The Centre for the Study of Democratic Citizenship presents:
Zachary Markovich (Massachusetts Institute of Technology)
Bundled Causal Inference
When: Thursday, February 24th at 1 PM.
Zoom link / RSVP: Contact vincent.arel-bundock@umontreal.ca to get the link
Abstract:Bundled variables are ubiquitous in political science research. Traits as wide ranging as race, democracy, and ideology can all be represented as high dimensional vectors. However, the conventional tools of causal inference have been developed with univariate treatments or low dimensional conditioning sets in mind. This talk will introduce a framework for conducting causal inference with such bundled variables based on a bounding approach, and apply it to model the effect of campaign spending on election outcomes and the effect of race on standardized test scores, demonstrating the utility of this tool for applied researchers.Zachary Markovich is a PhD candidate in the Department of Political Science at the Massachusetts Institute of Technology, focusing on the intersection of machine learning and causal inference with applications to electoral politics and public opinion. He has also studied the effects of minimum wage increases on voter turnout, electoral volatility, and the effect of conjoint analysis on social desirability bias.
See more Montreal Methods Workshops events here.