Method Hierarchies in Clinical Epidemiology

Accumulating evidence from diverse fields of inquiry suggests the existence of method hierarchies, where criteria employed by the same epistemic agent constitute a certain preference hierarchy. In this paper, we illustrate the phenomenon of method hierarchy by discussing several prominent studies in clinical epidemiology of coronary artery disease. The current “gold standard” in clinical epidemiology is the randomized controlled trial (RCT) method. Yet, in the absence of studies that satisfy the strict requirement of the RCT method, clinical epidemiologists often relax the requirements of double-blinding, complete follow-up, no treatment switching, and/or randomization. Instead, they sometimes employ less stringent requirements, such as the requirement to account for the potential imbalances between groups through statistical models. This suggests the existence of a certain method hierarchy. However, it is unclear how method hierarchies are to be conceptualized and documented. Specifically, it remains to be seen whether a method hierarchy is best understood as being composed of individual employed methods or as a single composite method with a complex system of if-s and else-s. Suggested Modifications [Sciento-2019-0013]: Accept the existence of method hierarchies. Accept the following definition of method hierarchy: Method Hierarchy ≡ a set of methods is said to constitute a hierarchy iff theories that satisfy the requirements of methods that are higher in the hierarchy are preferred to theories that satisfy the requirements of methods that are lower in the hierarchy. Accept the following question as a legitimate topic of scientonomic inquiry: Conceptualizing Method Hierarchies: should we conceive of a method hierarchy as being composed of individual employed methods/requirements, or should we think of it as constituting one composite method with a system of if-s and else-s, and-s and or-s?