Key points are not available for this paper at this time.
The authors propose and calibrate a “split hazard” model that allows simultaneous modeling of observed heterogeneity with respect to both probability and timing of eventual adoption. Model validation with data for the diffusion of automatic teller machines (ATM) across 3689 banks demonstrates the robustness of the model framework to conceptual and distributional assumptions.
Sinha et al. (Sat,) studied this question.