Chapter 3 Research Design and Method

In: Worship, Ritual, and Pentecostal Spirituality-as-Theology
Martina Björkander
Search for other papers by Martina Björkander in
Current site
Google Scholar
Open Access

The previous chapter described the theoretical framework of this thesis: a spirituality approach to Pentecostalism, a practice approach to theology, and a ritual approach to worship, thereby also providing a methodological basis for using empirical research methods in theology. This chapter answers concrete questions about how research was conducted in this particular study. I first delineate the research design, presenting it as a critical case study with two cases and several embedded units within a multi-layered context. I also briefly describe my methods of data collection and analysis. The rest of the chapter goes into greater detail, ending in a summary of the data collected. My hope is that providing a glimpse of the way I worked and the problems I had to solve can inspire others to undertake similar studies in other places. At the end of the chapter I discuss research ethics and my own positionality.

1 Case Study Design

According to Robert K. Yin, a leading scholar in the field of case study research, a case study can be defined in two steps. First,

A case study is an empirical method that

  1. investigates a contemporary phenomenon (the “case”) in depth and within its real-world context, especially when
  2. the boundaries between phenomenon and context may not be clearly evident.1

Secondly, a case study copes with situations where there are many variables and therefore “benefits from the prior development of theoretical propositions to guide the design, data collection, and analysis” and “relies on multiple sources of evidence, with data needing to converge in a triangulating fashion.”2 According to Yin, case studies are most suitable for answering “a ‘how’ or ‘why’ question” about “a set of contemporary events” over which “the researcher has little or no control.”3 The case is bounded in time and space, but the lines between case and context are defined by theoretical deliberations rather than evident as concrete reality.4 In other words, a case study is designed around a specific real-world, contemporary phenomenon set within a theoretical framework, uses the sources and data collection methods that are deemed most suitable to answer the research questions, and aims at creating an in-depth account.

Case studies can be designed as either single-case (one case) or multiple-case (two or more cases), and they may be either holistic (one level of analysis) or embedded (two or more levels). ‘The case’ constitutes the main unit of analysis and any embedded sub-units are selected in order to cast additional light on the phenomenon under study. Cases can be chosen on the basis of various rationales, for example, because they are unusual in that they deviate from theoretical expectations or everyday logic, or the opposite in that they are common or typical to everyday life.5 The critical case is a specific type where a case is selected on the basis of a certain theoretical framework and deemed critical for the development of that framework. According to Yin, such a case study may become “a significant contribution to knowledge and theory building by confirming, challenging, or extending the theory,” and even help “refocus investigations in an entire field.”6 This can also be called an instrumental case, a study that “uses a case to gain more insights into one issue or theme” and “attempts to understand something more general than the case.”7

In multiple-case studies the cases are juxtaposed to create a more solid investigation, although chosen to replicate each other in some way. “For ‘two-case’ case studies, you may have selected both cases at the outset of your case study, anticipating that they will either produce similar findings (a literal replication) or produce contrasting results, but for predictable reasons (a theoretical replication). With more cases, the possibilities for more subtle and varied replications increase.”8 In multiple-case studies, the analysis can be both thematic (considering the themes emerging from each case) and cross-case-based (comparing themes between cases). The report may present each case separately and then the cross-case analysis, or consist of only the cross-case analysis, with each case presented only in short vignettes.9

In case study research, data collection methods vary according to the phenomenon under study, and may include both qualitative and quantitative approaches and textual and non-textual sources. The key point is triangulation, where a combination of data collection techniques paired with suitable methods of analysis for each source create an in-depth account of the specific case in its real-world context.10 Due to the reliance on multiple sources and multiple methods, the case study approach is “interdisciplinary in nature and allows a broad variety of explanations and concepts in handling the cases.”11

1.1 Critical, Multiple-Case, Embedded-Case Study Design

The current study is organized as a critical case study, with a two-case, embedded design. Guided by two main research questions—“How and why do pentecostal churches practice worship as part of their spirituality?” and “How can worship be theologized from a pentecostal perspective?”—it seeks to construct a pentecostal theology of worship, based on empirical data from two concrete congregations and in dialogue with relevant theory from several disciplines. Given that worship is such a complex and rich ritual practice, an interdisciplinary case study design was deemed most appropriate.

The study can be seen as a critical case study for a number of reasons. First, and based on the theoretical framework described in the previous chapter, I see spirituality and theology as intimately connected in the pentecostal-charismatic tradition, and worship as a key ritual practice within it. Therefore it is critical for the development of theory on pentecostal spirituality to conduct an in-depth study of worship, with a special focus on ‘the rite of worship and praise’ as practiced in concrete settings. Second, as highlighted in the Introduction, churches on the African continent represent a vital part of the global pentecostal-charismatic tradition and indeed of contemporary Christianity itself. On the continent, East Africa and Kenya have become a particularly vibrant context in this regard, bursting with churches and worship music. It is thus critical that East African pentecostal churches are taken into account when studying pentecostal spirituality and worship. Third, as described in the next chapter, urban Pentecostalism has a long history and vast influence in Kenya: politically, religiously, and culturally. The capital city, Nairobi, boasts hundreds of charismatic churches, their music and liturgy constituting a main attraction; Mavuno and Woodley, and their respective church families, are influential members of this community. Lastly, it is also critical in the sense of using cases to interpret and discuss larger theoretical issues; in other words, it could just as well be described as an instrumental case study.

There are two main analytic units in this study, that is, two cases: ‘the rite of worship and praise’ in Mavuno Church and in citam Woodley, respectively. The cases are temporally bound to the time of fieldwork, 2013 to 2014, theoretically bound to the corporate ritual practice of worship, and spatially bound to the two churches and their respective church premises. This means that in a technical sense, it is not the congregation or the church that is the ‘case’ but, more specifically, the corporate ritual practice of worship, and only during a limited time and in a limited space. Nevertheless, for convenience, I sometimes call Mavuno and Woodley ‘case churches’ in the report; moreover, when I utilize the present tense in my account, it should be understood as an ‘ethnographic present’, describing the way things were at the time of fieldwork. For while I am aware that ritual practices, musical tastes, and ecclesial organisations change over time, it is not part of this study to track those changes.

Opting for a multiple-case design rather than a single case was a matter of strengthening the empirical base of the theoretical discussions. With two cases rather than one, the chances increased of creating a robust investigation and a more nuanced interpretation. At the same time, I did not want to have more than two cases, which would have risked making the study too shallow as my time in the field was limited. The cases were chosen according to a theoretical replication logic: they were similar in size, situated within the same city, attracted the same middle-class population, and belonged to the same broad Christian tradition. At the same time, they were different in terms of affiliation and membership (targeting different age groups) and in terms of ecclesial background and theology (rooted in different parts of the Christian tradition).

Within each case there are five embedded units: the ritual performance, the church leaders, the music team, the songs, and the congregation.12 Each of these units are part of the ritual practice of worship and each contribute to the whole. Based on my theoretical framework, ritual performance is the central sub-unit and the other ones all relate to that in one way or another. This means that the roles of leaders, music team, and congregation, as well as the lyrical content of songs, are triangulated vis-à-vis the ritual performance, but not vis-à-vis each other (except in a few instances). This also affected the data collection process, as lyrics (and to some extent surveys) were collected based on their connection with the services observed.

The context of the two cases can be described as overlapping layers: global (pentecostal-charismatic Christianity), regional (urban East African Pentecostalism), ecumenical (the Kenyan church landscape), and societal/cultural (Kenyan society and culture). These layers are all shared between the two cases and each context includes ritual practices, theology, spirituality, and music, among other things. Then there are several layers specific to each case: organizational (the church family and the local congregation), liturgical (the ritual life of the local church), and theological (the local theology). Figure 4 illustrates the relationship between the two cases, the embedded units, and the different levels of context.

In order to research the rite of worship and praise in Mavuno and Woodley, I used a combination of data collection techniques: participant observation, fieldnotes, audio- and video-recording, semi-structured interviews, a corpus of lyrics, and research surveys. The first four can be subsumed under the umbrella term ‘fieldwork’, which takes many of its cues from ethnographic research and ritual studies. The corpus compilation came out of a theological interest in the doctrinal content of songs, and the surveys were an attempt to include more of the congregational voice, as well as adding background information on the two churches. The details of all of these techniques will be outlined below.

It is important to point out at this stage that there is no direct link between the data sets produced by each technique and the sub-units described above. For although the sub-unit of ‘songs’ naturally connect to ‘lyrics’, and that of ‘church-leaders’ or ‘music team’ to ‘interviews’, all of them are also part of the ‘ritual performance’ and very much visible in the data sets ‘audio- and video-recordings’ and ‘participant observation’, to mention just a few examples. All the collected data have contributed insights to the two cases and the five units, as well the larger theoretical issues.

