I wanted to post an announcement and link to my new blog, “ep-ology” (the name is meant to cover the topics of epIDEMIology, epISTEMology, and eTHICAL pOLICY-ology). I have created that to be a home for longer analyses that do not fit well in this blog, as well as those that are not directly on-topic. I will continue to post here shorter entries that fit well, but my more in-depth analyses (even those about tobacco harm reduction) will appear over there.
The first of the latter is there. Following on Rodu’s recent posts about the misleading interpretations of the Deborah Winn dissertation data, particularly the nonsense claim that smokeless tobacco has been shown to cause a 50-fold increase in risk for oral cancer, I decided to post an analysis of some other things that appear in that data. I have published most of this before, but it was sufficiently obscure (and seems to no longer be available), so this is basically new.
In Part 1 I provide some background and outline the reasons that the Winn result is meaningless, even apart from the biased analysis. (For those who do not know, that paper is basically the only one that could be interpreted — if it were the only study that existed — as showing that some Western smokeless tobacco appears to cause oral cancer. This is a huge outlier from the rest of the science.) In Part 2, I report on what can be learned by reanalyzing the data, a smoking gun of highly-biased analysis.
I look forward to “seeing” some of you over at my blog.
– Carl V. Phillips