3.8 Approaches to interpretation of data
3.8.4 Types of coding
According to Newman (2000:420) and Sarantakos (2005:350), there are three types of coding that fall into the category of advanced coding; these are open coding, axial coding and selective coding. The study will give a detailed discussion of these in the next subsection. It is worth noting that all of these stages of coding were employed in the study. The coding process took a further two months (June and July 2010). During this process, there was development of categories and refining of themes and this then led to data analysis (Heather 2001:98). The table below gives a quick summary of these processes:
Table 3.7
Data collection and analysis timeframe
Stages Data Collection Strategies
Timeframe Population Sample
S ta g e O n e Focus groups, Semi-structured interviews December 2009 – January 2010 (2 months)
UoA, MMU, MuCo and UHAI groups
S ta g e T w o Transcription March 2010 – May 2010 (2 months and 1 week)
Field Data (wav files) Emerging themes S ta g e T h re e Coding and analysis June 2010 – July 2010 (2 months) MS Word Transcripts Development of categories, refining of themes, culminating with analysis
Open coding
Open coding is the initial step in the process of data analysis. The term ‘open’ implies bringing to the surface analytic categories that have been hidden in the raw data. Open coding is connected to a constant process of taking notes, sorting them out, looking for meanings and comparing notes. Open coding does not end at this point; it proceeds with the creation of labels/symbols/codes/tags in an attempt to discover patterns of similarities or differences (Neuman 2000:422, Sarantakos 2005:349). While the researcher was engaged in the phase of data transcription, which took approximately two months, he was constantly comparing data. The categories included lecturers, students, PLWHAs, others in the academy, Males and females, and graduate and undergraduate students. So when the transcription was complete, the study had approximately 48 tagged MS Word files, of 240 pages. From that point the researcher went into open coding ‘proper,’ where I read the raw data intensively and categorised it into 3 labels namely, Research Question One, Research Question Two and Research Question Three. So in each labelled group there were responses from interviews and Focus Groups Discussions (FGDs). The second label (Research Question Two) was voluminous because this research question is the heart of the study. At this point, no data cleaning or reduction was done; what was done was the formation of the first analytic categories.
Axial Coding
Axial Coding is a second bridge towards data analysis. It entails a further refining of categories and creation of new ones if it is deemed necessary. Neuman (2000:432) argues that at this stage the focus is not on the organised data, but on initial labels or coded concepts. Sarantakos (2005:350) argues further that axial coding does not look critically only at the initial labels or codes, but seeks to establish interrelations within one particular label or code, thereby formulating themes. When the researcher began this phase of coding, he was surprised by the explosion of themes which he had no ‘idea’ of when the first chapter was being written. It should be noted that at this point of data analysis; the researcher began doing some data cleaning. Data cleaning means retaining of information that is relevant to the research objectives (Babbie 1995:372). Also data cleaning requires one to retain ‘catchy’ phrases or paragraphs that are ‘heavily’ loaded with meaning. The
researcher had trouble in the process of data cleaning, because he did not want to leave out valuable gems of thought. All in all, at the end of this process, the researcher was able to reduce 240 pages to 200 pages of loosely integrated data. The table below summarises the themes that emerged during axial coding.
Table 3.8
The development of categories and
themes during axial coding
Research Questions Categories Themes To what extent have the academy and PLWHA adhered to stereotyping and social categorisation?
Introduction Stigmatisation (Major theme), Curriculum, Positions, Ignoring and Culture (Sub-themes)
What are the pointers that the narrative of 2 Samuel 13:1- 14:39 portrays in the context of HIV/AIDS?
Amnon Rape (Major theme), Identification, Curse, Clergy promiscuity, Peer pressure, Curriculum, Corruption, Culture, Church, Moods, Real-life stories and Way forward (Sub-themes) Tamar Church, Culture (Major themes),
Identification, Stigmatisation, Curriculum, Corruption, Economic power, Clergy promiscuity, Evading, Moods, Real-life stories, Way forward (Sub-themes)
What is the way forward towards an HIV/AIDS biblical hermeneutics?
