57 | P a g e Moreover, as 25 students were interviewed I was satisfied that data saturation had been reached and that sufficient in-depth data had been obtained to address the research questions and objectives and, ultimately, to achieve the aim of the study.
4.5 Data Collection and Analysis
58 | P a g e the university had implemented, therefore this was the only way I could ask a student to first do the screening before we begin with the one-on-one interviews.
In-depth interviews are recognized as an anthropological method of data collection. This data collection method is deemed pertinent in anthropological research as it permits the researcher to collect information on people's experiences using open-ended questions in an interview schedule.
According to Cresswell (2011), interviewing is “...a crucial part of qualitative fieldwork methodology. It helps to understand the meanings that individuals give to their lives and the social phenomena they have experienced”. This method efficiently enabled me as the principal investigator to probe for an in-depth understanding of depression among the students. The study used one-on-one interviews during which I asked these probing questions. The participants were comfortable when they narrated their stories, as urged by Reeves et al. (2013). According to Sutton and Austin (2015), in-depth interviews are ideal for obtaining data on individuals' personal histories, perspectives, and experiences, particularly when exploring sensitive topics as was the case in this study.
4.5.2 Data analysis
Qualitative data analysis is a process that moves away from qualitative data collection as the former elicits understanding and interpretations of the people and situation being investigated (Creswell, 2015). It further enlightens the research objectives by revealing patterns and themes in the data (Braun & Clarke, 2014).
In the current study the data were analysed using thematic analysis. According to Nowell et al.
(2017), thematic analysis is a qualitative method that can be used to analyse extensive qualitative data. It is defined as a method for identifying, analysing, organising, describing, and reporting themes found within a data set. Nowell et al. (2017) further posit that a careful thematic analysis can produce trustworthy and insightful findings. Clark (2006, cited by Nowell et al., 2017) argues that thematic analysis is a valuable method for examining the perspectives of different research participants as it allows the highlighting of similarities and differences in the data set and generates unexpected insight. Nowell et al. (2017) further argue that thematic analysis is flexible.
59 | P a g e 4.5.3 Steps of thematic analysis
Thematic analysis is a qualitative data analysis technique and is typically used to describe a group of texts, such as interview transcripts. The researcher studies the data carefully in order to identify recurring themes, subjects, ideas, and patterns of meaning (Caulfield, 2019). Six steps are usually followed in thematic analysis, and these are discussed below.
• Familiarization with transcripts
The first stage, according to Caulfield (2019), is to familiarize oneself with the data. Before beginning to analyse individual words, it is necessary to gain a thorough understanding of all the data obtained. This entails transcribing audio recordings, reading through the text, collecting first notes, and generally looking through the data to become familiar with it. In this study, the data were collected by the primary investigator (myself) which was an advantage because the process of writing/transcribing helped me to become familiar with the content of the data. Furthermore, as I wrote and listened to the audio recordings over and again, the rich data that had appeared complex at first began to make more sense. The data was initially recorded in the form of audio recordings using the Philips Voice Tracer Recorder, but these were transcribed by myself and translated (where IsiZulu had been used) into English. I had to listen to the audio recordings more than once in order to be sure and I re-read the transcripts so that I could get the meaning and the patterns of the participants’ responses before coding. This helped me become familiar with the data and to know when I had reached data saturation.
• Generating codes
After familiarising myself with the data, the next step was coding the transcripts of the organized data. Coding data has the additional benefit of saving time and making the process of analysis much more organized and efficient (Clarke & Braun, 2013). A code is
“a label assigned to a piece of text, and the aim of using a code is to identify and summarise essential concepts within a set of data, such as interview transcripts” (Seldana, 2013:32).