In terms of analysis, each data set has required its own approach. Data from participant observation, interviews and fieldnotes were carefully coded according to an iterative-inductive model, looking for both theoretical and emerging themes. The lyrics underwent a theological content analysis that included literary form and function, central motifs, key words, and name, title, and pronoun usage. Survey results were gathered in tables, looking at both total numbers and percentages. After each data set was analysed, a cross-case analysis also took place, comparing the two cases for themes, convergences, and dissimilarities. This book consists to a large degree of this cross-case analysis, but each of the cases is presented separately in Chapter 4 and the Interlude.

The details of each of the data collection techniques and the analysis of each data set are described below; suffice it to say that the overarching ambition for the analysis has been to answer the research question in the best possible way. Simply put, the ‘how’ has been analysed with a ritual studies approach, focusing on ritualization and embodiment, while the ‘why’ has been analysed with a pentecostal spirituality-as-theology approach, as explained in the orthodoxa model. Behind both questions is a practice approach to theology, where theologizing processes and ‘reading’ theology from practice take centre stage.

In sum, the current study is designed as a critical case study, wherein empirical data from two particular cases are used to develop and expand existing theory on the pentecostal theology of worship. The cases were chosen for theoretical (and practical, see below) reasons and data gathered using a range of data collection techniques. In the next section I go into details of each of the stages of research, moving now from research design to research process.

2 The Fieldwork Journey

In the following sections, I describe my research process in some detail, showing how data were collected and analysed. I do so in a fairly personal tone, underlining the reflexive and embodied character of ethnographic field research.13 The description will take the reader through the different phases of my work and the methodological questions connected to each of them. At the end, the type and amount of data are summarized. Once the concrete work of research has been described, I return to the question of reflexivity and my own role as an insider-outsider researcher.

2.1 Choosing a Topic, a Method and a Field

When I embarked on my PhD journey I had for long been frustrated by how little of the lived faith of pentecostal and charismatic Christians is actually spelled out and researched as theology, especially in sub-Saharan Africa where sociologists and anthropologists are much ahead of theologians in the field. It seemed necessary to find new ways of researching African pentecostal theology, as so much of it is expressed in genres other than books or even writing.

My specific interest in musical forms of worship began as a reflection on the importance of songs and music in pentecostal-charismatic churches around the world and especially in Africa. Growing up as a child of pentecostal missionaries to Rwanda, I had literally visited hundreds of pentecostal services and knew how vivid and engaging the music could be. Music has a universal character and can connect people across cultures and generations; at the same time, I knew from cross-cultural experience later in life, that even when you get the sermon translated for you, much of the overall message is lost if you do not understand the songs. I also knew that music and worship could lead to severe conflicts within a congregation. Disputes over the type of music and singing that a service should contain seemed to be about more than just culture or generation. They seemed to carry something deeper. Music and songs seemed to be crucial to people’s spiritual or religious life and in many instances functioned as carriers of theology.

With this in mind I set out to study the practice of worship within pentecostal-charismatic churches in Africa with the aim of saying something about pentecostal theology and ritual. Since I was interested in pentecostal “practices, not texts,”14 I needed a method that allowed me to come close to “the lived and the local.”15 Inspired by ritual studies and ethnography,16 I opted for a field-based method, which proved easier said than done.

The first problem was educational. As a theologian I was trained in text-based hermeneutics, historical methods, biblical languages, and philosophical reasoning, while I had not had any training at all in empirical research methods, fieldwork, or anything close to qualitative research of an ethnographic type. I had to learn everything from scratch and entered a steep learning curve that got even steeper as I came into the field. The second challenge was that of research design and data collection. How should one combine a focus on rituals and theology as part of the same study? Is there a specifically theological way to do field work? What data should be collected and for what reasons? The third problem was that of analysis: how does one analyse material that is so differentiated, taking its complexity into account and yet retaining one’s theological gaze? Undoubtedly, the analysis would be coloured by the theologian’s preference for texts and linguistic and hermeneutical training. At the same time, hopefully, the ethnographic attentiveness to ‘the lived and the local’ would add a distinctly new and intriguing dimension to it.

Although I was quite clear from the start what my topic would be and, later, also the kind of methods I would use, I encountered yet another problem. I needed to choose a specific setting in which to work, a ‘field’. In field research, this choice, of course, is pivotal. The location of both researcher and research field contributes in fundamental ways to the end result of the study. Here I am thinking of geographical, social, religious, denominational, intellectual location, and personal location. A field can be chosen in many ways and its boundaries are very much shaped by the interests of the researcher. Traditionally an ethnographic field would have been, perhaps, a tribal or rural community, whereas today it could just as well be a hospital, a school, or a group of young people, depending on the research interest. Karen O’Reilly also points out that frequently a field has to be chosen on the basis of practicality, and may largely depend on the researcher’s contacts and whether the researcher is welcomed into a specific setting or not.17

The initial topic gave me a general indication (Africa, pentecostal-charismatic churches), while practical and personal reasons led me to Nairobi and Kenya. First of all, I needed to find somewhere my linguistic skills would be sufficient for my purposes, hence my search for a place where there were many English-speaking churches; second, good initial contacts would reduce the time spent just getting in touch with suitable churches; third, I wanted a place where I could safely bring my family, where we could find good accommodation. Thus, since most of my contacts were in Eastern Africa, I checked opportunities in Kenya, South Sudan, Ethiopia, Tanzania, Zimbabwe, and Uganda. The choice fell on Nairobi, and I started to use my network to make contact with urban, English-speaking, pentecostal-charismatic churches.

2.2 Two Phases of Fieldwork and Two Cases

My fieldwork in Nairobi was conducted in two phases, totalling five months over the period of 2013 and 2014. Thus, the total time in the field is limited compared to a classic anthropological or ethnographic study, although not necessarily so for a case-study design where the quantity and depth of collected data is more important than the time spent in the field. My insider-outsider status also helped in this regard; I did not have to start from scratch when I entered the field, but could build on previous knowledge and experience. Mats Alvesson and Kaj Sköldberg point out that although the normal procedure in ethnography is to be submersed in the society or group for a year or more, the time requirements may be considerably shorter for an insider researcher due to his or her acquaintance with the context and general knowledge of the research subject.18 I think it is fair to say that this is the case in my research, for although I am not an insider of the particular groups taking part, I am an insider to the liturgical and theological tradition under study, as explained below.

The initial phase of fieldwork took place in January 2013 in the format of a pilot study;19 considerable time was spent during this period orienting myself among the vast number of pentecostal-charismatic churches in the city. Before I came to Kenya, I had asked people in my network for recommendations, and decided on some of the churches I wanted to visit. I had also tried to contact some of the pastors via e-mail, although with little success. Once in Nairobi, two people acted as my initial sponsors20 or ‘cultural brokers’,21 helping me on the ground, taking me to churches and introducing me to different pastors.

Unfortunately, the selection of churches and research participants was a little less varied than would have been desirable, as several churches belonged to the same two ‘church families’ (citam and Nairobi Chapel group, see Chapter 4).22 At the same time, the fact that people in my network, unbeknown to each other, all recommended them indicated the importance of these two church families within the charismatic community. The first visit was also a little less fruitful than I had hoped since I did not manage to schedule interviews with all the desired leaders; however, I got a picture of Nairobi churches and some very useful contacts.

After the first phase of fieldwork, I went through my data and reflected on which churches to choose and where to go from that point. To keep the study both deep and manageable, I had intended to select a single church and its ‘rite of worship and praise’ as my ‘case’. However, although there were many similarities between the practice of worship in the churches I had visited, there were also some interesting differences, leading me to choose two churches rather than just one. This would make the study clearer and richer as it would allow me to compare them to each other as well as see common patterns. Apart from the broader criteria described above (English-speaking, urban, pentecostal-charismatic), I wanted churches that could be regarded as critical cases in that (1) music and worship had a central place in the service; (2) they were well-known and often mentioned by my initial informants, a criterion underlining their standing within the pentecostal-charismatic community; and (3) they had both similarities and dissimilarities according to the theoretical replication logic described above. In addition, I also wanted churches that (4) welcomed me as a researcher and seemed keen to have me there. Mavuno and Woodley fitted these criteria best.

