If you are like most authors, when submitting a paper to a journal you may usually consult the instructions for authors and peruse a few articles published in that journal to see which format is preferred for reporting statistics and, in the case of hypothesis testing, the accompanying P values. What may have struck you over the years is that, with the exception of a few journals with very specific guidelines as to when and how to report P values, nothing specific to reporting statistics or P values is included in the author instructions. Paging through the issues of the same journal for a year often reveals a variety of formats. This conundrum finally prompted this statistician to conduct an Internet search using the key words “reporting p-values.” The references found there in combination with a widely recognized reference suggest that there is “almost” consensus about some guidelines for reporting P values and a total lack of consensus for others. The purpose of this article is to discuss these guidelines, where you do and do not find consensus, and how you decide to report P values in your own work.
Kay Savik, MS, Biostatistician and Research Fellow, School of Nursing, University of Minnesota, Minneapolis.
Correspondence: Kay Savik, MS, School of Nursing, University of Minnesota, 5-160 Weaver-Densford Hall, 308 Harvard Street SE, Minneapolis, MN (e-mail: firstname.lastname@example.org).