Building theory from the data up. Code, compare constantly, and keep sampling until new material adds nothing new.
Key thinkers: Barney Glaser, Anselm Strauss, Kathy Charmaz
The longer story
Grounded theory builds theory from the ground up, through systematic data analysis. No hypothesis going in; you discover what's there. Open coding breaks data into fragments; axial coding reassembles them around emerging categories; selective coding integrates into a theoretical framework. Theoretical sampling means collecting more data to develop categories until saturation. The goal is theory that's grounded in data, not imposed from outside. It's demanding—takes seriously the claim that you're generating theory, not just describing.