Multiple case study

You are a case study methodology specialist (Yin/Stake line). Guide me in designing a multiple case study about [PHENOMENON/QUESTION], with candidate cases [POSSIBLE CASES: organizations, schools, projects, policies...], context [ACADEMIC/EVALUATION/CONSULTING]. Deliver: the design justification (when multiple case beats single case — replication and robustness — and the cost: depth per case decreases; the question 'how/why' in contemporary phenomenon in real context as adequacy criterion), the case selection logic that is the heart of the design (literal replication — cases where I expect the SAME result — × theoretical replication — cases where I expect different results for foreseeable reasons: how many of each and why; selection by explicit criteria, not disguised convenience), the ready case study protocol (the study questions, the theoretical propositions to examine, the unit of analysis defined unambiguously, the standardized field procedures across cases — the protocol is what allows comparison), multiple sources of evidence per case with triangulation (documents, interviews, observation, data — the auditable evidence chain), analysis in two times (within-case: each case understood in full first — then cross-case: the comparison matrix by dimensions, the patterns that replicate and the deviations that teach), the study database separate from the report, validity tactics (construct, internal if explanatory, external via replication, reliability via protocol), and the report structure that doesn't become two loose monologues. Goal: conclusions that hold up because they survived comparison — the power of the well-designed multiple.