Enable AI Chat to speed up your analysis by assisting you in further rounds of coding.
With the help of AI Chat, your second and third rounds of coding can go much faster. Read these scenarios to see how AI Chat can help.
Scenario 1. You do a first round of loose coding. Then, use AI Chat to help clean it up.
We recommend that people start their first round of coding without overthinking it. That means, you create a few initial broad codes, and code any snippet that feels like it can loosely belong in it. Don't think too hard on whether or not your codes or snippets 100% fit together. Just code it, and move on. The point is to get familiar with your data and get a sense for where grouping might occur.
But after that first round of loose coding, you'll need to eventually organize and clean up your codes. That means refining those code definitions, organizing codes into top level codes and sub codes, and removing snippets that don't quite belong in each code.
In this article, we'll show how to use AI Chat to help with the steps after loose coding.
Step 1: Start with a loose first round of coding (Broad strokes)
Create a small set of initial codes. Read through each transcript and quickly apply snippets to the initial codes. Create new codes as you discover new commonalities. But keep your coding in broad strokes. Be more general with codes. It's better to have less codes with more snippets in each one, rather than many codes.
(Think of this as starting with a few large boulders, that you'll eventually break into smaller rocks)
Don't overthink it. Just read through all your data at least once with a quick pass.
Step 2. Use AI To help you break your codes into sub codes
Since you used broad strokes in your first step, use AI to help you break those codes into smaller subcodes. Subcodes can be nested under the top level codes and provide more nuance for your findings.
Click on AI Chat in the left hand menu.
Select a single code as a filter
Type in the following prompt: "Help me categorize each snippet into more nuanced sub-codes. Make sure each grouped category has commonalities. Give me a name and description for grouped categories, with snippet citations, presented in a table."
Evaluate the subcodes. You may agree or disagree with AI's choices here. If you do follow up and ask the AI to rework the subcodes. Give it feedback.
Once you are happy with the subcodes, create a new code and give it the name of the subcode AI provided
Click on the snippet citations from AI's answer and code each one
Now you have subcodes!
Step 3. Check which snippets don't belong in a code
When coding loosely and broadly, you might end up with snippets that don't quite belong in a code anymore. Use AI Chat to help determine which snippets to remove.
Click on AI Chat in the left hand menu.
Select a single code as a filter
Rank on a scale of 1-5 if each snippet belongs in this code with the code name "[COPY AND PASTE CODE NAME HERE]" and code description ,"[COPY AND PASTE CODE DESCRIPTION HERE", and provide an explanation. Give it to me in a table format.
Evaluate each of AI's answers
For each snippet you want to remove from our code, click on the snippet citation, then click on the code in the right hand menu to remove the code.
Want to learn end to end qualitative coding with AI?
Watch our 30 minute webinar where we break down the concepts of qualitative coding and how AI can be used to assist your process.