Here comes the fun part - starting the detective work! This process is like a thrilling game of hide-and-seek with the text. You begin by doing a few 'scouting runs', reading the text multiple times to get familiar with it. Think of it as the first time you enter a haunted house; you'll notice some creepy things immediately but others only pop out after a few rounds. These 'creepy things' are your initial, low-level themes - ideas or patterns that keep appearing in the conversation. You highlight these and add notes in the margins of your text.

To be extra sure about your findings, sometimes a fellow detective (independent coder) is brought in to verify your themes. Just like how you'd take a friend along on a ghost hunt to make sure you're not imagining things!

Now, the themes you've found are like puzzle pieces, and you need to assemble them to see the big picture. You start grouping related low-level themes into broader, high-level themes. It's like sorting out Halloween candy - deciding if KitKats, Snickers and Mars Bars all fit into the 'chocolate' category.

The catch is, these categories or high-level themes require some interpretation. To make sure you're not creating a 'chocolate' category where one doesn't exist, other researchers might check your groupings. It's like having a friend verify that yes, indeed, these candies are all chocolate-based!