Research data management for open science

Abstract

This dissertation investigates research data management practices in laboratories in the context of open science. To achieve that, we first seek to understand what kind of organizational and technological issues are impeding the planning, production, preservation, and dissemination of data in laboratories. Then, we propose a conceptualization of laboratory work using the lens of experimental systems theory, which provides a socio-technical perspective on the building blocks of scientific experimentation. Finally, we apply the lens of reproducible experimental systems further to design a laboratory forensics approach for investigating storage systems in laboratories. The laboratory forensics approach is a starting point of experimental resources discovery and evaluation in labs. Next, we draw upon the results of forensic investigations to shape open science readiness, which is an ensemble of RDM practices and technology that supports reproducible and open practices in laboratories. The goal of pursuing the design of open science readiness for laboratories is to foster evidence-based research data management that effectively achieves the preservation and dissemination of research data in an open and FAIR way. Therefore, the main research question of this dissertation is stated as follows: How can we organize research data management for preserving and disseminating laboratory experiments in a reproducible way? First, we start with organizational and technological issues among stakeholders involved in research data management. First, we examine the cooperation between researchers and data managers. By doing so, an agenda for open data in academia is proposed based on qualitative research highlighting issues such as lack of proper infrastructure, accountability, legal frameworks, and rewards in research data management. At the same time, new roles such as data stewards and the struggles with data management support are investigated. To further determine stakeholders’ needs and practices, a similar exploratory approach is used to discover how funding agencies and data management support develop a research data strategy in the Netherlands. Then, we elaborate on the concept of reproducibility in experimental science. To achieve that, we dive into data management issues from a technological point of view, showing what types of reproducibility issues occur in storage systems with laboratory forensics techniques. Moreover, we investigate reproducibility in research data management by mapping laboratory work and the scholarly infrastructure to a socio-technical model. As such, we obtain a more comprehensive view of reproducibility issues and refine organizational and technical aspects of reproducibility challenges in practice. Finally, we illustrate some applications of “FAIR technology”. First, we show the need for designing reproducible and reusable research software with the reproducible, research-oriented knowledge discovery in databases process (RRO-KDD). Then we present a strategy for open science readiness. The results of this work provide research laboratories and other stakeholders such as libraries, ICT, and funders with insights into reproducibility and open science challenges grounded into an investigation of laboratory work.