Suppose there is something you need to estimate or a hypothesis you need to test but due to ethical, technological, or practical limitations, you simply cannot collect the data you need. Let’s be honest, it’s probably because your paper is held up in peer review because a reviewer is asking about some potentially confounding factor that is impossible to measure. Regardless of why, what can you do about it? Besides giving up or making some unverifiable assumptions, are there rigorous solutions for statistical inference with limited data?

Presenter: Justin Silverman  |  College of IST