Statistical Errors in the Biomedical Literature
Tom Lang, Tom Lang Communications and Training International
 

Hypothesis testing began to be used in the biomedical literature in the mid 1930s, and studies of statistical flaws in the literature began to appear in the late 1950s and early 1960s.  Since then, hundreds of studies, mostly of leading general and specialty peer-reviewed journals, have found high rates of errors in articles reporting statistical analyses.  In this presentation, the history and nature of these statistical errors will be examined and some proposed solutions will be considered.