This article develops prevention misalignment as an empirical diagnostic concept for crime-prevention research. Its central claim is that weak, mixed, or unstable preventive outcomes should not be treated as a single undifferentiated category of failure. Instead, prevention analysis should distinguish among design failure, implementation failure, and processual misalignment. Design failure refers to interventions whose preventive logic is intrinsically weak or implausible. Implementation failure refers to interventions that are reasonable in principle but insufficiently or inconsistently delivered. Processual misalignment refers to interventions that are serious in design and at least moderately credible in implementation, yet underperform because their operative leverage is directed at the wrong processual level, introduced at the wrong temporal stage, or dependent on institutional conditions that are absent or only partially reproducible. Building on the theorem-driven framework of crime as a mediated, gated, and path-dependent process, the article operationalizes prevention misalignment through eight diagnostic dimensions: problem representation, claimed intervention mechanism, actual processual target level, feasibility conditions, implementation seriousness, mechanism reproducibility, outcome pattern, and main failure diagnosis. The framework is applied through a comparative qualitative six-case design comprising Sluta skjut in Malmö, wider Swedish Group Violence Intervention implementation, gang-exit support in Sweden, youth recruitment into criminal networks in Sweden, London Shield, and Glasgow CIRV. The comparison identifies several forms of misalignment, including level misalignment, temporal misalignment, institutional misalignment, translation-induced misalignment, and compound misalignment. The article contributes a structured empirical strategy for diagnosing what has failed when prevention underperforms. Its purpose is not to replace implementation research or effect evaluation, but to add a process-sensitive diagnostic layer capable of distinguishing weak intervention models from weak delivery and from interventions whose mechanism package does not fit the level, timing, or institutional conditions of the problem. This article forms part of the Theorem-Driven Crime Process Research Programme and represents the principal empirical-diagnostic development of the programme’s prevention fit and prevention misalignment extension.
J. E. Fröderberg (Fri,) studied this question.