(This article was originally published at Realizations in Biostatistics, and syndicated at StatsBlogs.)
This article is very important. Yes, p-values reported in the literature (or in your own research) need scrutiny, but so does every step in the analysis process, starting with the design of the study.
Please comment on the article here: Realizations in Biostatistics
