Through all the phases of coding, I made 3 distinct types of memos, which were reflections and notes to be elaborated on when creating and describing memos. First were the gerunds and adjectives that made up my focused codes, which I stored in a separate Excel file (for these types of memos, I found the separate file easier to navigate than any of NVivo’s functions). Charmaz explains that focused codes “pinpoint and develop the most salient categories in large batches of data.” (p.46), so this was an essential step to ensure I was doing this. I used NVivo’s ‘Memo’ function for general thoughts, especially when thinking of preliminary categories and points to describe them. ‘Annotations’ are created directly in the data (contained in each initial and focused code file), and I used these for memos that arose in response to more specific segments, such as examples I could use when describing my categories. Writing the memos for both creators, I found they had many similar approaches to their content, with the differences being the execution. Once I finished my focused coding for both Molyneux and ContraPoints, I merged the two projects so all the videos were in one place, and combined all ‘Memos’ so I could explore potential categories that described both of them.
I arrived at five categories, which I then tested on my codes for each video. To do this, I created ‘Nodes’ for each category with the preliminary titles, and used coloured highlights so they could easily be differentiated: facts vs feelings (blue); characters in the narrative (green); the creator as a character (purple); sarcasm and satire (red); and flow (orange). Soon into the categorisation process I found a need for a sixth – ‘setting the scene’ – to describe the creators’ tactics in guiding the audience towards a certain understanding, similar to how they sway their emotions and portray characters, but for the overall sense of the story rather than these specific components. The other categories I had considered in the ‘Memos’ were merged to broader categories, i.e. ‘humanisation’ was a component of ‘characters in a narrative’, ‘truth’ was a component of ‘facts vs feelings’, and ‘identifying the target audience’ and ‘weighted statements’ both contributed to the categories for characters.
I could easily sort each code into at least one category for each of the videos, and found myself with six full categories which accurately illustrated all aspects of both creators’ approaches to storytelling that I had identified. To decide where each code fit the best, I asked myself for each what is the strongest point I can make for this code in the context of one category. The easiest was for ‘flow’, since this almost exclusively related to the codes I created to describe the tone of the speakers. There were some overlaps with ‘the creator as a character’ and ‘characters in the narrative’, since sometimes the creator identified with the characters being described, so I could speak about one code in two different contexts. Similarly for ‘facts versus feelings’ and ‘setting the scene’, there were a few tough decisions on what I felt was an offering of evidence to strengthen the argument, compared to what was an aid for the audience to understand a broad context. In some of these cases, I created two codes describing different aspects of the same sentence, such as a statement of fact that was loaded with implications about how you should understand that fact, so I could enter one code as ‘facts versus feelings’ and the other as ‘setting the scene’.