Theoretical neuroscience
Theory of spike initiation, sensory systems, autonomous behavior, epistemology
Editor
Romain Brette
Abandon Statistical Significance (2017)
Blakeley B. McShane, David Gal, Andrew Gelman, Christian Robert, Jennifer L. Tackett
2 comments on PubPeer arXiv: 1709.07588v2 arXiv: 1709.07588
Recently, there has a been a lot of discussion about issues of reproducibility in the biomedical and psychological literature. Some people argue that the threshold for statistical signifance should be lowered, say p = 0.01 instead of 0.05. This paper argues, and in my opinion rightly so, that the use of statistical signifance should be abandoned. One of the main arguments, which I also defended in a blog post, is that the null hypothesis is not credible. When any manipulation on a living being is performed, it is unrealistic to hypothesize that the effect will be exactly 0. It might be small, yes, but not exactly zero. And if it’s not zero, then with a sufficient number of observations there is 100% probability that you find a statistically significant effect, whatever threshold you use. The term “significance” is misleading; a ridiculously small effect would still be statistically significant, provided enough observations. It doesn’t prove much, apart from the fact that you are ready to sacrifice hundreds of animals just to get published by glamour journals. Conversely, finding that something is not significant means literally nothing: it could be that you just haven’t repeated the experiment a sufficient number of times – in fact it must be this interpretation, given that the null hypothesis is not realistic. So, I agree with the authors that statistical significance should be abandoned, in favor of more meaningful statistical measures (for example effect size).
Subjects
A peeriodical is a lightweight virtual journal with you as the Editor-in-chief, giving you complete freedom in setting editorial policy to select the most interesting and useful manuscripts for your readers. The manuscripts you will evaluate and select are existing publications—preprints and papers. Thus, a peeriodical replicates all the functions of a traditional journal, including discovery, selection and certification, except publication itself.
Why set up a peeriodical? The traditional journal has changed remarkably little in centuries and many people feel that scientific publishing is stuck in a rut, subject to a corporatist drift, and is not serving science optimally. The advent of preprints in many fields beyond those served by the ArXiv is liberating the dissemination of research, but most other journal functions have not been replaced effectively. Now you—all researchers—have the opportunity to select and certify research according to your own criteria. We expect peeriodical subject matters and editorial policies to be extremely varied. Some peeriodicals may wish to target narrow domains, while others will adopt a generalist approach. Some peeriodicals will be inclusive, focusing on discovery, whereas others may aim to enforce stringent quality criteria, prioritising certification. The point is that all approaches are permitted and supported—we hope you will innovate! You can create multiple peeriodicals. It will be users and readers who decide which peeriodicals they find useful and interesting. Users can sign up to receive alerts from any peeriodical they wish.
A peeriodical has one or more editors. Anybody can set-up a peeriodical and either operate it alone or invite colleagues to form an editorial board or community. The editors can select "manuscripts"—existing papers or preprints—to consider, either spontaneously or through suggestions from other researchers, including of course the authors. Note that there is no obligation that the manuscript be recent; for instance, we expect that some peeriodicals could focus on underappreciated classics. After all, predictions about scientific impact are generally more accurate for the past than the future. If the editors wish, they can solicit reviews for the manuscript via the Peeriodicals interface. Reviews will be published and the referees will have the option of posting anonymously or signing their review. Editors may decide at any time to accept, reject or comment on the manuscript, taking into account the comments received. They may of course suggest improvements to the manuscript or underlying study. If they justify their decision, their editorial decision will also be published.
How will Peeriodicals fit into the publishing landscape? We see them as a space without entry barriers in which researchers can innovate and explore new approaches to scientific dissemination, in parallel to the traditional publishing industry. There are related and complementary initiatives, notably the overlay journals promoted by Tim Gowers, exemplified by Discrete Analysis, but also Science Open Collections, PLoS Channels, the APPRAISE initiative and Peer Community in... Each of these projects has their own specificities and goals. Nobody yet knows exactly what the future will look like, but we strongly believe that we are about to experience a period of rapid evolution in the dissemination of science and we hope that Peeriodicals will inspire and help you to share your imagination and expertise with the whole research community.
For those starting a peeriodical, you will discover that the hardest part is building up an audience. Unfortunately, we can't yet guarantee you the exposure you would get from a paper in a glamour journal. Reviews with scientific content will be mirrored on PubPeer, offering an audience through the PubPeer browser and Zotero extensions. However, it will be largely up to you to run your publicity, most likely through social media. We are on Twitter (@PEERIODICALS) and will of course help out as we can.
Get started now by requesting an invitation with the link in the top right menu.