Method-Conclusion Coupling
Analysis pattern clustering claims by method rather than conclusion
Method-conclusion coupling is an analysis pattern that clusters claims by the method that produced them rather than by the conclusion they reach. The insight is that different conclusions can share the same flaw, and identical conclusions can rest on very different grounds.
If three studies use the same flawed methodology and reach the same conclusion, that is not strong evidence; it is correlated failure. Conversely, if two studies use independent methods and reach the same conclusion, that is stronger than three studies using one method.
This pattern supports replication path separation: asking whether apparent agreement comes from independent evidence or from repeated use of the same approach. In Coscientist, it is implemented as a specialized agent for clustering claims by method.