LONDON, UK. The World’s top ten journals in the biosciences have come together to release guidelines that they believe scientists should follow to ensure publication, and The Science Web has gained exclusive access to these. We print them below, without modification:
In order to ensure that we publish only the very best news-worthy stories, we encourage all of our submitters to follow the guidelines below.
1. First, choose an interesting, though controversial, topic; such as “the impact of the microbiome on health”, or “trans-generational epigenetic effects”.
2. We recommend that you carry out experiments on only a few, well chosen individuals. Using a small numbers of subjects increases your chance of finding an effect; also, we simply don’t have the space to publish studies on larger numbers of samples. Be wary of increasing the numbers, or checking in an independent population – your effect may disappear completely, and we would definitely not publish a negative result. Don’t worry about the “small N” problem, we can simply extrapolate.
3. You should probably look at only a small number of well-chosen candidate genes. Again, by limiting the number of genes you look at, you will increase the chance of finding a difference. There’s really no need to look “genome-wide” – simply use your awesome encyclopaedic knowledge of the subject to pre-determine what the answer is – after all, you knew what you would find before even starting the study, didn’t you?
4. If at first you don’t find an effect, torture your data until you do. This can be done by dropping “pesky” outliers from the dataset. And remember, it doesn’t matter how small your effect is – simply fiddle with the scale on your graphs to make it look bigger!
5. Speculate wildly and inaccurately about the significance of your results; don’t worry if your results, only ever observed in a tiny population on a small number of genes, and completely unreplicable, have very little real-world significance – simply make stuff up about how it impacts our understanding of e.g. human evolution or the existence of alien life
6. Ignore statistics, almost completely. Definitely do not correct p-values, it only makes them less significant. We don’t have space for error bars
7. Please do not put any interesting methodological information in the main paper – our readers are not interested in *how* you did it; bury that in the supplement
There we are! The Science Web are proud to bring you this exclusive guide on how to get published in only the very best high-impact journals. Enjoy!