SCI Shortcourses will take place at the 2024 American Causal Inference Conference (ACIC), held in Seattle, Washington, on Tuesday, May 14th.
There will also be virtual Shortcourses held following the ACIC.
RATES
Student Member – $80
Non-Student Member – $110
Student Non-Member – $180
Non-Student Non-Member – $210
Short Course Workshops
Unlocking the Mysteries of Mixed Exposures: Targeted Learning for Robust Discovery and Causal Inference in Epidemiology
DATE: May 14, 2024
TIME: 8:30 A.M. – 12:30 P.M.
INSTRUCTOR:
David McCoy, University of California, Berkeley
DESCRIPTION:
| In epidemiological studies of high-dimensional data with mixed exposures (such as those involving air pollution, pesticides, pharmaceuticals, or nutrition), researchers face the daunting challenge of unraveling nuanced interactions, discerning susceptible subpopulations, and pinpointing optimal exposure thresholds for regulatory purposes. Standard methods struggle with the sheer number of potential interactions and lack the flexibility to discover nuanced (synergistic or antagonistic) relationships, which are intrinsic to mixed exposure studies. Read more |
A Hands-on Introduction to BART, BCF, and Other Bayesian Tree Ensembles for Applied Causal Inference
DATE: May 14, 2024
TIME: 8:30 A.M. – 12:30 P.M.
INSTRUCTOR:
Jared Murray, University of Texas at Austin
DESCRIPTION:
This course offers a code-and-data-first introduction to Bayesian tree ensembles including Bayesian Additive Regression Trees (BART) and Bayesian Causal Forests (BCF) for applied causal inference in experimental and observational settings. These methods are routinely among the top performers in empirical studies of machine learning methods for estimating causal effects – including the ACIC data competitions – and feature prominently in high-profile applications, particularly in the social sciences. Read more
Statistical methods for electronic health record data
DATE: May 14, 2024
TIME: 1:00 P.M. – 5:00 P.M.
INSTRUCTOR:
Susan Shortreed, Division of Biostatistics, Kaiser Permanente Washington Health Research Institute; Department of Biostatistics, University of Washington.
Yates Coley, Division of Biostatistics, Kaiser Permanente Washington Health Research Institute; Department of Biostatistics, University of Washington.
Jennifer Bobb, Division of Biostatistics, Kaiser Permanente Washington Health Research Institute; Department of Biostatistics, University of Washington.
DESCRIPTION:
Electronic health records (EHRs) are increasingly being used in causal inference and other clinical research to advance medical science, epidemiology, and public health, but implementing studies that leverage EHR data must address several unique challenges and potential sources of bias. In this course we will provide an overview of the strengths and limitations of using data gathered from EHRs for causal inference, specifically to estimate the effects of interventions or exposures on health outcomes. Read more
Introduction to data-driven selection of causal graphical models ("causal discovery")
DATE: May 14, 2024
TIME: 1:00 P.M. – 5:00 P.M.
INSTRUCTOR:
Daniel Malinsky, Department of Biostatistics, Columbia University
DESCRIPTION:
| This course focuses on methods for causal discovery a.k.a. causal structure learning or estimating causal graphical models from data. We begin with an overview of the classical algorithmic strategies (constraint-based algorithms, score-based algorithms) for learning causal DAGs. Relevant graphical and statistical concepts will be discussed, including Markov equivalence, faithfulness, conditional independence testing, consistency of the BIC score for selection, and theoretical properties of methods such as the PC algorithm and GES algorithm. Read more |
Highly Adaptive Lasso and Adaptive TMLE in Causal Inference
DATE: May 14, 2024
TIME: 1:00 P.M. – 5:00 P.M.
INSTRUCTOR:
Mark Van Der Laan, University of California Berkeley
Rachel Phillips, University of California Berkeley
Lars Van Der Laan, University of Washington
DESCRIPTION:
The field of Targeted Learning (TL) involves the construction of targeted plug-in estimators of specified estimands using super learning, the highly adaptive lasso (HAL) machine learning algorithm, and targeted maximum likelihood estimator (TMLE), where the latter represents a bridge from machine learning to statistical inference. In this short course, attendees will learn about the HAL estimator of a target function such as a prediction function, conditional treatment effect or conditional density, and its software implementation with the hal9001 R package. Read more