Statistical Mistakes

Misuse of statistical inference in medical research

The blog

The importance of statistical inference has increased continually in medical research during the last 75 years and is still increasing with the methodological development of statistics and the technical development of the computational possibilities. Many medical researchers are, however, ignorant about statistical methodology, a major problem that has been documented and discussed repeatedly. Unfortunately, little progress has been made over the years. Serious mistakes appear frequently. Without them medical science would have advanced more. Safer and more effective means and procedures for preventing, diagnosing and treating diseases could have been found, and inferior ones been avoided.

The blog is written to facilitate search for information on the quality and accuracy of modern medical research, but also to encourage efforts to provide a more adequate statistics education of physicians and medical researchers. The literature presented in the blog is restricted to scientific publications (having a PubMed ID) describing the prevalence or incidence of actually occurring statistical mistakes. The blog also includes references to literature explaining the errors methodologically and to literature describing how the errors can be avoided. These pages represent reading suggestions. They are subjectively selected references, collected without any ambition of completeness.

Should you know of any paper that you think should be referenced here but isn’t, please send an email to the author with your suggestion.