Summary
The episode works through the move from a set of codes to an interpretation, using an invented study of Saskatchewan grain farmers in which a summary of codes is rewritten as a claim that could be wrong and then tested with the cases that would complicate it. It examines how taxonomies and typologies fail when categories sit at mixed levels of abstraction, when two dimensions are coded from the same passages, or when cells hold only one or two cases. It considers what a single interview can show about a cyclical process model, and how repeat interviews, diaries or longitudinal survey data such as the Canadian Longitudinal Study on Aging, collected every three years, could complement it. A harder worked example tests an invented concept map whose boxes are all well grounded, and shows that one arrow has no supporting passages, another mostly runs the other way, and the main pathway is a composite that only one participant described. The hosts close with a disagreement about triangulation and Richardson's crystallization and agree that a study should report each check it ran and what that check changed.