Hosted by the Department of Biostatistics
Speaker:
Siyu Heng, PhD
Assistant Professor, Department of Biostatistics, NYU School of Global Public Health
Design-based inference offers a principled, model-free framework for trustworthy statistical analysis in causal inference, survey sampling, and missing data. However, its use in many modern applications has been limited by a fundamental obstacle: existing design-based inference methods typically assume that propensity scores (i.e., design probabilities) are known, whereas in observational studies, real-world surveys, missing data problems, and many other settings, propensity scores are usually unknown and must be estimated from data. In this talk, Dr. Heng introduces propensity score propagation, a general framework for valid design-based inference with unknown propensity scores. The framework uses a regeneration-and-union procedure to propagate uncertainty from propensity score estimation into downstream design-based inference without introducing super-population assumptions about the outcomes.
This event is open to the NYU Community (current students, faculty, and staff) for in person attendance. The general public will only be allowed to participate virtually.