Chapter 3: Research Design
3.5 Data Analysis
Analysis of the data in this study was an ongoing process. As with traditional ethnographic analysis, throughout the study I was engaged in writing fieldnotes and analytical memos based on my thoughts and experiences within the research setting (Spradley, 1979). Fieldnotes were the primary means of data. They were written while in the field and described what I observed firsthand. Analytical memos and reflexive memos were also written while in the field to connect my fieldnotes with concepts and theories that informed my research (e.g., a specific action of an individual may have appeared to me as working to disrupt an institution). Additionally, throughout my time in the field, I immersed myself in my written fieldnotes and analyzed them inductively for similar themes and occurrences. Reflexive and analytical memos were also analyzed for themes and patterns, which informed my data collection strategies upon re-entering the field by providing me with specific elements to examine while acting as participant-observer (Lofland et al., 2006).
Immersing myself in the data in this way not only provided me with patterns and themes – it also allowed me to gain deeper understanding of the research setting and the social relations in which I was involved during the research. As the critical institutional ethnographer, it was important for me to remember during data analysis and writing up the research that I wished to explicate people’s everyday experiences and their actualities, rather than just describing an account based on a local’s viewpoint (Campbell & Gregor, 2002; Nichols, 2014; Smith, 2005). Therefore, ultimately, through data collection, data analysis, and the writing up of research,
Getting to an account that explicates the social relations of the setting is what an institutional ethnographic account is about. This kind of analysis uses what informants know and what they are observed doing for the analytical purpose of identifying, tracing, and describing the social relations that extend beyond the boundaries of any one informant’s experiences (or even all informants’
experiences)…An institutional ethnography must therefore include research into those elements of social organization that connect the local setting and local experiences to sites outside the experiential setting. (Campbell & Gregor, 2002, p.
90)
This way of data analysis allowed me to appreciate how my data collection methods were inherently tied to my findings and data analysis. Additionally, it provided a clear and logical way to use critical ethnography and institutional ethnography to analyze the institutional work of the SSC.
When I arrived back in Canada in late August, intensive data analysis using the qualitative analysis software programme NVivo 10 took place. In order to provide a clear step-by-step procedure of my data analysis, an outline below is provided:
1) In early September 2016, I inputted/transcribed all of my collected data into Microsoft Word, including all written fieldnotes, reflexive fieldnotes, and the interviews I had transcribed using the transcription computer programme Inqscribe. The amount of handwritten notes included:
-‐‑ In-fieldnotes: 8 50-page small pocket sized notebooks, totaling 375 pages of in-fieldnotes.
-‐‑ Full fieldnotes: One large 200-page notebook and one large 154-page notebook, totaling 354 pages of fieldnotes.
-‐‑ Reflexive fieldnotes: 27 pages total.
2) As I transcribed my data and inputted it into Microsoft Word, I also maintained a small memo-pad for analytical purposes as I typed up my data. This memo-pad was where I wrote small analytical memos, reflexive notes, and observational comments as I transcribed. In late September/early October, all of my fieldnotes and interviews had been inputted into Microsoft Word. The following table provides an overview of my data:
Table 2 - Microsoft Word Transcribed Data
Type of Data Information Total
Transcribed Interview Thabani – 36 minutes 14 Interview Thembile – 46 minutes 13 Interview Jabulani – 58 minutes 14 Interview Margaret #1 – 65 minutes 20 Interview Margaret #2 – 64 minutes
(2 parts)
21 Total Time and
Pages:
310 minutes (5.2 hours) 297 Pages
3) After inputting all transcribed data into Microsoft Word, I then inputted all of my data (fieldnotes, interviews, and documents) into the qualitative analysis software programme NVivo 10 beginning in mid-October 2016. Documents that were collected were reviewed, with many being copies of each other in different
folders, or being repetitive based on the information within. Initially, documents were grouped in file folders according to the SMGSA and SSC. After removing documents based on repetition, copies, and relevance to the study, I personally categorized documents into groups in NVivo 10 based on similar information and purpose. The tables in Appendix G display the categories and number of pages of documents, as well as display the total amount of documents before and after I reviewed them.
