Paper short abstract:

Based on a comparative, secondary analysis of case studies on the production of scientific knowledge and the role of replication this paper seeks to contribute to a better understanding of the dimensions of replication across different epistemic cultures and how they link to open data.

Paper long abstract:

In recent years concern has been growing about the irreproducibility of findings published in the scientific literature. The open sharing of data underlying published results has been proposed as a remedy. In this debate, 'scientific replication', the ability of others to reproduce a result, has been highlighted as a crucial method to ensure the reliability and integrity of scientific knowledge. Some suggest an ethical imperative exists for scientists to share data to enable scientific replication. Others speak out against stricter requirements for the provision of data for what has been termed 'reproducible research', warning of the costs and effort involved for authors and potentially the entire scientific community (e.g. as referees.)

The call for the public sharing of research data to improve the reliability and trustworthiness of scientific knowledge aligns with a larger trend of mandating public access to scientific data. However, many statements that dismiss or promote data sharing to enable scientific replication focus narrowly on one type of replication and generalize across fields. Scientific fields differ in their epistemic practices, the kind of knowledge they produce, and their construction of data. This paper is based on a comparative, secondary analysis of empirical and historical case studies on the production of scientific knowledge and the role of replication. It distinguishes between explicit, implicit, exact, and conceptual replication and aims contribute to a better understanding of the dimensions of replication across different epistemic cultures and methodological traditions, and the extent to which it requires the sharing of data.