At the same time, my selection of Mavuno and Woodley was, in some ways, a matter of chance and coincidence. I happened to stumble upon information on the internet regarding Mavuno, which led me to search for an opportunity to visit the church during my pilot trip. Meanwhile, people in my network put me in touch with one of the pastors in Woodley, whom I also decided to visit. I was impressed by Mavuno’s high-quality music and production values, and fascinated by the way they adopted and adapted elements from youth culture into their services. Woodley also had a large and well-organized music ministry, and I was intrigued by the joy and energy that the music team conveyed in the services. In both churches I got a warm and friendly welcome from church leaders who seemed to understand my specific interests and were open to participating in the study. On a personal level, I liked what I experienced and wanted to return.

The second phase of fieldwork took place about a year later, from December 2013 through March 2014, when I collected the bulk of my data, focusing entirely on Mavuno and Woodley. The Senior Pastors of each church—the ‘gate keepers’23—had given their permission for my project and also appointed people to help me get around.

I came to the field with a distinct interest in worship, which I retained throughout my project. Due to the complexity of the subject—with worship a core ritual practice in pentecostal spirituality and yet a theologically laden concept with numerous connotations—I figured that triangulating the data would be advantageous. This would allow me to stay focused on the topic, yet enable me to see it from different angles while gathering rich material. Guided by my research topic, I chose to spend considerable time and energy observing and participating in Sunday services and other church gatherings. During the services I took notes and both videoed and audio-recorded activities. I also interviewed several pastors, members of the worship teams, and a few congregants in both churches, and conducted a survey using distributed questionnaires. I gathered so-called ‘worship sets’ (a list of lyrics to the songs sung in a service) from Woodley, although these proved harder to get from Mavuno where I needed to rely on my own notes and later internet searches. The details of each of these data-collecting techniques are discussed below.

3 The Data-Collection Maze

3.1 Observation and Participation

Given the embodied and enacted character of ritual, a central method for ritual studies is participant observation. Here ritual studies is heavily indebted to ethnography, wherein participant observation is a key feature,24 and uses similar methods, although focusing on particular practices rather than a culture as a whole. “The most direct way of studying a ritual is by attending it, opening up the possibility that all your senses will take in the event,” Grimes says. “Observing is more than watching”; if done well it is “multisensory” and requires “full presence.”25 Thus, the researcher is observing while participating and participating while observing. “It is almost impossible to observe without participating, especially if you observe for long or repeatedly. Researching, you are at first an outsider, but researching attentively bends the body forward … you become something of a participant by listening empathetically, even if you intend later to think critically.”26

Depending on the researcher and the research topic, participant observation can lean more towards observation or more towards participation, although the two are never entirely separate. Spradley developed a continuum model that distinguishes between five levels of participation—non-participation, passive participation, moderate participation, active participation, complete participation27—of which the ‘moderate participation’ type would best describe how I worked. I was present at the scene of the action—whether church service or prayer meeting—observing what was taking place and to some extent interacting with the other participants, yet at the same time I was clearly identifiable as a researcher and my goal was to watch, listen, and take notes, rather than interact with people or engage in the service. However, I did not hang out with members of the two groups on a daily basis and I did not observe them in their everyday lives outside of the church premises. I lived in my own house with my family and I commuted to the two churches for services and gatherings.

While observing I tried to use all my senses and take notes on what could be heard (voices, prayers, music, laughter, handclaps, volume, rhythm, disturbances, etc.); what could be seen (facial expressions, body movements and postures, interaction between people, clothes, screens, decorations, etc.); what could be felt (atmosphere, intensity, joy, grief, engagement, reservation, etc.); and also occasionally what could be smelled or tasted (food, garbage, tea, wine, bread). For me, this meant that I could not also participate in the services in a way that involved my heart and my spirit as I would have done had I come as an ordinary congregant. I did sing along with the songs (while noting down the words and looking around), although I did not involve my whole person in worship. I occasionally prayed and communicated with God during the services, but usually I was more eager to observe what others did when praying and to listen to their words rather than participating in the prayers myself. Thus, I was not fully participating, but at the same time not only observing. I laughed at the jokes, greeted people at the time of greeting, spoke to them before and after the service and participated in communion more or less in the same way as I would have in any service. In fact, the time of communion was the only occasion during services where I felt it was not appropriate to take notes or ‘spy’ on others (either for me or the people around me), although afterwards I made notes on it.

As can be seen from the above, participant observation involves all the senses and is thoroughly personal and embodied. The researcher has to use him or herself in a holistic and often exhausting way. At the same time, participant observation requires some sort of mental distancing, as the researcher must not only feel, hear, and see but also write down and reflect on what is said and done in an orderly manner.28 Participant observation may seem like ‘a paradox’ or ‘an oxymoron’, a method with built-in tension; however, the tension need not be resolved.29 As O’Reilly points out, “the apparent tension between participation (and involvement) and observation (and distance) does not have to be resolved: it is what gives participant observation its strength. Participating enables the strange to become familiar; observing enables the familiar to appear strange.”30 Participating and observing are at the heart of ethnographic fieldwork methods and neither can do without the other or without their constant companion: field notes.

3.2 Field Notes, Audio-Visual Recordings

Writing field notes is an important way of documenting fieldwork, and an integrative activity in participant observation. As Grimes observes, the only way to convey the multisensory experience of ritual, including “smell, taste or touch,” is to “describe, interview, photograph, or video, in effect rendering olfactory, gustatory, and tactile data as audiovisual data.”31 Therefore, taking notes and using audio-visual media is key for documenting, analysing, and presenting ritual studies research.32

Ethnographers advocate a variety of ways to systematize and record field notes, often distinguishing between different degrees of elaboration and different kinds of content: between mental notes, jotted notes, and full field notes, for example, and between descriptive notes, analytic notes, and autobiographical notes, the latter often referred to as a diary or journal.33 The researcher, as the producer of field notes, is inevitably part of the research itself; ethnographers do not just collect, but in a very real sense co-produce their own data, meaning that all field notes are both data and analysis, regardless of the content and style of writing.34

My own notes took several different forms,35 the simplest of which were in the notes app on my iPod: short jottings on something seen or heard in an everyday situation in Nairobi. A more elaborate format was the field diary. My field diary is a personal and reflective one, recording my own experiences and impressions in the field, while at the same time including both analytic and descriptive notes. Once I was back from the field, I continued writing occasional field notes, reflecting on methodological decisions or analytical ideas—something in the nature of what O’Reilly refers to as keeping an “intellectual diary.”36

The most elaborate form of notes that I used were records of observed church services,37 what I called ‘observation notes’. These were mainly of a descriptive character, carefully recording what took place in the service but always in a personal and reflective style, making use of all my senses. The ethnographic vignettes (see Interlude), as well as examples throughout the study, provide a clue to what they looked like. Observation notes were written in several steps. During church services I used an ‘observation guide’ (see Appendix 2) and continuously jotted down detailed notes on whatever took place both on stage and in the hall. The first part of the guide consisted of a one-page form where I noted things that I wanted to remember for each service, including time and location; the number of congregants and the approximate percentage belonging to different age groups and genders; the number, age, and style of singers and musicians; the number of songs in English and Swahili; Scripture verses read or quoted; the theme of the sermon; the number and type of altar-calls, and so on. According to DeWalt and DeWalt, this type of counting or listing of items is an important way to improve the level of description and may serve as a basis for later conclusions.38 The second part of my observation guide consisted of a two-page form for taking running notes in chronological order with the simple headings: When? Who? What? How? I used about 15–20 copies of the second part for each service I observed, writing on both sides. This means that I have approximately 30–40 pages of hand-written notes for each communal gathering that I observed. These notes were later digitalized as full records, with the help of audio and video recordings to jog my memory, producing 15–20 pages with detailed observation notes for each service.

Since my focus of research was the practice of worship in communal gatherings, I documented this very carefully with both elaborate notes, as above, and a combination of video and audio recording. Grimes underlines the importance of the latter, saying, “Fieldwork is greatly enhanced by video. It captures details we don’t see or can’t remember, and it engages the senses … in a way that writing alone usually does not.”39

On a practical level, my recording developed over time. For the pilot trip I used my iPod to record sound from communal gatherings, while during the second phase of field work I had a proper dictation machine. For both trips I used a video camera to record and take photos. During the first phase of fieldwork, I did not videotape whole services, only taking small clips and a couple of photos from each. During the second phase I decided to use a tripod to position my camera, which gave me more freedom to take notes and participate. Mostly I merely turned it on and left it running with the same perspective throughout the service. While this meant that I could not also take photos, I concluded that, of the two, video recordings would be more useful. In Woodley, I had the camera next to me for the first few weeks, but later the media team kindly offered to run it from their balcony at the back of the hall. This had the advantage of allowing me to sit wherever I wanted without the concern of camera positioning. In Mavuno, I kept the camera with me in the hall and struggled to get a good image due to the mass of people around. In both churches, I occasionally put the camera on the platform, facing the congregation, to get a different perspective.

