Evaluation without Experimentation

An overview of IPTW methodology

Mar 22, 2023 11:00 AM
Data for Progress Data Engineering Open Mic
Data for Progress Data Engineering Open Mic

In this four-minute lightning talk, I explain the motivation for using observational versus experimental methods when measuring the impact of GOTV (voter turnout) efforts. Next, I provide a high-level explanation of a specific method (inverse propensity of treatment weighting) and demonstrate the method’s ability to balance baseline traits and historical performance of a synthetic control population for credible effect estimation.

This talk is based on my related blog post and summarizes my work as a Bluebonnet Data Fellow with the Two Million Texans relational organizing campaign during the 2022 midterms.

Emily Riederer
Senior Analytics Manager