Methods: Mind the Gap

Webinar Series

Randomization-Based Inference in Prevention Research

October 15, 2021
Michael Proschan
Michael Proschan, Ph.D.

National Institute of Allergy and
Infectious Disease (NIAID)  

View the Webinar

About the Webinar

Clinical trials are the gold standard of medical evidence, and randomization is the cornerstone of clinical trials. In this presentation, Dr. Michael Proschan shows how to conduct a test of an intervention effect based only on the randomization scheme employed in the study. This “randomization test” is valid with virtually no assumptions and can be applied regardless of the randomization method—simple, permuted block, even covariate- or response-adaptive randomization (but avoid response-adaptive randomization in clinical trials!). With the most standard randomization methods, the randomization test is nearly the same as a t-test if the sample sizes are large. Randomization tests can also be used to construct confidence intervals. If your trial requires major unplanned changes and you have not yet broken the treatment blind, guess what approach can save you? That’s right—the randomization test!

About Michael Proschan

Michael Proschan, Ph.D., is a Mathematical Statistician at the National Institute of Allergy and Infectious Diseases (NIAID) with 32 years of experience in clinical trials. He received his Ph.D. in statistics from Florida State University in 1989 and is a Fellow of the American Statistical Association. He loves teaching and has been an adjunct professor at Johns Hopkins University and George Washington University. He is co-author of two books, “Statistical Monitoring of Clinical Trials: A Unified Approach” and “Essentials of Probability Theory for Statisticians,” and just completed a third book, “Statistical Thinking in Clinical Trials” (in press).

Last updated on December 3, 2021