From each service I observed, I have approximately two hours of audio and video recordings. During the pilot trip and at minor gatherings, I was not always allowed to use the camera and so I sometimes only have the audio recording and the notes.

3.3 Interviews

Interviewing is an important research method and one that is used widely across many disciplines. Often a difference is made between structured, semi-structured, and unstructured interviews, depending on the level of control given to the interviewee and interviewer, respectively. O’Reilly says that ethnographers tend to rely on unstructured interviewing, but may employ any of these categories depending on the research topic and the purpose of the interview.40 She further distinguishes between formal interviews conducted at particular times and with a particular purpose and informal interviews and conversations, which take place spontaneously and continually during fieldwork.41 Just as with participant observation and field notes, the researcher herself is intimately and inevitably linked to the data collected from interviews and a co-constructor of that data. His or her personal presence in the situation and chemistry with the interviewee affect the result in a very real way.42

I principally chose formal, semi-structured interviews, although I also had a number of informal, unstructured conversations, which occurred in connection with staff meetings and church gatherings, but also in public places, over a meal, or while commuting from one location to another. They mainly took place with staff members, worship leaders, or pastors in each church; I had virtually no informal conversation with ordinary congregants apart from some small talk going in or out of church (reasons for this are discussed in regard to the role of the researcher below). Informal conversations were documented as part of my field diary, and for the sake of clarity I do not refer to them as ‘interviews’ in the study.

Formal interviews, however, were conducted in a different manner. These were rigorously planned, individually adjusted, carefully documented, audio-taped and later transcribed. Since my research interest focused on worship as a communal, ritual, and spiritual-theological phenomenon, pastors, worship leaders, and worship teams were my key informants. Many were of higher status than myself—well-educated, busy people with weighty responsibilities; I could not risk wasting their time with informal conversations, and I could not count on getting more than one chance to interview them. Since most were also older than I, and male—an important element of status in Kenyan society—I had to ensure I behaved in a serious and respectful manner, using a formal interview style43 and, above all, preparing well. Steinar Kvale and Svend Brinkman note that a key component in interviewing the elite is knowledge. If the researcher can show that she is well-versed in the topic, masters the technical language, and is aware of the status of her interviewee, then a certain balance can be achieved that allows for an interesting exchange despite the asymmetry of power and status.44 In my case, I believe my general acquaintance with the pentecostal-charismatic tradition and my personal background helped me handle these situations. Nevertheless, I experienced a difference between the interviews during the first phase of fieldwork and those of the second. Not only was I much better prepared for the interview situation itself, I had also gained much more specific knowledge about the field, the topic, and my interlocutors, which resulted in engaging conversations and rich material.

I also conducted group interviews with people from the music team in each church and with a youth group in Woodley. Ethnographic group interviews differ from focus group interviews in that they are conducted with naturally occurring groups in the field, rather than groups formed by a sample of people who do not know each other.45 Pranee Liamputtong points out that in collective cultures, group interviews might be more suitable than individual ones as they are closer to the way people relate in ordinary life and generate positive input.46 The three groups I interviewed met, talked about Christian faith, and worshipped together on a regular basis, and their close relationships impacted positively on the conversations, rendering them more informal and personal than the individual sessions. This was especially true of the youth group interview, which started in an orderly fashion but became increasingly heated, while remaining friendly, turning into a lively and completely unorganised discussion over which I was no longer in control. Not that it was in any way unpleasant or irrelevant to my research, rather the opposite, but it was not what I had planned or expected.

Interviews with the two music teams landed somewhere between the very formal and the informal. Both involved rather young people who knew each other well; although they did not have formal positions in the church, their on-stage roles generated a certain level of loyalty to the church and knowledge of the research topic. Conversation with the Mavuno team was engaged and vivid, not as chaotic as with the youth group and yet not as formal as the one-on-one interviews. Almost the whole team took part and it was clear that they also enjoyed the conversation. With the Woodley team the conversation was more formal and less engaged, at least partly for practical reasons (everyone was tired after a long rehearsal session) and partly due to the participants being selected by the leader rather than volunteering. My plan was to do an interview with one of the ‘life-groups’ in Mavuno (small groups that meet on a weekly basis for prayer and Bible study), as a comparison to the youth group in Woodley, but I was unable to attend the planned interview for personal reasons, and it was conducted and partly video-recorded by the life-group leader. In the end, I decided not to include this interview in my analysis because of its limited quality.

During the pilot trip, I used more or less the same list of questions for each interview, although I adjusted them on the spot and asked follow-up questions depending on the answers. For the second phase, I developed a more elaborate ‘interview guide’ of two parts (see Appendix 1). The first part was the same for all interviewees, and contained a number of short questions such as name, personal background, and role in the church. Here they also signed their consent to being part of the study (see below). The second part contained an individualized set of main questions and a number of possible sub-questions, as well as some respectful sentences with which to frame the interview. Some questions were introductory, some direct or specific, and some more probing in character.47 Examples of topics covered include questions related to their personal background and role in the church, church organization, the church scene in Nairobi, and, of course, their views on worship. Depending on their role, I spent more or less time on different topics, but in general I was asking for facts, ideas, concepts, and experiences, more than narration or discourse.48

All of my own interviews were audio-recorded using either the iPod (first phase) or the dictation machine (second phase). In total, I conducted twenty-two formal, semi-structured interviews, each lasting between fifteen minutes and two-and-a-half hours, resulting in approximately twenty hours of interview material. Most second-phase interviews were later transcribed and coded but, from the pilot trip, only those that directly related to Mavuno and Woodley were so treated. For each interview I took hand-written notes and also made reflective notes in my field diary.

3.4 Lyrics

In case study research (as well as in ritual studies and ethnographic research), the idea of triangulation is an important one, as mentioned above. Using a broad set of methods and techniques is a way to paint as rich a portrait as possible of the case, topic, or culture under study. In my case, this meant collecting songs that were used in worship, since this would give me yet another perspective in addition those provided by participant observation and interviews. As was noted in the previous chapter, analysing the content of lyrics is a way to investigate not only the ‘indexical’ meanings in ritual but also the ‘canonical’ ones.49 This, however, raised some methodological challenges, both during fieldwork and at the stage of analysis. I describe these in more detail below in a special excursion into analysing lyrics in search of theology.

The process of collecting lyrics differed slightly between the two churches. Given that my ‘cases’ were the ‘rite of worship and praise’ and my main sub-unit the liturgical performance of worship, I decided to collect lyrics from observed services alone. In Woodley, I was invited to join a Facebook group that the music team used for communication, giving me immediate access to most of the lyrics used. Each week the worship leader posted the worship set for the coming service, which included lyrics for all the songs in their set order, sometimes the chords and names of artists or composers, but not the music sheets. Links to online versions of the songs were also posted in the group, mainly from YouTube, so that the team could prepare beforehand. I downloaded the documents for the Sunday and Wednesday services that I had observed, and followed links to some of the songs (especially those that I did not recognize). Comparing their worship sets to my observation notes, I could see that they generally followed the sets quite strictly, only deviating from them in a few cases. Browsing the internet for the songs that were not on the worship sets, I ultimately compiled a list of lyrics for 54 unique songs that were sung on one or more occasion at Woodley.

In Mavuno, I was promised the slides that displayed the lyrics on big screens during worship, but unfortunately did not receive them. Since I had counted on them, I did not take notes on lyrics, which meant that, once home from the field, I had to spend a lot of time browsing the internet to find the songs. The Mavuno music blog,50 where the music team has put some worship sets, was of no use to me, since it did not include the sets for the services I observed. I later contacted some of the pastors and got some information from them as well, ultimately producing a list of 36 unique songs sung in observed services in Mavuno.

Since both churches are English speaking, most songs are sung in English. But both churches also include Swahili songs in their repertoire, and I have had a total of 20 songs translated for the sake of analysis by Sahaya G. Selvam at the Don Bosco Research Centre in Nairobi.

3.5 Research Surveys

When I planned my fieldwork, I was hoping that I would get a chance to get to know people in the congregations and to interview ordinary congregants and not just leaders. It was, however, much harder than I expected to have anything but short conversations as people came in and out of church, and no one seemed interested in getting to know me as a person, still less inviting me home. I can only guess the reasons: the difference in culture and colour, my own shyness, and my presence as a strange person with a camera might have been part of it, but more than anything I concluded it had to do with the urban atmosphere. People seemed to interact mainly with their own peers or family, and left the mingling with newcomers to the teams appointed for the job. The short period in the field also contributed to my difficulty; I would probably have got to know people had I stayed longer and engaged more in different parts of the ministry. This lack of congregational voices is a real disadvantage in the study and one that was not easily compensated for.

