behaviour challenge normative femininity in the country?

The following sub-questions will be used to answer the main question: a) What violence in and around schools do girls experience? b) How is normative femininity defined in Burundi?

c) How do girls cope with the forms of violence they experience?

1.3 Outline of the chapters

This study is organized into eight chapters. The following chapter presents the theoretical framework, where the three main concepts of this study – violence, coping strategies, and femininity – are presented. In chapter three more information is given about Burundi, Gitega, and the three schools where this research has been conducted. Chapter four explains the research methodology, methods, and unit of analysis, and it also encompasses the limitations and ethical considerations that should be kept in mind while reading the results of this study. The three chapters thereafter show the research findings, focusing on school-related violence that girls encounter (chapter 5), ideas about normative femininity (chapter 6), coping strategies they use (chapter 7). The final chapter aims to answer the research questions and discusses these findings, followed by recommendations for further research and policy-makers.

(13)

13

2 Theoretical framework

This chapter explains the three concepts on which this study is based – violence, femininity, and coping strategies. The first paragraph provides an analytical framework of violence which focuses on the intertwined relation of violence and education. The second paragraph concerns femininity and helps to understand how coping strategies of girls can be influenced. The final paragraph focuses on coping, which is necessary to understand the different ways of how girls deal with violence. After framing these concepts the conceptual scheme is provided which shows the concepts and their interrelations.

2.1 Violence and education

Violence is a broad concept that can be analysed from different angles, hence perceptions and experiences of it may differ per person, society or culture (Mider, 2013: 703). Nordstrom (2004) argues that it is difficult to give one definition of violence, because it is embedded in (personal) experiences of the past, present and future which are conceived differently by different people (Nordstrom, 2004: 59). Salmi states that “most people think of violence in a narrow context, equating it with images of war, murders, or riots” (Salmi, 2008: 2). However, as defined by Galtung (1990), these examples are only direct and visible forms of violence. He distinguishes between direct, structural and cultural violence, which are interrelated and therefore called the ‘triangle of violence’. Direct violence is a visible event, which is caused by less visible underlying beliefs which are defined as cultural and structural violence (Galtung, 1990: 292-293). Structural violence is a process of “unequal exchange” which exists when there are persons or groups structurally exposed to violence (ibid.: 293). Cultural violence consists of prevailing attitudes and beliefs that are aspects of culture which cause violence, such as heroism, patriotism or patriarchy (Galtung, 2004). It is an invariant or permanence which makes direct and structural violence look or feel like it is not wrong (Galtung, 1990: 291).

Salmi built further on Galtung’s triangle and describes violence as “a multifaceted phenomenon associated with specific causes and responsible people or institutions” (Salmi, 2008: 18). He created an analytical framework that can be used to analyse patterns and relationships between these forms of violence and prevailing economic, social, and political power structures at different levels in society (ibid.). He applies the framework to education, as – although schools might not look like violent places – schools can be both victims and perpetrators of violence, and they can protect or fail to protect against violence (Bush and Saltarelli, 2000; Jackson, 2000; Salmi, 2006; Harber, 2007; Novelli & Lopes Cardozo, 2008). Based on the premise that education is a basic human right, Salmi classifies categories and forms of education-related violence which are not only caused by internal issues such as harmful pedagogical practices – which Galtung refers to as direct violence – , but also by external issues such as poverty and gender inequality – which Galtung refers to as structural and cultural violence.

(14)

14

The first category of Salmi’s framework is direct violence, which refers to physical and often aggressive or coercive acts – such as corporal punishment or verbal harassment – that can result in “the deliberate injury to the integrity of human life” (Salmi, 2006: 10). Indirect violence refers to harmful situations which indirectly violate the right to survival (ibid.). Salmi argues that this is mostly linked to illiteracy, which is a strong factor of poverty and has a big impact on the lives and future of children (ibid.: 17). Indirect violence is divided into violence by omission – lack of protection against violence – and mediated violence – results of human interventions whose harmful effects are felt in indirect or delayed ways (ibid.: 3). The third category is repressive violence, which refers to the deprivation of fundamental human rights and occurs when there are, for example, strong power structures in schools. The final category is alienating violence, which refers to deprivation of higher rights, whereby children in schools can for example be inhibited to speak their own language or where gender- or ethnic groups are being discriminated within the curriculum (ibid.).

Education-related violence can have a big impact on children’s lives, as it can affect “a child’s emotional and cognitive development and his/her ability to flourish and positively contribute to personal, family and wider community life” (Antonowicz, 2010: 6). When perpetrated in schools, it can lead to dropping-out of school, bad results, poor mental or physical health, a weakening of the social cohesion in the school and wider community, and it can furthermore affect gender (in)equality, economic opportunities, political stability and human development levels (Harber, 2007; Baghdadli, Harborne & Rajadel, 2008). UNICEF states that “denying children access to quality education increases their vulnerability to abuse, exploitation and disease. Girls, more than boys, are at greater risk of abuse when they are not in school” (UNICEF, n.d.). Particularly among girls, drop out- and out-of-school rates are generally higher, which is also the case in Burundi (UNESCO, n.d.). Illiteracy and poverty is also more widespread among Burundian women than among men (EAC, 2009: 9-10). As women “are regarded as vital stakeholders of national life”, educational deprivation for girls and women can have huge impacts on future generations (MPCD & UNDP, 2011; Uvin, 2009).

