Data and Causal Inference in Political Campaigns

For our February event, we're thrilled to have political scientist and political consultant Aaron Strauss, PhD, talk about two things that people in this town absolutely love. One of those things is political campaigns. The other thing is using data to estimate causal effects in subgroups of controlled experiments! Aaron will talk about how his experiences in high-stakes political campaigns apply to anyone designing real-world experiments where different groups may respond differently to different treatments. Probably with funny anecdotes.

Agenda:

6:30pm -- Networking, Empenadas, and Refreshments

7:00pm -- Introduction

7:15pm -- Presentations and discussion

8:30pm -- Adjourn for Data Drinks (Tonic)

Abstract:

Understand how political campaigns use data and causal inference beyond the media narrative. Explore the theory of electioneering with data, as well as how campaigns have advantages and disadvantages compared to their business counterparts. Delve into the details of heterogeneous treatment effects -- i.e., how to identify the voters most responsive to campaign contacts. Have the chance to ask big questions such as "Do any of these techniques make a difference on Election Day" and "Are campaigns invading my privacy?"

Bio:

Aaron Strauss (http://www.linkedin.com/pub/aaron-strauss-ph-d/1/1ab/614) is a freelance political data consultant who has been involved in three of the past four Democratic presidential campaigns and earned a PhD in political science from Princeton. As a practitioner, Aaron developed the first Google Maps overlay for political canvassers and helped introduce heterogeneous treatment effects into the Democratic campaign circles. As an academic, Aaron co-authored the first experiment to demonstrate the impact of text messages on voter turnout.

Sponsors:

This event is sponsored by Intridea (http://www.intridea.com/), Cloudera (http://www.cloudera.com/), Statistics.com (http://bit.ly/12YljkP), and Elder Research (http://datamininglab.com/).