Research
Articles
FREDERICK J. BOEHMKE,
University of Iowa
Approaches to Modeling the Adoption
and Diffusion of Policies with Multiple Components
Scholars have begun
to move beyond the dichotomous dependent variable—indicating whether
a state adopts a policy or not in a given year—usually employed
in event history analysis. In particular, they have devoted increasing
attention to the components of policies that states adopt. I discuss a
variety of estimators that have been employed to analyze the adoption
and modification of policies with multiple components, including various
forms of event history analysis, OLS, and event count models. With various
modifications, the researcher can estimate models that treat each component
as distinct, pool these models to leverage commonalities across components,
or treat the components as identical parts of the same process. Each of
these has its strengths and may be appropriate in certain circumstances.
Nonetheless, in the majority of cases, some version of event history analysis
for multiple or repeat failures is likely to be preferred. The different
approaches are illustrated by studying state adoption of various obesity-related
policies.
|
|