4) I organized the data in NVivo 10 based on type – for example, fieldnotes were organized as fieldnotes and then broken into ‘reflexive’ and ‘in-field.’ Interviews were organized by interviewee. Documents were organized based on the
categories outlined in Appendix H. In order to further review fieldnotes, I
organized the data into locations – for example, if a specific fieldnote was written based on something I witnessed at the SSC, I created a node titled ‘SSC’ and placed the data in that node. I also did the same for people – for example, a node titled ‘Thabani’ was where conversations and actions of Thabani were filed under.
I also organized data into nodes based on SFD programming, such as ‘LVP.’
Documents such as modules and conversations or interview parts based on ‘LVP’
would be placed in this node – the same was done for ‘LIT’, ‘TRRFCC Values’, and other prominent events or programmes at the SSC, such as the ‘Olympic Day Run.’
5) Following the use of word queries, I began to open code my data. This was an iterative process where I would continuously read my data, write analytical memos and annotations, as well as review connections between chunks of data.
Some nodes were created before open coding occurred – these nodes were based on research participant names (e.g., Thembile) or SFD programmes (e.g., LIT programme). Upon open-coding, I would go through each data source and create nodes as I coded, creating new nodes that I would look for in other sources. For example, one node I created was based on conversations about the SSC – it being a ‘safe place’ or ‘home’ to certain individuals. I then coded other interviews and conversations to identify whether this pattern was prominent in other chunks of data, which in some cases, it was. See Appendix I for an example of coding chunks in NVivo 10.
6) As I coded, I made notes and analytical memos about the difference in themes and patterns based on sources – for example, many modules of the LIT programme focused on HIV/AIDS education. However, when coding fieldnotes, HIV/AIDS education was rarely mentioned and if it was, little information was discussed in relation to the document modules. These analytical memos and comments would assist later in understanding the SSC. They were also important to examine how what is not spoken, observed, or written in such memos and comments may provide further insight into my own research and interpretation process.
7) Coding took approximately from mid-October to late November/early December.
The iterative process involved me going back and forth between my data sources and immersing myself in the data in order to achieve saturation. Around early December, I achieved saturation once coding became repetitive and new nodes were not being created based on my analysis. The final nodes, themes, and patterns were given descriptions based on the references located in each. For
example, when examining the ‘Leadership’ node/theme, references to each source that was coded based on this node would be able to be seen in NVivo 10 (see Appendix J). This provided easy access to each reference located in the node/theme.
8) Based on analysis and data saturation, nodes that I created before open coding as well as new themes that emerged as I analyzed my data were included in the final list of themes and patterns – overall, four prominent themes emerged, each with additional descriptive nodes in each theme classification. For an overview of each theme and sub-themes, please see Appendix K.
9) Throughout the process of data analysis and the utilization of NVivo 10,
interpretation of the data based on my ‘sensitizing concepts’ occurred through the use of analytical memos and comments. Additionally, as I analyzed the data, connections to the literature on SFD and international development was continuous in order to prepare for the writing phase of the study.
The process of data analysis was an intensive, iterative process where I conducted coding and reviewed my data sources that led to immersion in my data and data saturation.
Although I used the NVivo 10 analysis programme to assist in data analysis, a period of training and self-teaching occurred in order for me to be able to utilize this analysis tool.
Throughout the duration of data analysis, I constantly went through the three phases of data description, analysis, and interpretation – although it appears linear as described here, this process was iterative, where I went from one stage to the other randomly and at different times, as well as simultaneously back and forth. Once data analysis was
completed, the writing stage of this study began – however, throughout the process of this
research, including entering the field, collecting data, data analysis, and the writing up of this thesis, I had to ensure that I was conducting trustworthy research. In the next section, I discuss the important elements of trustworthiness, reflexivity, and ethics in any
ethnographic study.
