Summary
The episode works through four problems that students often find difficult in case-oriented qualitative analysis: telling a principled redefinition of the phenomenon from tuning, reading and minimizing a truth table, distinguishing necessity from sufficiency, and judging a decision tree's out-of-sample accuracy. Sarah and Kiffer apply analytic induction to invented cases from a community lunch program, including a case that forces a redefinition and a non-instance that shows why Robinson asked for cases where the outcome is absent. Sarah minimizes an invented truth table aloud, catches an illegitimate merge, and traces how the parsimonious solution relies on a logical remainder. The hosts then examine a trivially necessary condition, a held-out accuracy that matches the base rate, and the reuse of test cases during revision. They close by debating whether coding transcripts as zeros and ones disciplines interpretation or hides it, and they agree on a practical rule for tracing and stress-testing every coding decision.