2.2 Femininity

The second theoretical concept of this study is femininity. This concept is important for this study as ideas of normative femininity can influence coping strategies that girls adopt when facing violence. When thinking about what constitutes a girl or woman people often refer first to the anatomical configuration of women – also called biological sex. However, girls and women are not only identified by their physical anatomy, but also by their (expected) behaviour and roles they take in society. This is referred to as gender, which “builds on biological sex to give meaning to sex differences” (Johnson & Repta, 2012: 21). Femininity and masculinity are thus both part of gender. Paechter (2006: 5) states

(15)

15

that femininity is understood as “ways of ‘doing girl or woman’”, but as there is not one fixed idea of what girls do it is very difficult to define femininity. Moreover, it is also possible that a female person feels or acts masculine, while a male feels or acts feminine.

This fluid character of femininity comes forward in the work of Butler (1988; 2010). Focusing on a person’s agency, she states that gender is always performed, and that the external features of femininity (the way someone dresses or acts) are not fixed but can be chosen: “as a public action […] gender is not a radical choice or project that reflects a merely individual choice, but neither is it imposed or inscribed upon the individual […]” (Butler, 1988: 526). This choice can vary according to the situation in which someone finds itself. Therefore, instead of talking about femininity, it is argued that we should talk about multiple femininities (Behnke & Meuser, 2001: 157; Paechter, 2006: 20).

Paechter (2003) follows up on Butler, arguing that more emphasis should be given to communities of space “as a way of thinking about the formation and perpetuation of localised masculinities and femininities” (Paechter, 2003: 70). She argues that how someone understands itself as feminine is influenced by what she learns from her surroundings, and therefore this varies according to time, place and circumstances (Paechter, 2006: 20). From birth, most boys and girls are socialized in a way that meets the standards and dominant definitions of masculinity and femininity within their society (Nibafasha, 2013: 93). Not only the family, but also the media and religious, medical, political and educational institutions are important actors which teach or expect certain behaviour that girls adopt (Johnson & Repta, 2012; Leach, 2003; Dunne, Humphreys & Leach, 2003; Stets & Burke, 2000). There are thus a set of dominant ideas of femininity, but because it is socially constructed they can differ per country, society, family, or person – and even over time.

I agree with both Butler and Paechter that gendered identities of femininities can be influenced by someone’s surrounding and by her own agency. As mentioned in the introduction of this thesis, coping strategies can challenge norms and values of femininity if this coping behaviour does not confirm to normative feminine behaviour, which refers to a person’s agency and is called ‘rebellious femininity’ (Jaji, 2015). To understand this challenging and rebellious character of coping, we need to understand what the collective norms and values of femininity are – referred to as ‘normative femininity’ (ibid.). “Normative femininity refers to the culturally sanctioned representation of womanhood that is generally accepted as the ideal” (ibid.: 497).

Paechter (2006) states that many theorists who talk about collective ideas of the ideal-typical construction of masculinity refer to ‘hegemonic’ or ‘dominant’ masculinities (Sostakova, 2014; Connell, 1987), but this is not possible for femininities as, Paechter argues, there is no hegemonic femininity (Paechter, 2006: 5-8):

(16)

16

“Masculinity and femininity are not just constructed in relation to each other; their relation is dualistic. Femininity is, thus, defined as a lack, an absence of masculinity. There can be no hegemonic femininity, because being in a hegemonic position is also about being in a position of power; it is about being able to construct the world for oneself and others so that one’s power is unchallenged and taken (more or less) for granted as part of the other things.” Femininity and masculinity are thus constructed through their differences to each other, whereby femininity – both rebellious and normative – is seen as dependent on masculinity. Johnson and Repta (2000: 27) state that this dualistic relation is an important aspect of gender hegemony. I believe that this idea can also be applied to this study because Burundi is a patriarchal society. Chapter 3 will elaborate more on Burundi and its’ gender relations and associated masculinity and femininity in Burundi.

2.3 Coping strategies

The final concept of this study is coping. One of the most common definitions of coping comes from Folkman and Lazarus (2013: 313), who define coping strategies as “cognitive and behavioural processes that are invoked to reduce anxiety and other distressing emotion states in order to enhance or change these situations”. Skinner et al. (2003) state that “coping researchers argue that how people deal with stress can reduce or amplify the effects of adverse life events and conditions, not just on emotional distress and short-term functioning, but also in the long-term, on the development of physical and mental health and disorder” (Skinner et al. 2003: 216). However, it is difficult and not my intention to have an extensive research into the deeper effects of responses as violence, such as lowering stress, but I am interested to explore the everyday coping strategies girls use when responding to violence and how these ways are related to normative femininity. Furthermore, due to the limited amount of time wherein this research has been conducted, it is difficult to research if coping strategies enhance or change adverse situations as certain coping strategies can look like they lead to positive outcomes, but in the long term they lead to more negative situations. As coping in the above definition of Folkman and Lazarus relates to “cognitive and behavioural processes” by which girls respond to violence, I therefore believe that coping is a useful concept for this study in order to identify ways of responding to violence.