Jonadab Peer effect (Major theme), Identification, Behaviour, Social Justice, Demonology,
Church/State Accountability, Contemporary Jonadabs, Real-life stories, Way forward (Sub-themes) Servants Blind obedience (Major theme),
Identification, Setting, Class, Real-life stories, Way forward (Sub-themes)
Reporters False Reporting, Real-life stories, Way forward (Sub-themes) King’s sons Non-involvement (major theme),
Real-life stories, Way forward(Sub- themes)
Absalom Struggle for power (Major theme), Identification, Revenge,
Leadership style, Donor funding, Real-life stories, Way forward (Sub-themes)
King David Moral failure (Major theme), Parenting, Leadership style, Gender imbalance, Polygamy, Real-life stories, Way forward (Sub-themes)
Joab Reconciliation (Major theme), Genuineness of reconciliation, Gender imbalance, Culture, Contemporary Story, Way forward (Sub-themes)
Woman of
Tekoa
Behaviour (Major theme), Reconciliation, Recognition,
ARV use, Real-life stories, Way
forward (Sub-themes).
Epilogue Reconciliation, Reader response Hermeneutics, Ordinary readers, Empowerment, Fictional versus non-fictional authorship, Narrator’s point of view (Sub- themes)
Additional findings
Unexpected Findings
Selective coding
Selective Coding is the last bridge to cross before reaching the final destination of analysis of the data. At this point, major themes have been identified, using the previous methods like working through notes, diagrams and categories. This also includes looking for regularities, generalisations, and levels of abstraction. Also during this stage of data coding, there is high extraction of raw data (Neuman 2006: 423, Sarantakos 2005:350). So when the researcher began this last stage, I was working on a loosely integrated document of 200 pages. Also at this point the major themes in each piece of research were already identified (See the chart above). One of the crucial components in this phase of coding is selection. So the researcher was supposed to select themes that are relevant to research questions, and these themes had to meet the criteria of being ‘heavily argued.’ Upon meeting this daunting task, he used a diagram which has 6 columns; each represented the following categories, Theme, Sub-theme, Direct quote or Paraphrase, Findings and References. This diagram helped to reduce the 200 pages of loosely integrated data to 33 pages of finely selected data. This data is the ‘grey matter’ for the next chapter. Below is a sample of the selective coding process:
Table 3.9
Sample of findings as a result of selective coding
Main Theme Sub- theme Direct Quotations/Paraphrases Findings References Stigma tisation
- ‘In my view I would say to a large extent the church has been an instrument of stigmatising persons with HIV, which is directly against its mission.’
‘I think the church has been stigmatising persons with HIV through the use of the Bible.
‘The current situation is that we are stigmatising the persons with HIV in our midst. I want to confess that this is a weakness on our part.’
‘Sometime the
preachers say if you do not respect yourself then your days to live will be few!’ ‘Sometimes, the preachers say those who are not settled, have been attracting the possibility to have the disease themselves. Lecturers , students and persons with HIV all agreed that there is stigmatis ation of the sick but the persons with HIV were more emphatic about this. FGMMU (Married) S.1 FGHIV/AI DS S.1 FGMISSI ONS. S.3 FGPLWH As (Single) S.1
In summary, the data analysis required 6 steps, beginning with reflection on my personal experience as informed by the body of knowledge. The next step was the formation of an interview protocol by adapting Gerald West’s method of Reader Response. Further, after the collection of data, the researcher transcribed the responses and coded them (open and axial coding). After that, he re-read the data to discover major themes embedded in the data (selective coding). At that point he began the actual process of interpretation by comparing empirical evidence alongside abstract evidence in the body of literature. And finally the researcher drew conclusions and gave suggestions for further research. The table below gives a graphic explanation:
Table 3.10
Further steps in data analysis
Step 1 Reflection on my own experience and analysis of the literature that led to and provided confirmation of the research questions and issues that resulted from the case study.
Step 2 Development of the interview/focus group questions based on the contextual Bible Study method pioneered by Gerald West. My social location and exposure to the study also contributed to the formation of the instrument.
Step 3 Line by line of transcriptions interviews/focus groups were coded and later led to the development of themes and categories.
Step 4 Further checking of emerging themes to determine overlaps was done by re-reading the transcribed data and listening to the audio files.
Step 5 Comparing categories with one another to develop central findings that represent the lived ‘interpretive experience’ of the sample.
Step 6 Drawing conclusions, recommendations, and suggestions for further research.