June 2, 2021 | Hosted via Zoom
Causal Inference for Social Impact
Don’t Miss Out
Time & Location
Hosted via Zoom
About the Event
Causal questions are at the core of policy making and implementation of practices that profoundly affect people’s lives. Can limits on prescriptions curb opioid use? What is the effect of changing bail guidelines on crime? Can summer school repair some of the learning loss experienced by students who lost school time due to COVID? Answering these kinds of questions is complicated by challenges that arise when we develop research questions, design studies, gather data, and disseminate results in ways that involve and can directly impact the lives of the populations we are trying to support.
This event is an opportunity to start a conversation about how we, as causal inference researchers, can navigate the ethical, logistical, and methodological complications that we encounter on the front lines of causal inference. What ethical considerations should be considered that go beyond standard IRB protocols? How do we gain the trust of stakeholders and participants? What designs can accommodate the practicalities of existing practices? How can we ensure that we are doing more good than harm, particularly when we are focused on vulnerable populations?
We start the afternoon by highlighting the work of five causal inference scholars who are working at the intersection of causal inference in society. They will discuss topics ranging from sexual assault prevention, algorithm-assisted human decision making, experimentation in political polarization, treatment regimes for organ transplant, and randomized experiments for privacy.
The day will then culminate in an invited panel with three top causal inference researchers: Susan Dynarski, Chanelle Howe, and Eric Tchetgen Tchetgen. These scholars will discuss their own projects but also engage in a shared conversation moderated by Jake Bowers about the challenges and opportunities that emerge when trying to answer causal questions that have the potential for societal impact.
Event Schedule The schedule of the afternoon will be as follows (times are Eastern):
1:00 Event begins
Welcome from SCI leadership
1:10 Highlighted speakers
Mike Baiocchi: “Challenging causal inference problems from the front lines of sexual assault prevention:
Manjusha Kancharla: “Robust Randomized Experiments for Causal Effects Under Privacy”
Alex Volfovsky: “Online experimentation for studying political polarization”
Grace Lyden: “Transportability of causal inference under probabilistic dynamic treatment regimes for organ transplant”
Kosuke Imai: “Experimental Evaluation of Algorithm-Assisted Human Decision-Making: Application to Pretrial Public Safety Assessment”
3:40 Invited Panel
Jake Bowers, Political Science, University of Illinois
Eric Tchetgen Tchetgen, Statistics, University of Pennsylvania
Susan Dynarski, Economics and Policy, University of Michigan
Chanelle Howe, Epidemiology, Brown University
5:15 Special announcement about data competition
5:20 Closing remarks
5:30 Event ends