This study uses the distinction between problem-focused and emotion-focused coping strategies which has been first identified by Folkman and Lazarus in 1980. Problem-focused strategies are direct attempts to alter stressful situations, such as looking for social support from others or taking control. They actively try to improve the situation itself. Emotion-focused strategies aim to regulate

(17)

17

distress and controlling the individual’s emotions, and can imply avoiding situations or forgiving people (Folkman & Lazarus, 2013: 315-316). Both problem-focused and emotion-focused coping strategies can occur at the same time. This classification is found useful because it views people as active agents who try to face a certain problem when possible.

This study assumes that the choice of coping strategies that girls adopt can be influenced by both normative femininity as explained in the former paragraph, and girls’ own agency. Agency refers to the capacity of a person to act independently (Eriksen, 2001: 48-49). The debate over primacy of structure or agency has increased in the past decades and has even “come to be seen as the basic issue in modern social theory” (Archer, 1988 in: Ritzer, 2011: 520), because it questions how the social world is made of and what determines people’s behaviour. In this study coping does not necessarily mean that individuals can change certain structures, but that they have “the capacity to make purposive choice”, even if these choices can be limited due to structural constraints (Moncrieffe, 2009: 4). These structural constraints are, for example, certain norms and values that people have to deal with, e.g. children will not use physical violence against someone because they will be punished by their parents, therefore they can use other – less visible – forms of violence such as discrimination.

I am aware that the theory of coping strategies comes from the field of Western psychology, and its use in the field of development has been criticized (Lopes Cardozo, 2015), because in its original form it mainly focuses on the individual and his/her abilities (Lemay, 2005: 13). It is argued that coping strategies should not be the main focal point in development practice as coping strategies are not sufficient to overcome structural violations like poverty, corruption and violence. As lopes Cardozo (2015) argues:

“[…] an emphasis on resilience might lead development actors to prioritize strategies that focus more on immediate coping with adversity, rather than the (often longer-term)

addressing of inequalities and injustice in order to transform the structural conditions that

generate conflict in the first place. A possible concern is that a burden of responsibility is placed on individuals and communities, rather than governments and institutions as duty bearers.” (Lopes Cardozo, 2015: 5)

I believe that coping strategies can change or enhance certain violent situations, but it does not necessarily do so as wider structural condition are very difficult to change due to solely individual actions.

As mentioned in the introduction there is a growing body of literature which sees girls and women as active agents who also use violence as a coping strategy, instead of viewing them as passive and obedient victims of violence (Lopes Cardozo, et al., 2015: 18-19). Bhana (2008) conducted

(18)

18

research with girls in South Africa, and concluded that violence is also conducted by girls. As violence is present in their everyday lives, they “avoid it, negotiate it and indulge in it”, but just like Leach and Humphreys (2007) she states that it is mostly within the context which sets limits to their identities as girls (Bhana, 2008: 408). It needs to be emphasized though that someone who experiences violence will not necessarily adopt violent behaviour as a coping strategy. Korbin (2003) states that a part of the literature on violent children argues that children’s aggressive behaviour is socialized or viewed as normal child behaviour, but “children’s own view are rarely sought” (Korbin, 2003: 437-438). There is not enough understanding of children’s own experience of violence and how children’s own agency can lead to violent behaviour as well. This corresponds to Moncrieffe (2009) who argues that the concept intergenerational transmissions (IGTs) – which refers to the transmissions of for example poverty or behaviour from one generation to the other – can convey a path of dependence, while in reality it is more complex. It is argued that children are influenced by norms, values, attitudes and behaviours from other actors around them, but that children also have the capacity to influence others and that their agency is cultivated (Moncrieffe, 2009: 7; Tadros, 2009).

So far there has not been conducted much research on girls’ coping strategies in adverse situations in Burundi. As the data of this research will be shared with the PBEA programme which aims to address violence, it is important to explore girls’ ways of coping with violence and their views on how violence should be addressed in order to better adapt national and international support to the specific needs of children.

(19)

19

2.4 Conceptual scheme

FIGURE 1-CONCEPTUAL SCHEME.SOURCE: AUTHOR

The conceptual scheme in figure 1 shows the most important concepts of this study and their interrelations. The centre point of this study are girls, who are situated within schools where violence is present. However, violence is also present outside schools and schools can also have non-violent spaces or foster peace. Girls are influenced (the uninterrupted arrow) by ideas of normative femininity, which are produced in society and hence also in school, though girls also play a role in shaping notions of normative femininity. Normative femininity can influence girls’ behaviour and therefore their coping strategies. However, coping strategies do not always meet the standards of normative femininity, which is referred to as rebellious femininity. The interrupted arrow shows that coping strategies can also influence violence (diminish or aggravate), but does not necessarily do so.

(20)

20

3 Research context

The fieldwork for this study has been conducted from 18 January until 1 April 2015, in schools in and around the city of Gitega. This chapter will first give an overview of the country, followed by more information about the municipality of Gitega.

