Moving from Trust to Trustworthiness

Academia represents a paradox. On the one hand, many of the methods and techniques used across disciplines are cutting edge and constantly evolving. However, at the same time, our underlying cultures and working practices remain rooted in the 19th Century model of the independent researcher. Research groups are effectively small, artisanal businesses, each crafting outputs – often exquisite, but the product of the unique skills and processes of that group. This model risks poor reproducibility and replicability of research – closed workflows, closed data, use of proprietary file formats, and variability in skills across researchers. It also relies on trust – we have to trust that the researcher or group has fully disclosed all aspects of the process that generated an output. There is empirical evidence that this is not the case, a situation exacerbated by current incentive structures but also legacy systems such as journal article limits on word counts and display items – an echo of the print medium that is now largely defunct. However, existing technologies allow us to make transparent many (if not all) of the elements of a research workflow – protocols, data, code and so on. This transparency of process has the potential to make the system inherently more trustworthy, by allowing it to be scrutinised, thereby moving away from a model of trust in individuals.