“Integrating Models of Innovation Adoption: Social Network Thresholds, Utility Maximization, and Hazard Models.”

This paper shows how two different theoretical approaches to model innovation adoption, network threshold modeling and random utility modeling, both result in discrete-time hazard models, a popular statistical modeling approach to study innovation adoption and diffusion. The paper emphasizes the need for careful situational analysis prior to model building and illustrates the point with the diffusion oftetracycline, the drug analyzed by Coleman, Katz and Menzel in Medical Innovation. The situational analysis identifies marketing effort rather social contagion as a likely driver ofadoption. A new statistical analysis ofthe Medical Innovation data set, expanded with newly collected advertising data, support this alternative explanation for the wellknown Medical Innovation results.