MAP 1-MAP OF BURUNDI.SOURCE:CIAFACTBOOK BURUNDI

3.1 Burundi

Economy and politics

Burundi is a small landlocked country in Eastern Africa, bordering Rwanda, Tanzania, and the Democratic Republic of the Congo. The country is ranked in place 180 in the Human Development Index – among the ten lowest rated countries in the world – and has one of the lowest per capita GDPs in the world (UNDP, 2014). Sommers states that Burundi “continues to be plagued by punishing poverty and bleak development” (Sommers, 2013: 11). Concerning food security and undernourishment the country has been classified as “extremely alarming” (Von Grebmer et al., 2014). Furthermore, Burundi has a “poorly diversified economy” that has to deal with low investment rates, high production costs, low-skilled labour, climate change, and an unattractive business environment where corruption is widespread (Tokindang & Gbetnkom, 2014: 3; Uvin, 2009; Sommers, 2013).3_{The majority of economic activities are found in the agricultural sector. Although numerous}

3_{Informal conversation, 10/02/2015; interview with Bella (14), 23/02/2015, Mugutu; interview with Janine (10),}

(21)

21

hotels can be found in Burundi, the tourist industry is very small and most of the hotels are occupied by people from NGO’s, on who the country relies much for development aid.

As mentioned in the introduction, the country has experienced violence since independence, mostly due to unstable Burundian politics and internal, mostly ethnic (Hutu-Tutsi) divisions, which led to multiple coups and violent outbreaks (Uvin, 2009).4_{The Arusha peace agreements were signed in} 2000 and after the first ‘post-conflict’ elections in 2005 the country was relatively stable, though tensions raised from time to time. Despite the population’s fear5_{a tumultuous period broke out just} after I left Burundi, as demonstrations and a coup d’état against President Pierre Nkurunziza’s election bid for a third term occurred. The use of violence was not avoided by both protesters and the police, and dozens of people have been wounded and killed (Human Rights Watch, n.d.). At the moment of finalizing this thesis Nkurunziza has been re-elected despite national and international opposition. Numbers keep increasing, but more than 100,000 Burundians certainly fled the country (EU, 2015).6 It is noteworthy that – although some say that ethnicity does not play a role anymore in Burundi7_– the majority of those refugees are Tutsi (Buchanan, 2015). Furthermore, the constitution of Burundi mandates that “The National Assembly must be composed of 60% Hutu and 40% Tutsi MPs, with a minimum of 30% women and three representatives of the Twa ethnic group” (Sommers, 2013: 10). Ethnicity therefore still remains a relevant issue.

Education

Burundi has taken action on improving access to schooling and the quality of teaching, however, educational quality still remains low in public schools (Tokindang & Gbetnkom, 2014: 10). Due to the civil war and enduring poverty there are many challenges that schools have to face, such as a lack of instructional materials and resources, poor living conditions of teachers, and student-related challenges such as overcrowding of classrooms (Ndura-Ouédraogo, 2008: 45). Also, although primary schools are free, many schools charge additional fees (e.g. for examination paper or the building of new classrooms) which makes access more difficult for children from poor families (Sommers, 2013: 20). In the light of this research, which focuses on girls, it is noteworthy the ratio between boys and

4_{See also Daley (2007) for a detailed account of the continuities and transformations of violence in Burundi.}

5_{Annabella said, “We are now going to towards the elections. I would like them to be peaceful and that we get good}

leaders. I am afraid that there may be troubles during the elections.” (Interview with Annabella (12), 23/03/2015, Mugutu)

6_{The number of refugees increases every day, other sources are referring to 200,000 refugees (Spoor, 2015).} 7_{Informal conversation, 10/02/2015.}

(22)

22

girls enrolled in primary schools is almost equal, which might be a result of the government’s efforts in reducing the gender gap (Tokindang & Gbetnkom, 2014: 10-11; UNESCO, n.d.).

Religion

Religion is important in Burundi. It is estimated that 62.1% of the population in Burundi is Catholic8_, 23.9% Protestant, 2.5% Muslim, and the remaining 11.5% of the population has another or unspecified religion (CIA Factbook, n.d.). Especially the Catholic Church has a lot of influence in Burundian society. She is, for example, involved in the management of many primary- and secondary schools which are referred to as schools under catholic convention.9_{Furthermore, the church possesses economic} assets, has a large membership, international connections, institutional resources, and moral authority (Longman, 1998: 54). It is also noteworthy that during the current crises concerning the national elections of 2015 they had much influence on the government (Niyungeko, 2015).

Gender relations and femininity

When researching femininity in Burundi it is important to consider the fact that Burundi is a patriarchal society, meaning that men hold primary power. The domination of men in Burundi stems from before the Colonial period (Nibafasha, 2013: 11). The mwami (king), for example, was only assisted by men, and the Bashingantahe (wise men) acted as local mediators and judges10_{(Uvin, 2009: 62). Even now} the most important decision makers of the country are men. Although Burundian law states that political institutions should consist for 30 percent out of women, this has not yet been achieved and does not seem to be a priority. It is argued that the status of Burundian women reflects the patriarchal organization of society (UN CEDAW, 2007; Guerra, 2013), and Berckmoes (2013) argues that the patriarchal organization is also “generally perceived to support traditional family and gender values in Burundi” (Berckmoes, 2013: 22). She further states that this has an effect on socialization of children, as they learn “that girls and women occupy an inferior role in the household and in society. Masculinity is associated with power and responsibility, both in the home and in the public” (ibid.).

Nibafasha (2013) analysed the ways Kirundi proverbs construct gender identities and behaviour in Burundi, which gives an insight in traditional norms and values about masculinity and femininity. “Masculinity is expressed among other things by physical and moral courage, strength, economic

8_{These percentages are estimations from 2008.}

9_{Schools under catholic convention are part of public education and are administered and managed in accordance with}

the law of public education. Within these schools, the Catholic church is involved in the management and importance is being given to strengthening human and Christian values (BNEC, 2014).

