Measuring policy diffusion with automated content analysis

Abstract

Policy diffusion means that policy choices in one unit (such as states, cantons, and cities) are influenced by the policy choices of other units. This idea has been studied extensively in several social sciences and, within political science, in subfields such as international relations, American federalism, and public policy. While scholars have demonstrated convincingly that policy diffusion is a real and important phenomenon, much less is known about why and exactly how policies diffuse. We argue that this condition is due to the inherent limitations of existing research designs. This paper puts forward a new approach based on automated frame analysis. Theoretically, the paper focuses on how the perception of policy problems and solutions changes as a result of the adoption of policies elsewhere. We apply this approach to the spread of smoking bans among US states and Swiss cantons. Preliminary results from an analysis of thirteen American newspapers shows that media coverage intensifies when major legislation and illustrates how it prioritizes different consequences of smoking bans, such health, economic fallout for businesses, and enforcement issues.