This article is only available in the PDF format. Download the PDF to view the article, as well as its associated figures and tables.
Recommended Procedures for Biomedical Research Articles
Most biomedical research requires some form of statistical analysis of the experimental data. Although physicians who read the results of such research should not concern themselves with the details of the statistical methods, these physicians should be given some protection against incorrect data analyses and misinterpretation of results.In STA 635, "Biostatistics," given to graduate statistics majors at the University of Florida, the instructor ("somewhat randomly") selected from ten medical journals 11 articles to be reviewed by the students to afford them practical experience in reading the journals. Major errors were found in all but one of these articles. Where errors occurred, the conclusions drawn by the authors are suspect.To strengthen review processes, we recommend the following (added) procedures for editing medical journals:Any article containing even the most elementary statistical procedures should be reviewed by a competent biostatistician (the simplest techniques,
Any article containing even the most elementary statistical procedures should be reviewed by a competent biostatistician (the simplest techniques,
Shuster JJ, Binion J, Moxley J, Walrath N, Grassmuck D, Mahnks D, Schmidt J. Statistical Review Process. JAMA. 1976;235(5):534–535. doi:10.1001/jama.1976.03260310048026