During the last 25 years I have reviewed over 6000 research reports submitted to more than 75 different scientific journals, most of them medical. Some of these reports have presented reliable and groundbreakning research results, but far too many have just represented a serious waste of resources that could have been used to find better medical treatments, reduce suffering, and prolong life.
This experience has changed my life; trying to identify empirical support for presented findings but finding misunderstandings, inconsistencies, and obscurities, and trying to explain the discovered problems to mostly non-statistically oriented authors, have developed me as a statistician, and statistics is a difficult subject.
The experience has also made me sceptical. The recent irreproducibility crisis in medical research did not come as a surprise, and I expect that it soon will spread into many other scientific fields. An inadequate statistical education in combination with a rapid development of computational resources and a widespread "publish or perish" culture, have not only resulted in frequent mistakes and errors; an even worse problem is a systematic exaggeration of the presented findings and the diminishment of their limitations and uncertainty.
There is only solution to these problems, a better understanding of fundamental methodological principles among authors and readers. To that end, I have started this blog with comments inspired by my daily reviewing. These comments do not reveal any content of the reviewed reports; manuscripts submitted for publication are confidential.