Given the conditions, I needed a new strategy, and I decided to do a poll in each church to get a sense of the congregations’ demographic profile and the way ordinary congregants experience worship. It is not nearly as good as having in-depth interviews, but better than nothing. Again, it is a matter of triangulating data, using a different kind of method for a different perspective on the same subject. The surveys were later used mainly for background information on the churches and to a smaller degree to mirror the ‘congregational voice’ on worship.

My intention was to hand out standardized questionnaires (see Appendix 4) in conjunction with a service in each church, since this would give me comprehensive statistical material on actual church-goers and their on-site experience. This was approved in Woodley, and with the help of a research assistant and the ushers, I distributed questionnaires to people as they went into church and during the offering. After the service, the questionnaires were handed back and I got almost four hundred responses, providing very reliable statistical material. In Mavuno the process looked a little different since the church leadership did not approve of such hand-outs. They were afraid people would feel upset and unwelcome if they were met by a questionnaire as they came to church. I also got the feeling they thought questionnaires were a bit old-fashioned and did not fit their profile; they preferred an online poll and offered to help me execute it. Ultimately, no one helped and I had to rely on my research assistant’s negotiating with one of the staff members to give out the material in life groups and on buses taking people to church, which meant that I lost control over who filled in the questionnaires and in what situation. Therefore, the statistical material from Mavuno is weaker than that from Woodley. For example, it is possible that the representation is a bit lopsided in terms of demographics, with less affluent and younger people riding the bus to church rather than taking their own car. Nevertheless, I managed to get back three hundred responses from people affiliated with Mavuno, although I do not know how representative they are of the actual church-goers or to which service/services their appraisal relates.

The work of organizing, digitalizing, and analysing the surveys (using spss) was performed by a research assistant working for the Don Bosco Research Centre. We handed out more than a thousand questionnaires in total and ended up digitalizing and analysing 330 questionnaires from Woodley and 300 questionnaires from Mavuno. The numbers correspond to the estimated number of church attendants at that particular service for Woodley (30% of 1,000 people) and the estimated number of church attendants during an ordinary Sunday (two services) in Mavuno (15% of 2,000 people).

3.6 Summary of Collected Data

To sum up, during my first phase of fieldwork I observed nine Sunday services and a few other gatherings in a total of twelve Christian communities in Nairobi, including Mavuno and Woodley. I interviewed ten different church leaders, including pastors from citam and the Nairobi Chapel group of churches. In the second phase of fieldwork, I visited Mavuno and Woodley for the main service on six different Sundays. Since they had multiple services with a more or less similar program, this added up to ten unique services in Woodley and eight unique services in Mavuno. In addition, I attended a youth service in each church and number of other gatherings, adding up to a total of thirty-two observed services. During this period, I interviewed six individuals in leadership positions in Mavuno and Woodley and one local pentecostal theologian, and had a large number of informal conversations with leaders and academics in Nairobi. I interviewed the music team in each church and a youth group in Woodley, and conducted surveys collecting more than six hundred responses in total. I collected almost ninety songs from the two churches combined.

In numbers, the fieldwork looks as follows:

  1. I spent a total of 5 months in the field.
  2. I observed and participated in 45 services and church gatherings. Each service lasted between 1–3 hours.
  3. I have approximately 900 pages of fieldnotes, including observation notes, and 90 hours of video and audio recording, respectively.
  4. I interviewed 40 different people in 22 formal interview sessions. This amounts to approximately 20 hours of audio-recorded material, while 15 of the interviews have been transcribed.
  5. I collected the lyrics of 90 songs. I had 20 Swahili songs translated.
  6. I collected and analysed 630 questionnaires.

The study builds on all the collected data although not everything is directly quoted or explicitly referred to in the report. The detailed narrations in the Interlude function as exemplary descriptions of observed services, yet the themes I draw from them were present in other services as well. The same is true of interviews, where some voices are more present than others in the report, while the themes brought forward represent the whole set of interviews.

4 The Quagmire of Analysis

Above, I have described the methods I used in the field in a detailed and transparent way in order for the reader to understand how this particular study was designed and how the different methods relate to the purpose of the study and to the research questions. Each method produces a different set of data, and together they illuminate the complexity and richness of worship as it is practiced in Mavuno and Woodley. In this section I briefly describe how I have worked with analysis, and to what ‘analytic use’ I have put each of the data sets.

My point of departure is the ritual phenomenon that Albrecht refers to as ‘the rite of worship and praise’ and that for the sake of convenience, and in accordance with global pentecostal-charismatic usage, I refer to as ‘worship’. In general, my ambition has been to unpack both the practice and meaning of worship, and to do so from both ritual and theological perspectives (in line with the theoretical framing set out in Chapter 2).

The ritual analysis follows in the footsteps of Ronald Grimes, Catherine Bell, Thomas Csordas, Martin Lindhardt, and Daniel Albrecht,51 and focuses on the ‘performance’ and ‘embodiment’ of worship: that is, what people do in the rite, what they say, how they move, how they dress, what roles they have, what objects they use, and so on. I have looked at how worship is ritualized in terms of processes, structures, and elements of the rite, as well as how participants explain and experience these different aspects. The primary data sources were documented participant observation and interviews, although the surveys and lyrics were allowed to mirror them to some degree. I found video recordings especially helpful for the visual aspects of ritual, and easy to work with since they can be played repeatedly.52 Observation notes and interviews, on the other hand, proved very helpful in informing structural and experiential aspects.

The theological analysis follows in the footsteps of Steven Land, Mark Cartledge, Kenneth Archer, James K.A. Smith, Amos Yong, and Ulrik Josefsson,53 focusing on the relationship between pentecostal theology and spirituality, while at the same being time inspired by Clifton Clarke, Ogbu Kalu, and Kyama Mugambi54 in that it does so in an African context. Concretely, the doctrinal aspects of worship were analysed from lyrics and interviews, while the larger issues of the relationship between worship as a part of the liturgy and worship as a way of life—indeed, between pentecostal theology and spirituality—are at the core of my project and have been active throughout the process of analysis for each of the data sets.

In addition to these two main theoretical clusters, I have made use of theories and insights from ethnomusicology, anthropology, and sociology, depending on what I have seen in the data. In line with the iterative-inductive approach, I have aimed for a creative dialogue between theory, data, and analysis, allowing them to interact at each step of the process and so co-produce the end result.55

Technically speaking, my analytic process could be described as a ‘bricolage56 of analytic techniques suitable for each of the datasets. As mentioned above, the research survey was analysed using spss, and presented in numerical tables.57 Lyrics were compiled into an alphabetic collection of songs used in each church, and a chronological list of songs sung in each service. Then I made tables, using Word, listing such things as themes, metaphors, biblical texts, type of song, and pronoun and person in Godhead. This process is described further below. Lastly, for the interviews, observation notes and other field notes, I used NVivo to code the content according to themes, categories, and concepts, and to make annotations. The coding built on my four theoretical frames—orthopraxis, orthopathos, orthopistis, and orthodoxa—homing in on ritual, somatic-emotional, and doctrinal aspects, and relating the practice and interpretation of worship to the larger picture of pentecostal spirituality-as-theology.

While coding and writing, I occasionally went back to the video and audio recordings when I needed more detail, such as exact quotes or a clearer picture than provided by the observation notes. Since I started this coding rather late in the process—once I had all my data and had begun planning my dissertation—I chose to work closely with theory. The analytic process had naturally begun much earlier when designing the project, deciding on interview questions, creating observation guides, and so on, and is, therefore, at least partly built into the design.58

4.1 Analysing Lyrics in Search of Theology: A Special Excursion

The prospect of analysing lyrics in search of theology raised some methodological challenges for me, at the phase of collection as well as analysis. I have already described the collection phase above; here I explain how I worked with the analysis, since this has implications, especially for the discussion in Chapter 7.

For Woodley, I compiled a list of fifty-four unique songs that were sung during eight services. During my fieldwork, I observed a total of thirteen different services in the church, held on nine different days: ten Sunday services (unique occasions), two Power house (Wednesday evening) and one Family Carols Service (Friday night). Since the program is repeated in the three services each Sunday, I choose to count these as ‘one’ rather than ‘three’ services in my analysis. The Family Carols Service was a special occasion that did not follow the normal rhythm of a Sunday service (as explained in Chapter 5), which is why I chose to not include those songs in the analysis. Therefore, the analysis is based on songs sung in eight different services.59

For Mavuno, I compiled a list of thirty-six unique songs that were sung in six different Sunday services. During fieldwork, I observed a total of fourteen different services in the church, held on eleven different days; eight Sunday services (unique occasions), five staff meetings (Wednesday morning), and one youth service (Sunday). Since the program is repeated twice on each Sunday, these are counted as ‘one’ as above. Due to the more complicated process of finding lyrics for the songs sung in Mavuno, I have not included songs from staff meetings and the youth service; instead, only Sunday worship is considered for this church. Therefore, the analysis is based on six different services from Mavuno.60 One of them was a Christmas Service, referred to as the Alternative Christmas Service, that started off as a normal Sunday service but then included a concert of piano and voice in the middle, before ending with an altar call.61 In that service, a total of twenty-seven songs were sung, but given the fact that half of them were secular songs,62 only the thirteen songs that were functionally sung as worship songs63 were included in the theological analysis.

