I met a final year PhD student once, who told me a sad story. His supervisor had a plausible idea that exercise reduces the chances of developing bowel cancer. To test the hypothesis, the student made a transgenic mouse with an increased incidence of bowel cancer and got the mice to run (or not run) on hamster wheels. Unfortunately, there was no correlation between the amount of exercise and disease occurrence and progression and the student was unable to publish his results.
It has always puzzled me that while it is relatively easy to publish positive findings (A correlates with B), it is almost impossible to make public the opposite – C doesn’t correlate with D – on the same topic, despite carefully designed and meticulously executed experiments and their brilliant analysis.
I think, at least in part, this happens because nobody, least of all reviewers, likes a story without a happy ending. As a result of this, many a PhD student spends at least three or more years testing out a good idea discovering that no, it doesn’t work like that, and finishing his or her PhD without a publication. This means that, just as he is getting his degree, on the other side of the globe a new PhD student may be starting a similar project due to the negative results of the first study not being published. This repetition is a waste of funding, not to mention the negative effects it has on both PhD students’ careers.
However, by switching publishing from the printed model to digital publishing the “you don’t talk about negative results” model is changing. You can deposit your negative results in databases and even publish them as a paper in a peer-reviewed journal.
A part of digital biomedical publisher Faculty of 1000, F1000 research is an open-access, peer-reviewed journal. However, the reviewing takes place post-publishing – never again see your competitors miraculously publish very similar data while your paper languishes in the limbo of reviewing. You can update newer versions of your paper as frequently as you like.
PeerJ is an open-access peer-reviewed journal. It charges a one-off fee for an author, currently $99 for the publishing of one paper. Unusually, the reviews are published alongside the paper. The papers are indexed in PubMed and other major databases.
Figshare is a free database, which allows the uploading and sharing of data including papers. Each object has a unique Digital Object Identifier (DOI), which enables searching for and sharing the document.
Is a well-established database of preprints, mostly for maths, physics, and computational sciences, but they also include papers on quantitative biology. The papers are not peer-reviewed but they are curated to weed out pseudoscience and the papers can be found via Google Scholar. The option of later updating the manuscript is also available.
The view that negative results are not worthless and make an important contribution to scientific knowledge is gaining momentum among researchers. This will drive demand for publication methods similar to those detailed in this article and ensure important findings are published.
Do you know of other models or methods of publishing negative results?