As the purpose of this study was to anthropologically understand depression and attached
60 | P a g e social constructs to it from the views of university students, I systematically categorized excerpts to enable me to find patterns and themes for analysis. Coding allowed me to find what was interesting about the data and I was able to label the social constructs that emerged. As I was transcribing the interviews, I could carefully trace similar keywords and then mark and code the text. This was done manually using highlighters. This use of the inductive approach involved “...deriving meaning and creating themes from the data without any preconceptions” (Caulfield, 2019). I did not conduct the interviews based on expected outcomes, but I allowed the narration of social constructs to emerge from the data and from there find patterns in the texts. The generation of these codes allowed me to adhere to the ethical consideration of not tempering with the narratives of the participants.
• Searching for themes
The codes elicited exciting information and I could detect patterns in the data. Themes are broader than codes and involve active interpretation of the data (Clarke & Braun, 2013).
To generate themes, I looked at the list of codes while coding the data and collated them into broad themes. For example, as I was trying to understand depression, I looked for the participants’ experiences of depression as illuminated in the literature. By carefully reading the transcripts, I was thus able to create subthemes such as ‘feelings of loneliness and isolation’. The different codes were then grouped together into potential themes, and individual extracts of data were collated within the potential themes. Different themes were assembled in terms of their relationship with codes, and overarching themes were identified, as proposed by Mphambo (2011).
• Reviewing the themes
In this step, I read through all the extracts related to the codes in order to explore if they supported the theme, if there were contradictions, and to see if themes overlapped (Mortensen, 2020). Themes were reviewed in relation to the entire data set. Some codes were not used as they did not have enough data that supported them. Furthermore, I noted that some extracts were fitting into multiple themes and thus formed a broader theme altogether, while some had to be subthemes to avoid themes from becoming incoherent.
The re-reading of the themes was to ensure that the entire data set was captured. Mortensen
61 | P a g e (2020) notes that reviewing the themes means going back and forth among the themes, codes, and extracts until the relevant data have been coded and one has coherent themes that represent the data accurately.
• Defining and naming the themes
Each theme was reviewed in order to figure out what the topic was about, and what data it captured. This was accomplished by examining the coherence and narrative tale of each theme's collected data excerpts. Each theme was described in length, both in terms of the narratives associated with it and how it fitted into the larger ‘story’ of the data. I also determined how each theme linked to the study's objectives and research questions.
• Writing up of the themes
Finally, the findings were constructed in explanatory text. This means that extracts from the data set were presented in writing in order to provide enough evidence to show that each theme was relevant to an objective of the study.
4.5.4 Validity, reliability and the generalisability of study findings
The term validity refers to the truth in the research and is the manner in which the study's findings are accurate (Babbie & Mouton, 2005). The dependability of a measurement instrument, which is the extent to which the instrument produces the same results on repeated trials, is referred to as reliability (Mwelase, 2020, cited in Terre Blanche et al., 2006:52). The extent to which it is possible to generalize the data and context of the research study to larger populations and settings is referred to as generalisability (Terre Blanche et al., 2006:91). Because this was a qualitative study, it aimed for credibility, dependability, transferability, and trustworthiness (Vosloo, 2014).
To achieve credibility, truth-value had to be obtained. Thus the perspectives of the sampled participants had to be captured thoroughly and accurately. This was accomplished by spending sufficient time interviewing the participants and discussing issues associated with the topic with them (Mwelase, 2020). To ensure credibility, notes were taken during and after data collection through a rereading of the transcripts in order to make sure they made sense. One data collection
62 | P a g e method was used namely individual semi-structured interviews. This gave credible results as I was sensitive to the participants’ feelings and needs and they were comfortable during the one-on-one interviews. Furthermore, the credibility of the study was ensured as my interpretations and the report were based on meticulous thematic analysis and I ensured “...that the report is logical and easy to understand for other readers" (Shenton, 2003:64). Polit and Beck (2010) believe that the purpose of qualitative research is to provide a rich, contextualized understanding of various elements of human experience through intense study of specific topics or cases, rather than to generalize. The findings of this study can therefore only be generalized to the views of the UKZN- PMB students who participated. The findings are thus not transferable to another set of students but remain the contextual experience of the students that constituted the study sample and whose volunteered narratives were captured as an original collective.