10_{Uvin states that the bashingantahe is still alive, but it functions more haphazardly and in competition with many other}

(23)

23

power, competitiveness, domination, toughness and control of emotions, while femininity is expressed by hard work, humility, kindness, tenderness, patience and perseverance” (Nibafasha, 2013: 189). She states that Burundian culture shape men and women in opposite but complementary ways. Men are most important, and women remain under the authority of the husband. However, without women a household cannot exist and women are therefore also important actors within Burundian society (ibid.). When someone does not conform to masculine or feminine norms and values he or she is considered to be “deviant, rebel or a ‘missed’ male or female [ikirumbirane]”, meaning that someone lacks something in order to be a man or a woman (ibid.: 93-94).

Butler (1988) and Paechter (2003) state that gender is not static but that it is subject to change and constructed in interactions between people and institutions. Uvin (2009) argues that in Burundi “traditional expectations of men and women continue to make up the core of Burundian identity, even among young people. But they are extremely hard to achieve” (Uvin, 2009: 181). He says that this is due to the conflictual past: “after a decade of civil war and many more years of population pressure and economic stagnation, young women and men find it very hard to live up to these [traditional] expectations” (ibid.: 142). Because many women for example lost their husbands, some of them had to take over the roles of the men (ibid.: 158). Wallacher (2012) states that by empowering women, Burundian man will give up one part of their masculine ideal – being ‘the king’ who has power and does what he pleases – but he attains another part of the masculine ideal – being successful, prosperous and trusted (Wallacher, 2012: 45).

(24)

24

3.2 Gitega

This study has been conducted in and around the city of Gitega. The eponymous municipality in which the city is located consists out of 38 collines11_{and covers an area of 315km}2_{(PCDC, 2013: 4). Having a} population of about 150,000 people in 2012, it is the most populated municipality of Burundi (ibid.: 7-8). Gitega urbain is also the second largest city in the country, counting among others a beer brewery, several administrative institutions, a prison, a hospital, a big market, shops, and public services. Nevertheless the majority of economic activities can be found in agriculture (ibid.: 9-10). Just like the rest of Burundi, young people make up the majority of the population with around 65% being younger than age 25 (ibid.: 7-8). Being the former colonial capital of the country and the place where the king lived, Gitega has cultural and historical value. It is furthermore famous for the national museum, art and music (see photo 3).

Gitega counts dozens of primary and secondary schools and a campus of the University of Burundi. Jackson (2000: 6) states that Gitega is one of the favoured provinces concerning educational provision, where access to education is generally good. In 2012-2013 there were 48 primary schools, with around 33,219 enrolled pupils of which 51,6% were girls (PCDC, 2013: 31). A class counts on average 77 pupils with on average 35 pupils per teacher, but the biggest class in the municipality counts 118 pupils (ibid.).12_{The municipality also counts 8 private primary schools, who teach 1860} pupils (of which 51,9% are girls). Moreover, Gitega counts 35 secondary schools where 14,374 pupils are enrolled (ibid.: 32). The PCDC states that the biggest problems that schools encounter are the overcrowded classrooms (especially in rural Gitega), lack of teachers, lack of materials, no hygienic circumstances, an irrelevant curriculum which does not

focus on future

employment, and school facilities which are in poor condition (PCDC, 2013: 34-35).

11_{English translation: hills. These are also the smallest administrative units in Burundi.}

12_{These numbers have decreased as in 2006 the average number of pupils in a grade was 89, with an average of 37}

pupils per teacher (MPDRN, 2006). It should furthermore be noted that there are often 2 teachers in a class, one teaches in the morning and one in the afternoon.

PHOTO 4-ROAD TOWARDS THE CENTRAL MARKET IN GITEGA. SOURCE:FLOOR ZWEERINK

(25)

25

4 Methodological design

This chapter first explains the methodological positioning applied to this study. Thereafter the research methods that have been used during this fieldwork are being presented, followed by a description of the schools where research has been conducted and the unit of analysis. The chapter will be concluded by highlighting limitations and ethical considerations that were encountered in the field that should be considered while reading the results of this study.

PHOTO 5-A SCHOOL CLASS AT EPMUSHASHA 1.SOURCE: AUTHOR

4.1 Methodological positioning

This study takes the epistemological position of interpretivism, which requires an “understanding of the social world through an examination of the interpretation of that world by its participants” and is thus focused on grasping the subjective meaning of social action (Bryman, 2008: 366). The ontological position of this study is constructivism, which asserts that social phenomena and their meanings are continually being accomplished by social actors, that they are produced through social interaction and are in a constant state of revision (ibid: 19). As such, Burundian girls in primary schools construct social phenomena and give meaning to happenings – such as violence and femininity – differently at different points in time. “People are complex and multifaceted at any given time, for any given reason, they may only reveal part of themselves” (O’leary, 2004.: 60) which makes reliability of this study more difficult, but “[…] no one is wrong. It is simply a case of multiple realities – all valid, all true – at least from the perspective of those doing the describing.” (ibid.: 61). This study does therefore not strive to find one single truth about these topics but rather aims to explore the different views of different people. In order to ensure as much thoroughness and rigour as possible I have collected data from

(26)

26

different stakeholders (girls, boys, parents, teachers, organizations) using different methods. Although triangulation of both quantitative and qualitative methods is often recommended in research because it is said to give a more complete and less subjective understanding of a phenomenon being studied, following from the methodological position the chosen methods are qualitative. Qualitative research “usually emphasizes words rather than quantification in the collection and analysis of data” (Bryman, 2008: 366) and aims to find ‘thick descriptions’ of experiences of girls in the context of Gitega. The used qualitative methods – observations, focus groups and interviews – will be further explained in the next paragraph.

