The Problem With Fragile Results

One of the goals of scientific literature, including that published within the Journal of Arthroplasty, is to help improve our arthroplasty profession and the care we provide to our patients. Publishing an article is challenging – it requires a great idea, a solid research design, as well as execution and writing in a manner that allows for the reader to correctly interpret the results within the context of their clinical practices. It should also help direct future research endeavors. It is an arduous process and one that is often followed by a feeling of success when authors discover that their primary outcome is significantly different from the control group. In a properly powered study, this indicates that if the study was run 100 times, the difference found would occur in 95 of those studies. It is not a perfect statistical test and there has certainly been much written in other journals about its utility and flaws. Still, “reaching statistical significance” continues to be the main goal of many researchers when conducting randomized controlled trials and this finding continues to influence the reviewer’s decision when determining whether to accept a paper for publication. Despite the aforementioned reliance on “statistical significance”, we are constantly reviewing our acceptance policies at Journal of Arthroplasty. Occasionally, an article comes along that highlights an important facet of our editorial responsibilities.