Inferring Candidates’ Issue-Specific Positions from Itemized Campaign Contributions Using Supervised Machine Learning

• In Event: JSS SESSION: Ideal Point Estimation

Brief Overview

To assess the informational value of itemized donations to voters--a key assumption made by the Supreme Court in upholding disclosure--we train a supervised model on itemized donations to candidates to predict how they vote across policy domains.

Authors

• Adam Bonica, Stanford University

• Zhao Li, Princeton University