Faster Ideal Point Estimates with Less Wait for Political Scientists

Ideal points help explain voting and policy preferences. New methods tackle massive data challenges quickly.

➡️ New Fast Methods

• Standard EM algorithm extended to handle different outcome types (binary, ordinal, continuous)
• Added variational EM algorithms for dynamic/hierarchical ideal point models

➡️ Computational Advantages

• Produces identical estimates as standard MCMC but in minutes instead of days
• Works with various data sources including roll calls, surveys and social media text

➡️ Open-source Implementation Available