When we talk about ‘high impact’ in science, we are really talking about two related, but distinct entities. The intuitive meaning refers to scientific research that dramatically improves or changes our understanding of any given topic. This may involve using a novel technique to address an existing problem, or providing a detailed account of a signalling pathway to inform future research etc. The secondary meaning of high impact refers to publishing work in a platform with a high ‘impact factor’. In biology, these basically refers to publishing in the closed-access journals of Nature, Cell and Science. Increasingly, it is recognised that these two definitions of ‘high impact’ have very different scientific realities, and that these differences are hurting investigative research.
High impact journals influence high impact science by restraining the definition of good science
To provide an example, think about the rise of the ‘characterisation’ paper. With the birth of Jackson labs and the increased ease of generating GM mice, a huge number of papers are now based around a mouse model of some sort. Now, I am not knocking mouse models in general, but I have become aware that many of the papers making ‘high impact’ journals are based around the following notion: Generate a GM mouse, blindly characterise a miscellaneous subset of phenotypes from a broad range of perspectives (biochemically, behaviourally, ultrastructurally…), proceed to rake in brownie points from the scientific community.
I can see the appeal in this sort of broad stroke, holistic approach to research and to publishing. It means a single paper can link findings at the molecular level with behavioural changes that, as a neuroscientist, is something of a holy grail. However, it brings with it many issues.
Mostly importantly, this attitude normalises the way all research should be done. The reality is that the techniques and approaches used to address questions in science depend of the questions you ask. It seems fundamental, but nowadays it seems like a paper has to have some form of behavioural data or electrophysiology data in order to publish in the big journals. This has two important consequences: labs are punished for specialising and large, collaborative labs with access to vast resources are rewarded; it diminishes the importance of the question in science and how best to tackle that question.
This would all be a minor headache for scientists if it wasn’t for the fact that catering to the wishes of these ‘high impact’ journals is par for the course in an academic career. There is no way of avoiding it. High impact journals influence high impact science by restraining the definition of good science. If history has taught us anything, it’s that we do not know where the next big breakthrough in science will come from and to think that we do would be foolish.