The construction of interest communities: Distinguishing bottom-up and top-down models

Two theories are now used to explain the size and composition of interest communities: Olson's (1965, 1982) bottom-up collective action model and Gray and Lowery's top-down ESA model. Unfortunately, one of the common ways of testing these models-aggregate-level analyses of interest populations-cannot easily distinguish between them. Empirical evidence supporting one model will almost inevitably support the other. We explain why this is so, illustrating the problem with data on state populations of health interest organizations. We also discuss how the two models might be better distinguished.