This article introduces a novel perspective on policy diffusion, shifting focus from adoption to the crucial yet understudied issue-definition stage.

New Approach: The authors propose examining how policies are framed during definition using topic modeling, rather than just when they are adopted. This method analyzes media coverage and identifies distinct ways policies are presented before decisions are made.

Smoking Restrictions Case Study (United States): Using a unique dataset of over 52,000 newspaper paragraphs from U.S. states between 1996-2013 covering smoking restrictions, the study investigates how issue framing influences policy outcomes within and across governments.

Key Findings:

The analysis reveals two distinct types of frame influence:

* Concrete Implications Frames: How policies affect people or businesses are strongly predicted by prior state adoptions.

* Normative Justifications Frames: Why a policy is morally right or aligns with certain values appear independent of previous adoptions.

Methodology: The research employs topic modeling (NLP technique) on newspaper text, analyzing 49 U.S. states over an extensive period to identify these patterns in issue framing during the diffusion process.

This study demonstrates that understanding media frames is essential for comprehending policy decisions and provides a replicable framework for examining policy diffusion across various contexts.