New PLOS ONE Statistical Reporting Guidelines

Read the complete Statistical Guidelines here

The cornerstone of research credibility and reproducibility lies in ensuring that research is conducted and reported in an accurate manner. These values are reflected in PLOS ONE Criteria for Publication #3 and #7 (https://journals.plos.org/plosone/s/criteria-for-publication) which require that experiments and statistics are described in sufficient detail to allow others to reproduce the work, and also cover the need to comply with appropriate community standards for reporting.

But how does one determine the information needed to allow research to be replicated while still ensuring the paper provides a concise and self-contained summary of the findings? This can vary from one discipline to another and reporting guidelines provide a useful framework for the information that is relevant for specific study designs. As part of our commitment to good reporting practice, PLOS ONE supports the use of reporting guidelines, for example, we require that authors adhere to CONSORT and PRISMA guidelines for reporting clinical trials and systematic reviews and meta-analyses, respectively (https://blogs.plos.org/everyone/2017/06/14/promoting-reproducibility/).

We are keen to hear feedback from the research community on steps we can take to support reproducibility efforts. Some months ago, Zen Faulkes, University of Texas Rio Grande Valley, noted that many journals do not have guidelines for authors on how to report statistics, and we listened. In response, and to help guide authors in the preparation of their submissions, our staff editors have collaborated with statistics experts on our Editorial Board to develop a set of Statistical Reporting Guidelines. These guidelines are divided into two sections, which cover general and specific guidelines for methods and results reporting, and aim to further enhance our requirement for high quality research conduct and reporting.

The Statistical Reporting Guidelines provide general principles that authors should consider in order to provide enough detail for another researcher to reproduce the study. Given the variety of study designs and statistical methods used by PLOS ONE authors, we are not able to offer detailed advice on reporting every statistical test. For that, we suggest that you consult your favorite statistics text or statistician.

The guidelines do provide instructions for reporting some specific statistics that we have noted commonly cause problems for authors. For example, in order to increase consistency, we have provided guidance for reporting p-values, regression results, and measures of variance.

We hope the new Statistical Reporting Guidelines will be the key to getting your research published soon. Happy reporting!