3.6 Trustworthiness
As a qualitative researcher, the term trustworthiness plays a much different role than how it has been traditionally defined by quantitative scholars, whose focus relies heavily on generalizability, validity, and reliability (Sparkes, 1998). The issue of validity and trustworthiness in qualitative research has long been a question for researchers – some qualitative scholars believe it is near impossible to achieve validity in their work, whereas others believe that there are ways to effectively display trustworthiness (Kvale, 1995; Sparkes, 1998). In this study, I enhanced trustworthiness in various forms based on different conceptualizations. In one way, trustworthiness was accomplished based on certain ‘criteria for excellent qualitative research’ (Tracy, 2010); in the other way, a different, less ’criteria-based’ form of trustworthiness was also recognized that relates to arguments of the legitimization of knowledge.
Tracy (2010) outlined eight criteria of excellent qualitative research – although she suggested these criteria, she also argued that each, “may be achieved through a variety of craft skills that are flexible depending on the goals of the study and preferences/skills of the researcher” (p. 839). The criteria are to represent a tool for researchers when conducting qualitative research and also increase dialogue in relation to methodological trustworthiness. The criteria include: worthy topic; rich rigor; sincerity;
credibility; resonance; significant contribution; ethical; and meaningful coherence. This
study achieved various ‘criteria’ of qualitative research - for example, as discussed in the Case Study section, acquiring a large amount of data (documents, interviews, and
participant-observation) for this study as well as spending four months in the field (as both researcher/intern) is two ways of achieving ‘rich rigor.’ Additionally, ‘credibility’
was accomplished by providing “thick description, concrete detail, explication of tacit (nontextual) knowledge, and showing rather than telling” (Tracy, 2010, p. 840). Due to the ethnographic approach used in this study, and my methods while in the field, detailed description and nontextual knowledge was gathered (see Chapters 5 and 6). Finally, the two criteria of ‘sincerity’ and ‘ethical’ were attained throughout this research process.
Specifically, in terms of sincerity, I engaged in “self-reflexivity about subjective values, biases, and inclinations of the researcher(s)” (Tracy, 2010, p. 840) as seen by my
methodological approach and, as further addressed in this section and Chapter 4, my reflexive process. I also emphasized a relational ethics approach to research – a component of the ‘ethical’ criteria for qualitative studies that Tracy (2010) discussed.
Before discussing my methods of reflexivity and relational ethics to achieve trustworthiness, it is important to also recognize how scholars adopting a critical, poststructuralist, or postmodernist stance have questioned trustworthiness, most especially due to my own critical and postcolonial theoretical frameworks.
When discussing the role of validity in contemporary postpositivism and from a poststructuralist stance, Lather (1993) emphasized how the “conditions of the
legitimization of knowledge” (p. 673) have to be re-thought in an antifoundational discourse context. Rejecting the notion that methodology and the research process can be detached objectively from the researcher, Lather (1993) tried to loosen the code that
positivist science has shaped in relation to validity in research – and thus, argued for a dialogue with readers about how to approach reality construction, which, “involves making decisions about which discursive policy to follow, which ‘regime of truth’ to locate one’s work within, which mask of methodology to assume” (p. 676). Therefore, although I outlined specific ‘criteria’ and the techniques used in my study that enhanced the quality and trustworthiness of the findings, my central approach to trustworthiness was through engaging in reflexive, ethical practices of qualitative research and
knowledge production in order to account for power issues in the conduct of research. By recognizing how trustworthiness and other forms of ‘validity’ in research have been socially constructed, trustworthiness is achieved in different ways that focus on power dynamics inherent in research rather than those ‘checklists’ of qualitative methodology (Kvale, 1995). In this study, I enhanced trustworthiness by ‘increasing consciousness’
(Morrow, 2005):
Increasing consciousness involves identifying sources of inequality and
representing the perspectives of those who have been silenced or disempowered.
It also explores and makes visible who benefits from power and how power is exercised. Research participants are collaborators in the action/investigation, and researchers attend to the power issues and relationships between and among researcher and researched. (p. 253).