4.2 Research methods and techniques

Observations

Observations were a good starting point to grasp the context and get used to everyday life in primary schools in Gitega. I spent approximately 45 hours officially observing classes, but besides that I have spent many more hours in and around the schools. The classroom observations mostly took place in 5th_{and 6}th_{grade at Mugutu and Mushasha. The age of children in these grades vary from 9 till 18 years} old. The choice which grade would be observed was often made at the day itself and depended upon the activities that the children were involved in and if the teacher agreed to have me present in class. Although I tried to engage as much as possible with my surroundings during observations, I was definitely not always participating in ways that could define it as participant observation. Being a white girl that does not speak the local language Kirundi, it was difficult to fully engage with the children and teachers in schools. My appearance caused children and teachers to be distracted from their lessons and from their activities at the playground, and this caused them to behave differently from how they would if I would have not been there which probably influenced the results of this study as well.13

Despite the fact that I was not able to be a full participant observer, observations gave a lot of information about daily practices in and around schools. I mainly focused on teacher-pupil and pupil-pupil interaction, disciplining practices and, insofar I could understand it, the content of the lessons. Moreover, observations were necessary in order to build trust with children and teachers and let them get used to me, which also made focus groups and interviews with them easier.

13_{Especially in the first weeks children immediately gathered around me when I entered the playground. In other cases}

I can only guess that they were probably behaving different because I do not know how they would have behaved if I were not there. For example, one day a teacher suddenly gave an English lesson for only ten minutes whereas that was not scheduled for that day and participants from that grade told me during interviews that they hardly learn any English. I felt that the teacher did that because she knew that I speak English (observation, 17/02/2015, Mugutu).

(27)

27

Focus groups

I conducted seventeen focus groups with girls which were divided into four themes (see Table 1). Every focus group started with an introduction whereby the participants had the opportunity to ask me questions. I felt that the participants thoroughly seemed to enjoy this moment, whereby they curiously asked about my life and that of my translator, but also sometimes

asked for advice or help. For me this felt like an icebreaker, but also a little bit like a form of ‘reciprocity’. According to Sumner and Tribe (2008) there is always a power relation between the researcher and participant. An attempt to resolve that is by having a reciprocal relationship, whereby you give the participants something in return for the information that they give. Although I believe that the reciprocal relationship between me and my participants was never completely equal, I think that openness from me created more trust between us. Besides receiving information from me, another way in which I tried to create a more equal relationship was by giving the participants a small gift - often a pen - after the activity.

The first focus group with girls was an introduction and its main purpose was to find out how a focus group would work out and to get to know more about some girls and their lives. The second focus group was about violence in school, in the neighbourhood, and in the family. Participants wrote down violations that they experienced themselves or that other people they know have experienced. Afterwards the group discussed which violations they find the worst and what could be a solution to solve or a way to prevent this. The third focus group did not have one specific theme, but we discussed several statements and situations about school, family, violence and gender relations. The fourth focus group discussed violence and conflict resolution in relation to gender.

I intended to do many participatory activities to make the focus groups fun and informal, and to involve the participants as active agents rather than treating them as solely objects of the research. Moreover, participatory activities are an interesting method to analyse group dynamics (Leach, 2006: 1135). However, this did not work out as I expected as time wise it was not possible to have a long discussion after an activity because the children had to go back to class. Furthermore, the main purpose of focus groups is to have a dynamic discussion between participants whereby the researcher

Introduction/test 12 girls Mushasha

Violence 12 girls 18 girls 6 girls Mushasha Mugutu Nyakibingo Statements and situations 12 girls 5 girls 12 girls Mushasha Mugutu Nyakibingo Gender, violence and

peacebuilding

12 girls 12 girls

Mushasha Mugutu

Parents 6 women Mushasha

Boys 9 boys Mushasha

Gender equality 14 men, 3 women

American Corner, Gitega

TABLE 1-LIST OF THEMES, NUMBER OF PARTICIPANTS AND LOCATION OF FOCUS GROUPS.

(28)

28

has less control over the process (Bryman, 2008: 475), but it turned out that many participants were very shy and there was not a lot of mutual interaction. Therefore the focus groups turned out to be more like a group interview (see also paragraph 4.5).

As suggested by Morgan it is better to use smaller groups in focus groups, especially when topics are controversial or complex (Morgan, 1998 in: Bryman, 2008: 479). Therefore the groups always consisted out of six girls. Preferably I would have worked with the same groups per school for each different theme, but for practical matters this was not possible. Therefore, I have conducted each focus group with new participants (see paragraph 4.5). The youngest girl who participated was 9 years old, and the oldest 19, and the average age was 13 years old. Most participants were from fifth and sixth grade and came from Mushasha. See also Appendix II a.

