Computational Social Science Research Notes

Research notes on computational approaches to legislative politics, covering spatial models for ideal point estimation (NOMINATE), NLP-augmented vote embeddings, and hierarchical models that combine roll call votes with bill text and speeches to study political polarization.

Year	Paper	Key Idea
1985	A Spatial Model for Legislative Roll Call Analysis	NOMINATE probabilistic spatial model for ideal point estimation
2015	Tea Party in the House: Legislative Ideology via HIPTM	Hierarchical Ideal Point Topic Model combining votes with bill text
2018	Party Matters: Enhancing Legislative Vote Embeddings	Neural architecture combining bill text embeddings with roll call votes

Computational Social Science

NOMINATE spatial plot showing Senate vote on Balanced Budget Amendment (1995) with legislators positioned on liberal-conservative dimension

A Spatial Model for Legislative Roll Call Analysis

This paper introduces NOMINATE, a probabilistic spatial model that recovers metric coordinates for legislators and roll calls from nominal voting data, demonstrating that a single liberal-conservative dimension explains the vast majority of Congressional voting behavior.

Computational Social Science

Visualization of party-based legislative embeddings

Party Matters: Enhancing Legislative Vote Embeddings

This paper introduces a neural architecture that combines bill text embeddings (CNN/MWE) with sponsor ideology metadata to improve vote prediction accuracy, particularly in out-of-session contexts where political dynamics shift.

Computational Social Science

Hierarchical Ideal Point Topic Model visualization showing political polarization

Tea Party in the House: Legislative Ideology via HIPTM

This paper introduces the Hierarchical Ideal Point Topic Model (HIPTM) to analyze the 112th U.S. Congress. By jointly modeling votes and text, it uncovers how Tea Party Republicans and establishment Republicans differ in both voting records and how they frame specific policy issues.