In the complex, institutional social setting of the SSC, I was aware of how certain
knowledge or ideas of my own had been selected and utilized in the research process and sought to be reflective about how and why I chose certain ways of research methodology (i.e., critical institutional ethnographic case study) and particular conceptual frames (i.e.,
institutional work). Although research participants were not collaborative in each step of this research process (such as data analysis), my own reflexivity and ethical approach ensured that the voices of those who have been traditionally marginalized were heard and represented. Additionally, I strived to constantly engage in issues of power between the researcher and the researched – for example, by acknowledging the power I had in the process of this research, I would also work to display to people in the SSC that they were the ‘experts’ (see Chapter 6 for a specific example). In this way, and influenced by my critical institutional ethnographic approach, I sought to give voice to those at the local SSC who had been traditionally marginalized in the organization and also ‘Othered’
historically. Approaching trustworthiness in this way was suitable for conducting a critical institutional ethnographic case study, as both critical and institutional
ethnography emphasize the need to recognize knowledge production when conducting research. In the following sections, I outline how I engaged in ethical practices and reflexivity while in the field and throughout the research process in order to facilitate trustworthiness.
3.6.1 Relational Ethics
Critical ethnographers are concerned with how ethnographies are completed for political purposes (that of research)34 and how they must recognize that these politics do not stand outside of the ethical procedure (Davies, 2008; Madison, 2005). I wish to emphasize the importance of ethics and reflexivity and expand on how I strived towards trustworthiness. In relation to ethics and informed consent, Davies (2008) discussed how researchers may face multiple difficulties that must be addressed when presenting their
34 Another political reason for conducting research is as part of the completion of a Master of Arts’ degree.
research to potential participants. Researchers must think of how to present their research in a way that is meaningful to participants and use language that is free of mixed
interpretations. Although in Swaziland they have two first languages, one of which is English and the other siSwati, I remained conscious that a person in Swaziland may interpret what certain words and phrases may mean completely differently than my interpretation – for example, ‘fishing’ (see Chapter 4). Recognizing this, I presented my research in a language that is free of specialized terms drawn from specific theories and concepts – although the participant may have recognized the terms, I wished to present the research to all participants in the same manner and hoped that many interpreted it meaningfully.
Davies (2008) also commented on how researchers may face the dilemma of explaining certain aspects of the study to individuals – for example, in my case, it was sometimes hard to specifically identify what I was ‘looking for’ or hoping to uncover, as I wished to have my research problematic arise out of participants’ knowledge and experiences. As a critical institutional ethnographer, I explained clearly that for me it was a process of discovery where the implications were not known at the outset and that, as I was involved in the field, the study evolved as I learned. Furthermore, once I explained the research to participants, I noticed that some participants may have tended to forget that I was also acting as a researcher, and I would occasionally remind them of my role, especially when trusted with information that may have been provided to me due to friendships or informal encounters (Davies, 2008). For example, as an intern and friend to many at the SSC, some would confide in me about their activities outside the centre or about their dislike for certain people in the organization – at these times, I would
sometimes change direction of conversation or subtly remind them of my dual role as intern and researcher. Like Davies (2008), I saw consent as a process and not a one-off event. However, relational ethics is only one layer of ethical commitment – in addition, moments of situational ethics occurred while in the field. These moments involved respecting other’s viewpoint or assumptions and reflecting on my own ethical standing and place in the organization. For example, discussed more fully in Chapter 4, situational ethics were needed when Margaret would say something with potential racial
connotations – in order to deal with my own feelings, the relations of myself and others, and ethical procedures, I would reflexively negotiate my position based on certain situations of these kind.
Participants of any research must also be informed of the risks. Although there were no clear risks of the research beyond the risks present in everyday life in the SSC, I informed participants that the research findings, when written and presented, might be interpreted or used outside of my control. If a participant did not wish to be involved in the research or answer any questions I had, I respected their wishes and ensured that they were not involved in any aspect of the study – however, this never occurred. Another ethical consideration in social research is confidentiality. Confidentiality “essentially concerns the treatment of information gained about individuals in the course of research”
(Davies, 2008, p. 59). Researchers must be cautious when informing participants that their participation is kept confidential, as researchers must be realistic to the extent that they are able to protect one’s identity (Davies, 2008). Especially in ongoing research such as participant-observation, I had to ensure that confidentiality, like informed consent, was an ongoing process as I built relationships with those with whom I was involved and