Tonight I watched the Delve video “Intro to Qualitative Data Analysis using AI Tools.” The video narrator stressed that there isn’t one way to “do” qualitative coding. They defined qualitative coding as “a process of taking unstructured data and systematically categorizing excerpts in order to find themes and patterns for analysis.” Qualitative coding is deeply personal since there are interpretations of transcripts. A code is a tag or label assigned to excerpts of text. There are multiple approaches for coding. The narrator noted that in their approach, familiarizing yourself with the data is the first step. They noted that you always need to verify AI’s outputs. Step 2 is to create codes. Research questions can lead to codes. Adding code descriptions is important to help manage and analyze the codes. You can also create codes based on your familiarization with your data (step 1). The three examples that the narrator shared for initial coding methods were:
- Descriptive code: topic summaries. Use nouns.
- In vivo codes: use participants’ language. Use “quotes”
- Process codes: Show action. Use verbs.
The next step is to take those codes and have AI apply codes. They demonstrated this with Delve Tool. The narrator noted that verifying the codes created by AI is critical. Chatting with the AI can help summarize and refine the codes you can include. The narrator described using AI as a sounding board or colleague. They described coding as an iterative process that is continuously refined. AI can be used to suggest topics for codes. The Delve Tool AI Chat (Beta) seems very easy to use! For example, in one case the chat created a table. The narrator used the AI Chat to group codes. Step 7 is to turn high-level codes into themes. Next, you can write a narrative using supporting codes and explaining your theme. AI can also write a finding section to “inspire you” to refine and create one. I appreciate how the narrator used Delve and also noted that there is an AI blog.
