NISS co-sponsored the University of North Carolina's (UNC) 2017 Atlantic Causal Inference Conference (ACIC) held on the UNC campus, on May 23-25, 2017. The ACIC brings together statisticians, biostatisticians, epidemiologists, economists, social science and policy researchers to discuss methodological issues with drawing causal inferences from experimental and non-experimental data. Other sponsors at the conference were UNC Gillings School of Global Public Health, UNC Department of Biostatistics and SAMSI.
Defining Causal Inference, Fan Li, Associate Professor of Statistical Science at Duke University and a member of the 2017 ACIC planning committee, says, "Causal inference concerns designs and methods of analyses to evaluate treatments, interventions or actions in randomized experiments or observational studies. It is also known by other names in different disciplines, for example, comparative effectiveness research in medicine and in health care policy, and program evaluation in economics."
The interface of causal inference and machine learning/high-dimensional data is an emerging "hotspot" in both causal inference and machine learning. So, "with the aim to bring together leading researchers to present this cutting-edge technology, NISS facilitated a workshop titled "Causal Inference and Machine Learning/High Dimensional Data," adds Li.
The conference was bustling with 240 participants and included a series of workshops, short courses, and poster sessions. NISS awarded three presenters with travel grants, namely Stefan Wager, assistant professor at Stanford Business School; Alex D'amour, Neyman visiting assistant professor at UC Berkeley Statistics Department; and Alessandra Mattei, associate professor at University of Florence (Italy), Statistics Department.
- Wager talked about "Approximate Residual Balancing: De-Biased Inference of Average Treatment Effects in High Dimensions”
- D'amour talked about "Overlap and Deconfounding Scores in High Dimensions"
- Mattei talked about "Selecting Subpopulations for Regression Discontinuity Designs in High Dimensional Settings"
Two short courses were also delivered by world-class experts at the conference - one course on Precision Medicine Through Optimal Treatment Regimes delivered by Marie Davidian, Butch Tsiatis and Shannon Holloway from North Carolina State University, and the other course on Matching Methods delivered by Jose Zubizarreta from the University of Pennsylvania.
The conference was well received by participants, more so, because the interests in causal inference across varied disciplines - industry and government - is rapidly growing in the changing world of “big data” that we live in today, says Li. The 2018 Causal Inference Conference will be held in Pittsburgh in May 2018. ACIC continues to provide an excellent platform to bring researchers across many disciplines to come together and advance the research in causal inference.