Besides focus groups with girls I have conducted one focus group with nine boys at Mushasha. During this activity we discussed several topics that often came up during conversations with girls. Furthermore, I had two extensive conversations with parents at Mushasha.14_{During these} conversations I asked questions about norms and values they give to their children, forms of punishments, and several topics that also came up during the focus groups with girls. Finally, I participated in a discussion group about women’s history and gender (in)equality which was organized by the American Corner in Gitega.15

14_{I preferred to have one focus group with many parents, but unfortunately only few parents showed up at the}

scheduled meeting. Therefore I have asked some female teachers to participate. In total I spoke with six mothers divided over two groups.

15_{This discussion group has not been organized by myself but I was invited by the American Corner. The participants of}

the group were fully aware of my role as a researcher.

(29)

29

Semi-structured interviews

The third method are semi-structured interviews. For this study 26 primary-aged schoolgirls have been interviewed between 9 and 19 years old, of which 11 from Mushasha, 11 from Mugutu, and the remaining four at Nyakibingo. The average age of the participants was 14 years old and most of them were in 6th grade (see Appendix I and II b). I used an interview guide containing questions about their family, upbringing at home, their neighbourhood, the school, their experience of and attitudes towards violence, and their opinions about gender roles and gender (in)equality. This guide was based on the operationalization of my research questions, but also on information that came up during the focus groups. However, depending on the participant and her answers I adjusted, added or removed questions. The duration of the interviews varied from 30 minutes until almost 1 hour and 30 minutes. Furthermore, I have conducted eight interviews with teachers and headmasters of the schools and a couple of people from organizations who are involved in education, women’s rights or child protection (see Appendix I).

4.3 Research locations

During the first week of fieldwork I visited several primary schools that could be useful for this research. Some schools were suggested to me during informal chats and others were chosen in consultation with my colleague from the University of Burundi. I met with the headmasters and informed them about my study, after which every headmaster allowed me to conduct research at their school. For methodological reasons I intended to have a balance between an urban and a rural area as I thought that the socio-economic background of children from those two areas would differ from each other, which would lead to comparing results. The three main schools – one from an urban area and two from a rural area – where this research has been conducted were finally chosen according to practical reasons (reachability) and cooperation of the school staff. The following map (Map 2) shows the location of these schools, and the following sub-paragraphs will give more information about the schools.

(30)

30

MAP 2–MAP OF THE CITY OF GITEGA, INCLUDING THE RESEARCH LOCATIONS.SOURCE: GOOGLE MAPS.

EP Mushasha 1

The first school where I started observations was École Primare (EP) Mushasha 116_{, which is located in} the eponymous neighbourhood in urban Gitega (see Map 2). The neighbourhood of Mushasha was the first to be constructed by the German missionaries, and with its churches, chapels and convents it is still well-known for religious activities. Moreover, the neighbourhood counts numerous schools: five primary schools and at least three secondary schools, both private and public.

Mushasha is a public school under catholic convention, and it is also an école satélite which is open for children with disabilities such as deaf-mute.17_{560 pupils attend this school (meaning an} average of 93 pupils per grade) who are taught by 25 teachers. There is also a preschool where 2 teachers teach sixty children aged 3-6 years. The school is located in a big, old building18_{, which is} shared with another public primary school. On the opposite is a public secondary school located. These schools surround a big shared sports- and play field (see photo 7). Problems that Mushasha face are

16_{From now on referred to as ‘Mushasha’.}

17_{Few information about satellite schools can be found from official institutions. The headmaster of Mushasha}

explained that a satellite school is a school which also includes children with disabilities, but – compared to an inclusive school – does not get any support at all from the government or other organizations (interview with headmaster, 25/02/2015).

18_{The headmaster and teachers from Mushasha all said that they are afraid that because the building is so old and has}

many cracks in the wall, it would one day collapse on them (Interview with teacher, 04/02/2015; interview with teacher, 25/02/2015; interview with headmaster, 25/02/2015).

(31)

31

overcrowded classrooms, lack of materials and lack of a fence.19_{Several participants said that this} school is known for their good teaching practices, and during observations it was notable that this was the only school where they pay attention to subjects such as singing and drawing whereas other schools found those lessons less relevant.20

Because Mushasha is located in the city of Gitega, I assumed that the pupils would live in the city. However, the more focus groups and interviews I conducted, the more I realized that most pupils actually live in the country side. This was confirmed by the headmaster, who also said that the school does not organize any evening classes because pupils are from poor families and they do not have time and/or money to attend these classes.21_{Despite my aim to make a distinction between children} from different socio-economic backgrounds, due to this reason it has been difficult to make a rigid distinction between pupils from urban and rural areas for this study.

EP Mugutu

The second school where I gathered data was the EP Mugutu.22_{The school is located at the} eponymous colline in rural Gitega, about 15 minutes travelling by motor bike from the city of Gitega (see Map 2). Mugutu is a public primary school under convention of the Catholic church. Moreover, the school is one of six schools in Burundi which participates in a pilot project for inclusive education, where children with diverse handicaps are welcome to study. The teachers receive training from the NGO Handicap International as part of this pilot project. There are 543 pupils (on average 91 pupils

19_{Interview with teacher, 04/02/2015; interview with teacher, 25/02/2015; interview with headmaster, 25/02/2015).} 20_{Interview with Susanne (12), 12/03/2015; interview with Majorie (12), 12/03/2015; interview with Marlène (15),}

11/03/2015; observation, 30/01/2015. All at Mushasha.

