This article presents a general mathematical model of the diffusion of innovations, which incorporates mass media and interpersonal influence. The model is applied to three classic diffusion data sets: (a) use of hybrid corn, (b) knowledge of Eisenhower's stroke, and (c) doctors’prescription of a new drug. Nonlinear regression is used to estimate the mathematical model. The results show that diffusion of hybrid corn occurred via interpersonal influence, whereas the diffusion of knowledge of Eisenhower's stroke occurred via the mass media. For the diffusion of the new drug, the model shows that doctors who subscribed to few medical journals learned about the drug primarily through interpersonal influence, while doctors who subscribed to many medical journals learned about the drug through both mass media and interpersonal channels. Policy decision-makers can use diffusion models to (a) evaluate the effectiveness of media versus interpersonal campaigns, (b) make comparisons between subgroups, and (c) evaluate the effect of a policy.