3.5 Analysis of interview transcripts
3.5.2 Analysis using coding techniques
The coding strategy was decided upon before the start of the pilot interview, as indicated in the research timeline in TABLE 6 and TABLE 7 starting on page 66. The coding process was guided by coding techniques recommended for grounded theory methodology by a leading world expert on coding techniques, Professor Johnny Saldaña (Saldaña, 2013). Saldaña’s coding techniques were more extensive, clearer and better argued (Saldaña, 2013) than the coding techniques by Strauss, one of the inventors of grounded theory (Strauss & Corbin, 2008). Saldaña recommended first and second cycles of coding techniques for grounded theory methodology and they were all used in the current study except for the Longitudinal
and Elaborative coding techniques (Saldaña, 2013). Longitudinal coding would have required a series of interviews with each participant and longer timescales beyond the time limits of the EdD programme. Elaborative coding was not appropriate because it requires further
exploration of an existing substantial theory, of which there was none available. The first cycle of coding techniques used in the study were the Descriptive, In-vivo, Process, Initial, Versus, Evaluation and Causation coding techniques, while the second cycle included Pattern, Focused, Axial and Theoretical coding techniques.
I created a table with 13 columns and probably about 800 rows spread across several A3 pages to assist the coding analysis for each participant interview. Each coding analysis sheets for each participant were printed on differently coloured papers to help me more easily identify
between the sheets from different participants during analysis. A copy of the table showing the first and second cycle of coding techniques used are shown in TABLE 3 and TABLE 4 below, where TABLE 4 is a continuation horizontally of TABLE 3. The table has been split into two to
make the font bigger on an A4 paper.
Table 3: First Cycle of Coding Techniques Used
Transcript First Cycle of Coding Techniques Used
Descriptive In-vivo Process Initial Versus Evaluation Causation
Meaning of different fonts Italics - my personal comments) In Bold - Significant quotations Describing what is happening Peculiar words in the interview Describing the activity occurring and usually ending in ‘-ing’ Breaking into discrete parts & exploring their relationships Conflicting concepts/people /system, processes/pheno menon +ve or –ve evaluation of a topic, subtopic, recommend ation What is causing the phenomeno n 1st Sentence 2nd Sentence 3rd Sentence Etc.
Table 4: Second Cycle of Coding Techniques Used Transcript First Cycle of
Coding Techniques Used Concepts from Literature
Second Cycle of Coding Techniques Used
Pattern Focused Axial Theoretical
Meaning of different fonts Italics - my personal comments) In Bold - Significant quotations Identifying patterns in the 1st cycle of coding The most salient codes/ codes that make the most analytic sense Stitch together scattered codes around central themes Umbrella code to accommoda te all other codes 1st Sentence 2nd Sentence 3rd Sentence Etc.
A transcript for a participant was copied and pasted into the first column in such a way that each sentence or paragraph was placed in each row and permitted particular coding
techniques to be applied to each sentence or paragraph. The next seven columns were for the first cycle of coding, while the last four were for the more advanced second cycle of coding techniques. There was a column, between the first and second cycles of coding techniques, for concepts mentioned by the interviewees related to concepts in the Literature Review of Module 5 in APPENDIX 29 from page 299, the Literature Review in Chapter 2 from page 7 or concepts which required a fresh literature exploration. Some sentences or paragraphs in transcripts were coded with more than one code, but some had none. Some codes were allocated to a number of sentences or to multiple paragraphs.
Themes and concepts that were related were connected with pencil or pen drawn up lines and memo comments were written on the coding analysis sheet.
FIGURE 7 below shows the coding analysis sheet of the first participant as an example.
Figure 7: An example of memo notes and lines drawn on a coding analysis sheet
The emerging themes were then drawn into a coherent and summarised memo diagram for each participant and the themes were underpinned by supporting quotations from the particular participant. An example of a summarised memo diagram is FIGURE 8 on the next page.
Figure 8: An example of a summarised memo diagram
NB: The diagram was printed on an A3 paper size and explains why the font is too small. The sentences not in bold were the supporting quotations from the particular participant.
The eight summarised memo diagrams were useful in finding patterns across the eight participants and were aggregated to make a master summarised memo diagram illustrating the major pedagogical themes of the eight transcripts. The radical constructivism epistemology allows the knowledge of individuals to be ‘taken-as-shared’ knowledge (Cobb, Yackel, & Wood, 1992).
Strong strands of pedagogical themes emerged after analysing the third participant onwards and the data was saturated from the sixth participant onwards, i.e. no new pedagogical concepts arose after the sixth participant. Initially, I was hoping to interview about 15 participants in the Research Proposal in APPENDIX 31 on page 359, since it was the point at which I had expected to see data saturation of the concepts to emerge. The major themes that emerged from the master summarised memo diagram were the three Pedagogical Backdrop (starting from page 84), Pedagogical Approaches (starting from page 87) and the Pedagogical Timings (starting from page 99) of the Findings Chapter. The major themes were grounded and anchored in the interview transcripts by the location and compilation of 142 supporting quotations that have been placed throughout the Findings Chapter and assisted in preserving
the thoughts of the participants during the analysis stage, as recommended by Charmaz (Charmaz, 2014). The coding analysis stage took about 300 hours.
I seriously considered using software for the coding analysis as an alternative to conducting manual coding analysis. I attended a formal training session on NVivo software for managing the grounded theory, (Nvivo training booking is in APPENDIX 8 on page 236). I also watched several training videos on Nvivo and the most notable was (Rowe, 2014). I, however, preferred the paper-based approach because I could more easily put my own annotations and arrows all over the pages to connect themes and I felt that the software restricted the connection of some ideas and themes because it was not programmed to do so.