2019 Atlantic Causal Inference Conference

  • Start Date: 05/22/2019
  • End Date: 05/24/2019
Location: 

McGill University

Description: 

The Atlantic Causal Inference Conference (ACIC) is an interdisciplinary conference designed to bring together researchers and students of causal inference methodology and application. Focused on statistical theory and methods (including biostatistics and econometrics), but with regular contributions from epidemiology and computer science, ACIC has existed in some form since 2005. Since then, it has grown into a small annual conference with established awards (Tom Ten Have Award) and competitions (the Causal Inference Data Analysis Challenge).

 

 

Causal Inference

 

The focus of this conference is on the theory, methods, and practice of causal inference. In this section we give a brief overview of some aspects of this interdisciplinary field.
Causal inference in the data sciences is a broad topic where an analyst seeks to combine external knowledge with collected data in order to learn whether one variable is a cause of another. Different frameworks (such as structural equation models, directed acyclic graphs, potential outcomes, and independence assumptions) have been used to structure and justify different approaches to causal inference data analysis. More recently, some of these theoretical structures were found to be partially or entirely harmonious and lead to common estimation approaches. However, additional debates are still raging.

 

Because the learning of causal relationships has a wide range of subject matter applications, various groups developed methods appropriate to their respective application areas. For instance, while causal inference in statistics was primarily based on the randomized experiment, instrumental variables were widely used in economics studies as the primary tool to establish causality. Subsequently, instrumental variables were incorporated in the analysis of randomized controlled trials with imperfect adherence and even in the analysis of observational data in pharmacoepidemiology.

Other Information: 

More information will be made available soon.

 

Details on the 2018 conference can be found here.