Duration Dependence, Functional Form, and Corrected Standard Errors: Improving EHA Models of State Policy Diffusion

Abstract

AbstractDiscrete event history analysis (EHA) is the analytic tool of choice for many scholars of policy diffusion across American states. Unfortunately, the policy diffusion literature largely ignores several important specification issues for EHA models: duration dependence, choice of functional form, and the computation of standard errors corrected for temporal and spatial dependence. We use data from Berry and Berry's (1990) seminal study of state lottery diffusion to demonstrate ways to deal properly with these issues.