Thus, the ratio of songs in each service is quite astonishing. In Woodley, an average of 6.75 unique songs were sung in each service. Only two songs were sung on more than one occasion, pointing to the high level of planning and organization surrounding the liturgy and citam’s broad musical repertoire.64 It does not necessarily imply, however, that these songs were all new to the congregation; my impression was rather the opposite. It seemed like the songs were well known in most cases, either because they were taken from the large Christian song tradition with which many congregants are familiar, or because they are often played on Christian radio or distributed via other media.65 English hymns were frequent in the repertoire, as were Swahili choruses, Afro-American Gospel, and contemporary worship music by various performers, such as Reuben Kigame, VaShawn Mitchell, Kurt Carr, and Kabeira & the Klan.

In Mavuno the thirty-five songs are distributed unevenly between the six Sundays, thirteen of them belonging to one occasion as described above. This means that the average number sung at ‘normal’ Sunday services is 4.4 songs, considerably fewer than at Woodley. This is also how worship leaders explained the process to me, that they include four or five songs in each service. Three songs were sung on more than one occasion, the rest are used only once, again pointing to a high level of organization and planning of services. Apart from the Christmas service, where several hymns were included, the repertoire is more streamlined in Mavuno. African, Afro-American, and Australian contemporary worship music dominates, with original performers including Sinach, Gospel Fathers, Kanjii Mbugua, Marvin Sapp, and Tye Tribbett.

Although the repertoires in the two churches differ, there is still considerable overlap when it comes to original artists: Hillsong, Israel Houghton, Hezekiah Walker, Chris Tomlin, Don Moen, and William McDowell are all used frequently in both churches. Five songs occur in the lists of both churches: two hymns, both sung as congregational worship (Away in a Manger, Oh Come All Ye faithful), and two contemporary worship songs (Our God, Withholding Nothing). One song, a hybrid of hymn and contemporary worship, exists in both English and Swahili, (The Solid Rock/Cornerstone/Cha kutumaini sina), and both churches sang versions of it. In Woodley, fifteen out of fifty-four songs were in Swahili, the corresponding number in Mavuno is five out of thirty-five. It is thus a small proportion, but still important.

The sheer volume of songs forced me to choose between selecting just a few for in-depth analysis or taking a helicopter perspective and trying to uncover theological patterns in the whole corpus. Both methods raised questions: in the former, which songs to include or exclude and on what criteria; in the latter, whether or not these 54/35 songs of disparate origin could be treated as a corpus representing the theology of each church, and on what grounds. Ultimately, I chose the second path as I was interested in the overall theological pattern and thought a broader picture would do more justice to the local theology. The difficulty of choosing which songs to analyse when only a very few of them were ever repeated (if frequency were to be a deciding factor) also affected the choice. The question remained, however, of whether the selected corpus represented local theology.

The key here is the context and the particularity of field research. As a theologian working empirically, my focus is on the “lived and the local,”66 especially local practice. My study does not take into account all the worship songs ever sung in citam Woodley and Mavuno Church, only those that I observed during church services; neither does it take into account the origin and use of these songs in other settings. Instead, I see the church services as the context that defines the texts/lyrics and makes them what they are theologically. These specific songs have been chosen by this particular church (according to a fairly standardized process), during a particular time (winter 2013–2014), to serve a specific purpose (honouring God and mediating his presence to the congregation); therefore, they represent the theology of these local congregations, at least in some way, and may be analysed by me as a researcher. Had I come at a different time, I would have collected a different set of songs (for example, fewer Christmas hymns), and probably come to slightly different conclusions. In a similar way as one can speak of ‘Lucan theology’ in exegetics, while knowing that much of the Lucan corpus is also present in Matthew and Mark and not original to Luke, one may also speak of a ‘local pentecostal theology’, tentatively treating the collected songs as a textual corpus, despite their wide use and disparate origins.

An important aspect of this is the level of control. To what extent is the repertoire—and its theological profile—a matter of conscious choice? And if it is, who has authority to decide what songs to include? In citam, leadership is controlled by formal democratic and hierarchical structures, and through them formalized control is also exercised over what is preached in terms of doctrine. In Mavuno the control is exercised informally through a tight and diligent service-planning process and the charisma of the top leaders. Likewise, there is a control over the sung theology. In Woodley the worship leader in charge decides on what to sing on a specific Sunday, then sends the worship list to the worship pastor for approval before rehearsing with the team and presenting the song(s) in church. Should the Senior Pastor, the elders and deacons, or even the Bishop, have any objections to a specific song, they may disapprove of further use. In Mavuno, the tight planning process of services also includes the songs; hence, the worship team’s choices are checked in advance, and again afterwards in that the leading pastors may choose to remove a song from the repertoire if they see it as unfitting in some regard. Comments and suggestions from congregants may also be taken into account when worship leaders decide on the repertoire, meaning that an organic process operates to build a local consensus. If a song deviates too much from the general (explicit or implicit) local theology, it will soon be replaced by another.

In terms of method, the process of analysis went as follows: for each church, I listened to the sound recordings with the compilation of lyrics in front of me and made notes: marking, for example, the names and titles used to address God, Trinitarian references, ways to describe God, references to Scripture (implicit and explicit), personal pronoun usage, type of songs, words for worship and so on. Then I made charts displaying the outcome in tables to see the overall pattern (see examples in Appendix 3). While working on the analysis and writing stages, I kept moving back and forth between lyrics, charts, and internet sources such as,,,, and many others.67 Analysis thus focused on the theological content of songs, and not, for example, on linguistic qualities or stylistic form (but see form-functional typology, discussed in Chapter 7). This distinctly theological focus is, however, consistent not only with my own research interest, but also with the songs themselves and their doxological, liturgical, and sacramental purposes.

Having now toured through my research process as a fieldwork journey, a data-collection maze and a quagmire of analysis—and it was indeed a journey, a maze, and sometimes a morass of frustration and confusion—it is time to address ethical considerations and my own role as a researcher.

5 Ethical Considerations

5.1 Access, Anonymity and Informed Consent

I used the used the aaa ethical guidelines and general recommendations for ethnography when designing and conducting fieldwork,68 although later had to adjust my writing to the stricter gdpr legal obligations in Sweden.

In my first phase of fieldwork, I carried a letter that introduced me and my research project and one of recommendation from Lund University, which I handed to the people whom I met for interviews; I then explained my project and asked them for their oral consent to take part in research. As I was interviewing well-educated people in elite positions, I counted on everyone having understood the information. I also asked the leaders if they would be willing to receive me in their church for a second round of fieldwork. As both Mavuno and Woodley are large organizations, formal consent could not be gained from each individual coming to church, but Senior Pastors in both churches granted their general consent to my presence. This access was confirmed via email before I returned to the field and continuously during my time in the churches.

When I met with interviewees in the second phase of fieldwork, I asked them to fill in a form with their names, information about their role in the church or organization, age, and gender; here they also signed their consent to taking part in the research and ticked a box to indicate whether they wanted their names to be anonymized in the study or not. Most did not want to be anonymous, rather the opposite; they were happy to take part in the study and to articulate their views and the views of their church. However, the gdpr regulations make this impossible, since views on religion are counted as sensitive, personal information and must be protected according to Swedish law. Therefore, I have decided to give all research participants pseudonyms (whether I met them for formal interviews or informal conversations). As an extra precaution, I have chosen to designate all church leaders as ‘Pastor’ regardless of their role in the church, but readers should be aware that I interviewed people at all levels of pastoral leadership. The list of interviewees in Appendix 7 gives a general indication of each person’s relative position.

I made two exceptions to this rule. First, I did not anonymize local academics and theologians, since they did not speak of their own personal faith but talked in general about the situation in Kenyan churches or Mavuno/Woodley. Second, when I describe church services, I keep the real names of people in leadership positions acting on the platform, but anonymize the names of ordinary congregants when applicable. The church services are open to the public and, in the case of Mavuno, partly made public on the internet, so it is not a matter of breaking confidentiality. In effect, it would be very difficult to anonymize the services in any real sense, since those who know of these two churches would immediately recognize the people involved from my description of their demeanours.

