Failure is an option (english article)

Reporting about mistakes and failed research is part of open science. This does not include non-results but instead, the methods and designs that went wrong due to one’s own decisions or external conditions should be published. In journals, at least in Public Management, there has been no room for this and normally scientists do not like to talk about their failure (in public). Who likes to admit to not having considered something or making mistakes.

Error Culture

However, this handling of mistakes is surprising since most research is based on trial and error. With the exception that one hardly ever speaks about trial and error when the final result is in. Especially to newcomers studies that directly come to the right solution seem like magic. It seems that only their own research does not work out.

Additionally, cumulative dissertations and the general pressure to publish tempt researchers to publish every study instead of conducting a series of (failed) studies.

Open Lab Notebooks

One solution to these problems is already common in some disciplines: open laboratory notebooks. In such diaries, the scientist regularly notes the progress of the study but also, the things that did not work as planned. When these notes are made accessible openly (e.g. on GitHub) they can be easily retrieved by everyone. The problem with this is, that most likely only a few scientists will look into these project-logs before they plan a study. The same applies to personal blogs or websites where scientists document the progress of their scientific projects in more detail. This openness is commendable, however, will probably not reach enough readers.

The ‚Failure‘-Section in Journals

This is the reason why failed trials, in my opinion, belong to journals and conferences. A forthcoming special issue of the International Journal for Qualitative Methods offers a great start. For a special issue on „Qualitative and Mixed Method Failures“ researchers had been invited to report on situations and events „in which [their] choices, presence, or influence contributed conceivably to an adverse or undesirable research process or outcome(s)” (IJQM 2019). Such articles one normally looks for, even in method journals, in vain. Especially these journals could use a section on learnings. And also in disciplinary journals, it seems a good idea to me when problems refer to specific populations (e.g. public administration) or discipline-specific methods.

It would also be possible to integrate a section of “learnings” into the actual research paper. In my experience, most researchers report fairly transparent about their research process, but mostly these passages are eliminated first from their articles. And in the end, no one can read up on it. So why should this issue not be included in a structured abstract? Besides ‘results‘, ‘method‘ and ‘practitioner points‘ there would also be ‘failures and learnings‘.

The ‚Fuck Up‘-Panel

In addition, and even better before, failed research designs and methods should be discussed more intensely at conferences. Inspired by so-called fuck up nights, in which founders talk about failed start-ups, stories on failure could be told in conference presentations. Either in an own panel or even better in the regular tracks. Especially in relatively fixed special interest groups that regularly meet up and are loyal communities, no fear should arise that admitting to such mistakes could hurt one’s career.

This possibly could disenchant some scientific stars but at the same time, it would make them more approachable and human. Because: failure is success in progress.

---

This article has already been published in a german version here.

Translation by Lisa-Sophia Preller.