01
Senior Honors Thesis · 2026
Method
- Constitutional Vacuum Diffusion
- Computational Text Analysis
- TF-IDF Vectorization
- OLS Regression
- Policy Diffusion Theory
Networks of Influence: Legislative Text Diffusion in State Abortion Policy After Dobbs
This thesis examines how abortion-related legislation diffused across U.S. states in the first full legislative year after Dobbs v. Jackson Women’s Health Organization (2022). Drawing on a corpus of 188 enacted bills across 46 states, it uses computational text analysis—TF–IDF cosine similarity computed across 104,445 unique. n-grams—together with bill-level and dyadic regression to identify the structural predictors of textual convergence.
Key Findings
Three findings invert conventional expectations. First, protective coalitions, not restrictive ones, deployed standardized legislative templates most aggressively in 2023: protective bills are significantly less novel than neutral bills (β=−0.076, p= 0.022), while restrictive bills are statistically indistinguishable from the neutral baseline. Second, the highest textual convergence across state lines coexists with a near-complete absence of direct PAC contributions between coordinated legislators, indicating that diffusion operates through informational rather than financial channels. Third, at the state-pair level, shared advocacy organization co-presence (PAC Jaccard β = +0.039, p < 0.001) and shared protective coalition membership (β = +0.037, p<0.001) predict textual similarity while geographic contiguity is null (β = +0.003, p= 0.734).