At the beginning of my project, I applied for a research permit from local authorities at the National Council for Science and Technology (a department of Kenyan government later re-named nacosti). After a bit of hassle, and many visits to different offices, the permission was legally granted.

5.2 Copyright Issues

Since part of my data consists of lyrics to songs that come under copyright regulations, I have discussed the ethical and legal side of this work with staff at Lund University Libraries.69 Following their recommendations, I have taken pains to identify the author in each case and used an official version of the song whenever I quote lyrics. This information has been sought via, for hymns, and via, for contemporary worship songs in English. The situation was more difficult for the Swahili songs, since I did not have any song books or general online resources to consult; instead I had to rely on searches via Google and YouTube for each song. I have done my best in this regard, and if I have made mistakes, authors and copyright holders are welcome to contact me with corrections. For a full list of references to songs, see Appendix 7.

The authors (song writers) are referenced in footnotes when I quote lyrics unless I have not been able to identify who they are, whereupon I have mentioned this in the footnote. Normally, I quote only parts of the songs unless the analysis requires the whole text. This is within the legal requirements for an academic work, and I have not sought specific permission to use the lyrics.70

For quotes from the Bible, I have stated which version is used in each case. I mostly use the New International Version in English and the Swahili Union Version in Swahili.

5.3 The Role of the Researcher

In field based research, ethical considerations also include reflections on the role of the researcher vis-à-vis the research field and research participants. What kind of power relations were at play in the field? What position did the researcher take: in the field, before, and after? How did her age, gender, ethnicity, and religious affiliation affect the research? What biases did she bring with her? Did she in any way impact the research site with her presence? How does she represent the people she studied in her writing?71 Important as these questions may be, it is hard to give anything but partial answers to them, given space constraints; nonetheless, let me try add a reflexive take on who I am and the position I have taken in this research.

As I mentioned in the Introduction, I grew up as a child of Swedish pentecostal missionaries to Eastern Africa. I was born in Kigali, Rwanda, went to boarding school in Burundi, and spent eight out of my first twelve years in Africa. For vacations we travelled across the region, visiting countries like Tanzania, Uganda, Kongo/Zaire, Zimbabwe, Zambia; we often went to Kenya, enjoying the coastal life of Mombasa or visiting Nairobi where my sister went to boarding school. When we came back to Sweden my father kept travelling back and forth to the region and later my parents worked in South Sudan for several years. In this sense, I was in no way a new-comer to East Africa when I embarked on fieldwork, although I had never before lived in Nairobi. I was also no newcomer to African Pentecostalism, having been infused since childhood with the sounds and sights of praying congregations, preaching pastors, dancing choirs, and singing Sunday schools. Their spirituality fascinated me early on, although it was not until much later that I could put words to what I had experienced. Together with my family, I literally went to hundreds of services, in dozens of churches, lasting for hours and hours, often in languages I did not comprehend. Although I did not know it then, my participant observation of African Pentecostalism had begun already. Later in life, I have worked as a Bible teacher in pentecostal and evangelical churches in Sweden, travelled widely in Africa and across the world, and visited many different kinds of pentecostal and charismatic communities. I consider myself a pentecostal theologian, with my roots firmly in the pentecostal-charismatic tradition and a desire to develop and deepen the theological reflection among Pentecostals. Thus, in terms of spirituality and theology, I am a complete insider to this research.

However, in terms of culture and ethnicity, I am an outsider. I was born in Africa, but I did not grow up as an African. I was immersed in Swedish culture, went to Swedish schools, spoke barely a word other than Swedish, celebrated Swedish holidays, and had mainly Swedish friends. When I came to the field, I had never before lived in Kenya, and I had not visited the country as an adult. I was also an outsider to the two congregations; I knew no one in church and had never participated in any of their services. In very many ways I was a stranger in the field, and had to learn everything from scratch: simple things like how to travel around the city, and more complicated ones such as how to relate to the church elite. And not least, how to get to know the congregants. Growing up, I had an acute sense of being different from everyone around me (and it was not a pleasant one, I would have much preferred to fit in). This feeling of being strange kicked in when I came to Kenya and, paired with a personality that rather observes than participates, I had a hard time finding my role in church and often retired to my notes and my camera rather than trying to interact with congregants. The urban atmosphere in Nairobi did not exactly help either. Interacting with pastors was somehow easier. They were often already aware of my presence in the church and curious about the research topic. Most of the leaders I met in Nairobi were very welcoming, open to cooperating with me, and encouraging me in my research. My insider status did help in this regard, that I spoke the ‘pentecostal language’ and asked questions that revealed my knowledge of the tradition. However, asymmetries remained. For most of them, I was a guest, a student, and a stranger, not a friend or colleague. I was there due to their goodwill, nothing else, although with some I felt a different kind of connection, one where they enjoyed the conversations with me and even thanked me for spurring their reflections on an important topic.

So, am I an insider or an outsider in this research? The simple answer is that I am both. Like any other position it has both advantages and disadvantages. The insider status helped me gain access, observe nuances, and ask relevant questions, the outsider status helped me look at the congregations with a critical eye. At the same time, the outsider status made fieldwork emotionally challenging, and the insider status will always come with a risk of being uncritical and inattentive to details all too familiar. I have been aware of these risks throughout the project and taken pains to reflect on them at each stage of research. In writing this book, I have opted for a transparent and reflexive way of writing—where my own process and my own role as co-producer of data are made visible—so that readers can judge for themselves whether I have succeeded in handling my insider-outsider position in a responsible manner.

Did I cause any harm in the congregations where I did my research? I certainly hope not, and I have no reason to fear that I did. Did I in any way affect the congregations? In some small way I might have. The Senior Pastor in Woodley was reportedly thrilled with the results of the survey, which I had shared with the leadership, and said that he now understood better who his congregation really was, but I do not know if this had any concrete effect on church ministry. Some of the interviewees said that my questions had cast light on the importance of music and worship, and given them ideas for future sermons, but whether they implemented them or not, I cannot say. Other than that, I do not think that my presence in Nairobi had any impact whatsoever; these churches are much too solid to be affected by the temporary presence of a Swedish observer. And that is as it should be.

One last note must be made concerning my own role, and that has to do with the economics of research. What are the ethical implications of European-based and funded research mediating sub-Saharan voices? What right do I have to ‘represent’ their voices to the world? On the one hand, I have none. They have graciously given me access to their world, and I can only try and be as faithful as possible to their intentions as I write. On the other hand, money talks. The research is de facto funded and published in the West, for a Western-oriented academic audience, and that affects both the content and the form of what I write. Is it ethical? Well, the old order—where European theology dealt only with European issues in European languages—was certainly not. And the future order—where academic institutions around the globe fund and publish high-quality research on topics of interest to people around the globe—is yet to be fully realized. Much work must still be done to strengthen academic institutions in Africa. That said, the ideal cannot be one where only Kenyans can study Kenya, or only Swedes study Sweden. As an academic community, we must be able to study things that are outside our own backyard. This is, after all, one of the points of academic work: to learn new things. Personally, I would much rather funnel the voices of sub-Saharan Christians through European-based and funded research than not at all.

In this chapter I have discussed the research design of this current project, the details of fieldwork, data collection, and analysis, as well as the ethical dimensions of my work in some detail. For methodological and ethical reasons, my ambition has been to be as transparent and clear as possible about how the study was conducted. Ultimately, I hope readers will sense that respect, curiosity, integrity, and thoroughness have been my watchwords throughout the project, and that the working procedures can inspire other researchers as I myself have been inspired.


Yin, Case Study Research and Applications: Design and Methods, 15. List in original. Yin’s case study method have been applied to practical theology in Cartledge, Practical Theology: Charismatic and Empirical Perspectives, 82–102.


Yin, Case Study Research and Applications: Design and Methods, 15.


Yin, 13.


Yin, 31.


Yin, 47–61. Compare Jaison, Qualitative Research and Transformative Results, 44–46; Asbjørn Johannessen et al., Introduktion till Samhällsvetenskaplig Metod, Upplaga 2 (Stockholm: Liber, 2020), 195–98.


Yin, Case Study Research and Applications: Design and Methods, 49.


Jaison, Qualitative Research and Transformative Results, 45.


Robert K. Yin, ‘How to Do Better Case Studies’, in The sage Handbook of Applied Social Research Methods, ed. Leonard Bickman and Debra J. Rog, 2nd ed (Los Angeles: sage, 2009), 259. Emphasis in original.


