‘How do we define success?’ – Rethinking failure and success in science

This post was collaboratively written by PLOS staff (Ines Alvarez-Garcia, Phil Mills, Leonie Mueck and Iratxe Puebla)

Independent of the context, failure is a word that hardly ever leaves us indifferent. Fear of failure is human nature, and it is common that we prefer not to talk about failures if we can avoid it.  When we think about this in a professional context, failure can have clear and immediate ramifications for reputation and career progression and – as with any other professional – researchers are not immune to this fear of failure.

Part of this approach to failure in research is due to the fact that the research system has traditionally rewarded those who are the first to report a finding over those who are second, and those who report a positive result over those reporting a negative one. However, research generally involves a trial-and-error approach and a plethora of negative findings, or protocols that require troubleshooting before they are fine-tuned. Thus ‘failed’ experiments are common; more so than is often recognized or reported. Much effort and many hours of meticulous research endeavour go unrecognized by the current research assessment frameworks, resulting in a considerable squandering of potentially important research outcomes.

The ‘Failures: Key to success in science’ event at the Cambridge Festival of Ideas 2018 aimed to reflect on these considerations in a conversation involving our five panelists and the audience around the notions of failure and success in science.

Our five panellists kicked off the conversation by giving their perspective on what a successful research career should look like.

‘How do we define success?’ asked Cathy Sorbara (Co-chair of CamAWiSE, Cambridge iation for Women in Science and Engineering), maintaining that science does not have a defined endpoint, that collaboration should be a key part of the research process, and that scientists need to think about how they communicate their work, particularly to those unfamiliar with research. Tapoka Mkandawire (PhD candidate, Sanger Institute) felt that a key aspect of success is to work on something that you feel passionate about and are keen to share. The audience was interested in the forms that communication of research could take and the panellists noted that communication about research should not be restricted to publications, putting forward ideas around visual formats such as videos. Tapoka noted that her research group has developed a comic book to more easily describe their work to children.

A common theme was that the binary classification of success vs failure is somewhat unfair. Should a result be tagged as failure only because it’s negative and not been published? Fiona Hutton (Head of STM Open Access Publishing at Cambridge University Press) advocated the development of a more collaborative open pathway for research, with more openness at all steps of the process, such as that demonstrated with open lab notebooks, to capture the incremental steps that make up the research process. The sharing of negative and null results should be encouraged as well, as a move away from frameworks that rely on impact factors to assess the quality of research; Fiona mentioned DORA as a good initiative in this space, which is gaining support from institutions and funders.

Arthur Smith (Deputy Manager of Scholarly Communication (Open Access), University of Cambridge) and Stephen Eglen (Reader in Computational Neuroscience, University of Cambridge) tackled the challenges with the current research system and acknowledged that this places Principal Investigators (PIs) as the ‘survivors’ of the system, with only a few reaching the top of a steep pyramidal career structure. Stephen stressed that the driving force for getting into research should be a genuine interest in science and not the goal to eventually become a PI. Arthur noted that there are many other career paths available after a PhD and that the skills gained can be used in many other areas, such as the private sector. The training of PhD students should include aspects that go beyond publishing, and should balance this with the development of communication and other skills.

To round up the discussion we asked panellists to provide recommendations for steps that can help shift perceptions about success and failure in science. Here is what they told us:

• More support for early career researchers, so that they can have an informed, broader view of their career, and of the options after a PhD.
• Further recognition for the wide range of different roles that scientists play beyond the publication of research findings – for example, peer review activities, mentorship, etc.
• Provision of credit for recording and reporting troubleshooting, for any work that may not follow the shape of a conventional publication but which would help others engaged in related research.
• More training for those in a research path, to help them develop a variety of transferable skills, and to recognize the value of those skills.
• Increased diversity – higher diversity can only be beneficial in driving change towards how success is defined.

Achieving these aims and helping to sway current views about failures in research represents a formidable task, but – much like science itself – change progresses one step at a time and we hope that the engaging conversation at the Festival of Ideas provided one such step to shift how we define “success” in research. As we pursue initiatives towards such change, let’s remind ourselves of Arthur Smith’s definition of success: ‘Success is what makes you happy’.