improving Reproducibility In SciencE (iRISE)

Structured understanding of the drivers of irreproducibility and presenting concrete solutions of tools and interventions will help to increase the quality, reliability and re-usability of scientific evidence. To this end, iRISE proposes to provide theoretical and empirical evidence of the effectiveness of specific interventions, and a framework for a robust, evidence-based road map for the development, assessment and implementation of interventions intended to improve reproducibility. iRISE brings together qualitative and quantitative expertise, from academia and SMEs, including meta-science, statistics, economics, artificial intelligence, research ethics and integrity, quality assurance, and project management. iRISE proposes the development of a general framework for diagnosing and addressing reproducibility problems using analytical and computational modelling, simulations and meta-studies. Data on existing interventions will be systematically curated and evaluated, and stakeholders will be consulted to collaboratively identify practices and tools that should be prioritised for implementation. iRISE proposes to conduct empirical studies of both technical and practice-based solutions to increase reproducibility. Across all iRISE activities, the influences of research culture will be investigated, with a focus on mainstreaming systematic integration of equity, diversity and inclusion practices. A comprehensive Stakeholder Forum will be engaged to provide advice, and iRISE will commit to open and reproducible practices. The different types of evidence generated will be integrated into an open knowledge base to support the community in decision-making to identify, test, and implement effective and feasible solutions for reproducibility. The members of iRISE have made pivotal scientific and policy contributions relating to robustness, rigour and reproducibility in the past and have the skills and tools to succeed in this ambitious project that has potential...