21_{Interview with headmaster, 25/02/2015.} 22_{From now on referred to as ‘Mugutu’.}

PHOTO 7-VIEW ON THE BIG SPORTS- AND PLAYFIELD AT

EPMUSHASHA 1.SOURCE:

(32)

32

per grade) who are taught by 14 teachers. About 24 of those pupils are children with disabilities. Problems that the school face are overcrowded classrooms, lack of teachers and materials, no fence and no running water.23_{Just like Mushasha, the school was in a state of disrepair, with broken walls} and windows. The school was surrounded by cultivated land and there was no real playground, but just a relatively small open space (see photo 8).

EP Nyakibingo

The third school where I had the opportunity to gather data was EP Nyakibingo. I was invited by the headmistress, however due to practical reasons I visited this school only twice and I have only conducted a couple of focus groups and interviews here. The school is located in a very rural area and the easiest way to get there is by motorbike, which takes about 25 minutes from Gitega centre ville (see map 2). Because I was bound to the schedule of my translator it was not possible to go there for only a few hours and because of the raining season it was also discouraged to go there every day as the roads were not safe due to mud (see photo 2). Nyakibingo was also the biggest school I visited. 732 pupils attend this school (an average of 122 pupils per grade) who are taught by 18 (only female) teachers. Compared to the other schools, the classrooms of Nyakibingo are most shabby and dark. It was the first school where I have seen classrooms without enough desks and where even windows were missing (see photo 9). Finally the school receives rice from a NGO a couple of times a week,

23_{Interview with teacher/school responsible, 25/03/2015.}

PHOTO 8-

EPMUGUTU.

(33)

33

which they distribute among the pupils. Problems that the school faced were a lack of water, latrines and a fence.24

4.4 Unit of Analysis

The unit of analysis of this research follows partly from the request of the PBEA programme, which focuses on children in primary schools in Burundi. Burundian children can attend primary school from 7 to 12 years old. However, due to poverty and the conflictual past many older children are still enrolled. The intention of this research was to focus on 9-16 year old girls who are enrolled in primary schools in Gitega. It turned out however, that classes for observations and participants for focus groups and interviews were mostly chosen by convenience sampling and sometimes purposive sampling.25_{I preferred to have a balance between younger and older girls, but along the way I} experienced that it was generally easier to talk to older girls, therefore the unit of analysis was expanded to 9-19 years old.

4.5 Data analysis

During observations I wrote quick notes in my notebook, which I typed out and organized on my laptop afterwards. If teachers and participants agreed I also made photographs. Interviews and focus groups have been recorded (with agreement from participants) and transcribed afterwards in Word.

24_{Interview with headmistress, 12/02/2015.}

25_{The choice of participants depended upon in which activities the pupils were involved and if the teacher wanted to}

cooperate. If the teacher agreed to cooperate, he or she first asked in front of the class which pupils wanted to participate voluntarily. From those pupils participants were chosen by the researcher or by the teacher.

PHOTO 9-A SCHOOL CLASS AT EP NYAKIBINGO.SOURCE: AUTHOR

(34)

34

Additional data, such as posters made by participants during the first focus groups, have been photographed so that the children could take their own work home. After transcribing the recordings the data was imported into Dedoose, an online software package. In Dedoose they have been organized through thematic analysis which helped identifying categories and themes. This was done while moving back and forth between data and theory, while looking for relations, similarities and dissimilarities.

4.6 Limitations and ethical considerations

The main limitation I encountered in the field concerned language. The official languages in Burundi are French and Kirundi. French is a language which I do not speak fluently. All the interviews with adults, except for one, were conducted by myself in French, but as my French is not strong my sentences were basic and it was difficult to understand all given answers. Therefore, most French translations have been done afterwards with the help of a friend. Nordstrom argues that “local knowledge is crucial to understanding” (Nordstrom, 2004: 15), but because of this language deficit it was difficult to easily receive interesting or important information from informal conversations, which I believe to be a crucial source during research. Pupils, however, mostly speak Kirundi, and during interviews and focus groups my translator Denise therefore translated into English. Though her level of English was not very good26_{, and – especially when there was much to translate or when several} participants spoke at the same time – she was not able to translate everything literally. Apart from these limitations Denise was very valuable, as she was a calm and nice translator who made sure that participants were at ease, who thought with me about which questions I could ask or not, and who explained me many things about Gitega and the Burundian culture.

An ethical consideration, but also limitation, was the fact that I conducted interviews and focus groups during school time, meaning that the participants were taken out of class. The teachers rejected my suggestion to conduct these activities after class or during breaks, because they said that the pupils needed to go home to do chores. Teachers therefore agreed that I conducted interviews and focus groups during class time as long as the pupils would not be involved in important activities, but because I did not want to take too much time from participants the activities could not last longer than 45-60 minutes.

26_{For example, instead of using the word ‘sometimes’ Denise often said ‘most of the time’, and instead of saying ‘what}

do you think about that’ she often said ‘do you think that is good’. These differences are quite crucial during interviews. When I asked Denise or the participant for more clarification of an answer it often became clear if what she translated was literally what I said or not. Furthermore, her English sentences were basic and not always grammatically right.

Girls in Burundi : a study of primary-aged schoolgirls coping with violence