Yin, Case Study Research and Applications: Design and Methods, 226–28. Compare Johannessen et al., Introduktion till Samhällsvetenskaplig Metod, 198.


Yin, Case Study Research and Applications: Design and Methods, 126–30. Triangulation is common also in ethnographic field work: Karen O’Reilly, Ethnographic Methods (London: Routledge, 2012); Mats Alvesson and Kaj Sköldberg, Reflexive Methodology: New Vistas for Qualitative Research (Los Angeles: sage, 2009), 86; Grimes, The Craft of Ritual Studies.


Jaison, Qualitative Research and Transformative Results, 44.


The sub-unit ‘congregation’ is weaker than I had hoped for due to problems at the gathering phase as described below. Nevertheless, I have retained it as one of the sub-units since it constitutes an important part of the whole.


Amanda Coffey, The Ethnographic Self: Fieldwork and the Representation of Identity (London: Sage, 1999); James V. Spickard, J. Shawn Landres, and Meredith B. McGuire, Personal Knowledge and beyond: Reshaping the Ethnography of Religion (New York: New York University Press, 2002); compare Grimes, The Craft of Ritual Studies.


Smith, Thinking in Tongues: Pentecostal Contributions to Christian Philosophy, xx.


Ward, ‘Introduction’, 9.


Ritual studies is heavily indebted to ethnography and anthropology, see for example Grimes, The Craft of Ritual Studies, 11–54, 165–210.


O’Reilly, Ethnographic Methods, 23–24, 43.


Alvesson and Sköldberg, Reflexive Methodology: New Vistas for Qualitative Research, 85.


Corrine Glesne, Becoming Qualitative Researchers: An Introduction (Boston, MA: Pearson, 2011), 56.


Kathleen M. DeWalt and Billie R. DeWalt, Participant Observation: A Guide for Fieldworkers (Walnut Creek, Calif.: AltaMira Press, 2002), 37.


Pranee Liamputtong, Performing Qualitative Cross-Cultural Research (Cambridge: Cambridge University Press, 2010), 69.


During my pilot study I visited the following churches, congregations and ministries: Nairobi Pentecostal Church/citam (Valley Road, Woodley and Kiserian), as well as Nairobi Baptist Church, Nairobi Chapel, Mamlaka Hill Chapel, Mavuno Church Bellevue, Jubilee Christian Church, International Christian Church, House of Grace and Joe Kayo Ministires.


Glesne, Becoming Qualitative Researchers: An Introduction, 57; DeWalt and DeWalt, Participant Observation: A Guide for Fieldworkers, 36.


Alvesson and Sköldberg, Reflexive Methodology: New Vistas for Qualitative Research, 84–85; O’Reilly, Ethnographic Methods, 86–115; DeWalt and DeWalt, Participant Observation: A Guide for Fieldworkers.


Grimes, The Craft of Ritual Studies, 45.


Grimes, 45.


Referred to in DeWalt and DeWalt, Participant Observation: A Guide for Fieldworkers, 19–21.


O’Reilly, Ethnographic Methods, 96–100.


DeWalt and DeWalt, Participant Observation: A Guide for Fieldworkers, 24.


O’Reilly, Ethnographic Methods, 106.


Grimes, The Craft of Ritual Studies, 45.


Grimes, 78–92.


Robert M. Emerson, Rachel I. Fretz, and Linda L. Shaw, Writing Ethnographic Fieldnotes (Chicago: University of Chicago Press, 1995); DeWalt and DeWalt, Participant Observation: A Guide for Fieldworkers, 141–62; Glesne, Becoming Qualitative Researchers: An Introduction, 71–80; Grimes, The Craft of Ritual Studies, 101–5.


DeWalt and DeWalt, Participant Observation: A Guide for Fieldworkers, 143.


Observation notes and fieldnotes are referenced when directly quoted or paraphrased, not otherwise.


O’Reilly, Ethnographic Methods, 104.


Compare DeWalt and DeWalt, Participant Observation: A Guide for Fieldworkers, 141, where this is referred to as “records of prolonged activities and ceremonies”.


DeWalt and DeWalt, 72–73.


Grimes, The Craft of Ritual Studies, 79.


O’Reilly, Ethnographic Methods, 120.


O’Reilly, 116–27.


Grimes, The Craft of Ritual Studies, 47–51.


O’Reilly, Ethnographic Methods, 125.


Steinar Kvale and Svend Brinkmann, Den Kvalitativa Forskningsintervjun (Lund: Studentlitteratur, 2009), 163.


O’Reilly, Ethnographic Methods, 136.


Liamputtong, Performing Qualitative Cross-Cultural Research.


Compare Kvale and Brinkmann, Den Kvalitativa Forskningsintervjun, 150–57.


Kvale and Brinkmann, 167–76.


Packiam, ‘Worship’, 323.


The Mavuno Music Blog, (accessed 2020- 02-25).


Grimes, The Craft of Ritual Studies; Bell, Ritual: Perspectives and Dimensions; Csordas, ‘Somatic Modes of Attention’; T. Csordas, Language, Charisma, and Creativity: Ritual Life in the Catholic Charismatic Renewal, Contemporary Anthropology of Religion (Basingstoke: New York: Palgrave Macmillan US, 2012); Lindhardt, Practicing the Faith: The Ritual Life of Pentecostal-Charismatic Christians; Albrecht, Rites in the Spirit: A Ritual Approach to Pentecostal/Charismatic Spirituality.


Glesne, Becoming Qualitative Researchers: An Introduction, 81; Lars Kaijser and Magnus Öhlander, Etnologiskt Fältarbete (Lund: Studentlitteratur, 2011).


Land, Pentecostal Spirituality: A Passion for the Kingdom; Cartledge, Encountering the Spirit: The Charismatic Tradition; Cartledge, Mediation of the Spirit: Interventions in Practical Theology; Cartledge, Testimony in the Spirit: Rescripting Ordinary Pentecostal Theology; Archer, The Gospel Revisited: Towards a Pentecostal Theology of Worship and Witness; Smith, Thinking in Tongues: Pentecostal Contributions to Christian Philosophy; Yong, Spirit of Love: A Trinitarian Theology of Grace; Yong, ‘Conclusion: Improvisation, Indigenization, and Inspiration: Theological Reflections on the Sound and Spirit of Global Renewal’; Josefsson, Liv och Över Nog: Den Tidiga Pingströrelsens Spiritualitet. Abundant Life: The Spirituality of the Early Pentecostal Movement in Sweden.


Clarke, Pentecostal Theology in Africa; Kalu, African Pentecostalism: An Introduction; Mugambi, A Spirit of Revitalization: Urban Pentecostalism in Kenya.


Compare O’Reilly, Ethnographic Methods, 30.


Kvale and Brinkmann, Den Kvalitativa Forskningsintervjun, 251. Emphasis in original.


See Appendix.


Kvale and Brinkmann, Den Kvalitativa Forskningsintervjun, 253–58.


See Appendix 7.


See Appendix 7.


I discuss this in a forthcoming book chapter: Björkander, ‘Who Got the Rite Wrong? The Mavuno Alternative Christmas Service and Charismatic Ritual’.


By ‘secular’ I mean songs that are normally sung in non-ecclesial settings and that do not have explicitly Christian lyrics, in accordance with local usage.


I.e. congregational song, sung according to the ritual patterns of ‘the rite of worship and praise’ described in Chapter 5.


Kidula, ‘Singing the Lord’s Song in the Spirit and with Understanding: The Practice of Nairobi Pentecostal Church’, 133–47.


In Nairobi I heard Christian music played on buses and in shopping malls, and Christian cds were available on every corner, not to mention the internet. See also Parsitau, ‘“Then Sings My Soul”: Gospel Music as Popular Culture in the Spiritual Lives of Kenyan Pentecostal/Charismatic Christians’, 3.


Ward, ‘Introduction’, 9.


It has sometimes been hard to find reliable sources for the origin of a song or its original lyrics, but I have done my best to cross-check between different sources and report them here. In this I have followed the advice of librarians Annakim Eltén and Andrea Mervik, Lund University Libraries (Consultation 2017-01-24).


Compare discussion on ethical considerations in Grimes, The Craft of Ritual Studies, 37–44; Jaison, Qualitative Research and Transformative Results, 56–79; O’Reilly, Ethnographic Methods, 62–72.


Consultation with Annakim Eltén and Andrea Mervik, Lund University Libraries (2017-01-24).


Henry Olsson and Jan Rosén, Upphovsrättslagstiftningen: En Kommentar (Stockholm: Wolters Kluwer, 2016).


Grimes, The Craft of Ritual Studies, 35–46; O’Reilly, Ethnographic Methods, 62–85, 208–30.

  • Collapse
  • Expand