Steve Ruberg
It is difficult to define the intention-to-treat (ITT) effect, and thereby the ITT estimand, since the use of the ITT approach can involve treating patients with many different treatments for many different time periods. Because of this mixture of treatment effects, the ITT approach is more of a...
Steve Ruberg
Even with the publication of ICVH E9(R1) and so may efforts globally for explaining and implementing the estimand framework, too often statisticians, and clinical researchers in general, focus their efforts on how to handle incomplete data, which is my term for the missing or confounded data...
Steve Ruberg
Randomization and complete data on the randomized experimental units are needed for cause-and-effect inference. What happens when that logic is broken? Introduction For those who follow this blog, you may have noticed a rather long hiatus in blog entries. I have been absorbed in writing a...