Openness guides discovery

Abstract

In reality, research projects grow through an evolutionary process. Variation — the substrate of evolution — is provided through the emergence of new questions and avenues of investigation. The research team must then choose which directions to pursue, a process akin to natural selection¹. Along with this evolution at the macro-scale, a much more orderly process is required at the micro-scale. Within each step, thoughtful study design leads to robust experiments and analyses. This complementarity between evolution and design best encapsulates the process of discovery: in ‘night science’ we evolve ideas for the next step, while ‘day science’ tests them².

A project’s evolutionary history is generally obscured in the resulting scientific publication. The publication’s function is to justify and communicate the project’s main results; it is not a historical account. Instead of recounting all of the project’s dried-up branches, publications zoom in on a single lineage in its evolution: the steps that led to the most interesting result (Fig. 1c). Publications typically describe the discovery as it ideally should have happened, reporting only the evidence relevant to the proposed claims³.