It has been increasingly clear that most published research findings are either false (1) or not reproducible. The prevalence of irreproducibility may be up to 90%. A recent analysis suggested that the lack of reproducibility in biological research could cost up to $28 billion per year (2). Since becoming an editor of Journal of Bone and Mineral Research, I have taken the issue up with the Editorial Board. I thought that we should do something in our part to improve the reproducibility of science. I have been assigned to write the editorial in which we set out new guidelines for data reporting and statistical analysis (3).
The increasing concern about the lack of replicability of published research in the biomedical field has produced some initial consternation in the research community and in the public at large, but it has also fueled robust discussions among journal editors, funding agencies, academic institutions and researchers on how to address the problem and improve the quality of scientific publications. The editorial leadership of the Journal of Bone and Mineral Research, in concert with our sister journal, JBMR Plus and the ASBMR Publications Committee has developed new guidelines for reporting and analyzing research data with the purpose of improving rigor and transparency of published material, thus minimizing reproducibility problems. The new guidelines have been developed by two task forces appointed by the Publications Committee, with input from the Editors. Members of the task forces were chosen among the JBMR / JBMR Plus Editorial Board membership, and their final recommendations were unanimously approved by the Publications Committee. The new guidelines are now incorporated in the instructions for authors for JBMR and JBMR Plus; reviewers and Editors will be asked to verify that submitted manuscripts comply to the new instructions. This editorial introduces the changes made, provides rationale for specific choices and offers some advice, especially about statistical analysis. We focus on five major key areas: (i) Study design; (ii) Reporting of genetic studies; (iii) Presentation and analysis of pre-clinical studies; (iv) Fundamentals of statistical analysis; and (v) Interpretation of P values.
As with any changes, there will be some initial hurdles to overcome and details to settle down, but we are certain that we will be able to achieve the goal of having all published content compliant to the new guidelines by early 2020. Some may find such detailed instructions too prescriptive and cumbersome. However, ensuring transparency, integrity and rigor in published research is paramount to the continued success and stature of our journals. We are confident that our readers and the musculoskeletal field in general will be appreciative of the effort, and our authors will quickly learn how to navigate the process.
(1) See Ioannidis, PLoS Med. 2005;2(8):e124.