A general framework for diagnosing and addressing reproducibility problems

In light of the current state-of-the-art of understanding reproducibility, WP1 will offer a clarifying framework that will support the other work-packages through the following objectives:

Obj. 1.1 Clarify, for the purposes of the project, what “reproducibility” means in different contexts and how it can be measured, how to synthesise and organise the current understanding of the possible causes of irreproducibility, and how to measure them.

Obj. 1.2 Assess how the effects of interventions to improve reproducibility may vary depending on the phenomenological and methodological characteristics of a study and the context in which the research operates.

Obj. 1.3 Validate statistical metrics of reproducibility that take into account the different possible types and causes of reproducibility.

Obj. 1.4 Assess how well current theoretical understanding of reproducibility allows forecasting effects of reproducibility interventions.

WP1 Leads

Rachel Heyard

Rachel Heyard

ORCID

Postdoctoral Fellow (Center for Reproducible Science)

University of Zurich: Zurich, CH

Postdoctoral researcher at the Center for Reproducible Science at University of Zurich, Switzerland. She received her PhD in Biostatistics at the University of Zurich in 2019. After her PhD she joined the Data Team of the Swiss National Science Foundation (SNSF) as a Statistician where she gained first-hand insights in the working of a funding agency. It is here where she started getting interested in science policy and meta research. In Summer 2022 she left the SNSF to redirect all her efforts towards research with primary focus on metrics for replication success, promoting good research practices, and advocating responsible research evaluation. In iRISE, Rachel is co-lead for WP1 and participates in WPs 4, 5, 6 and 7.

Daniele Fanelli

Daniele Fanelli

ORCID

Meta Scientist

London School of Economics and Political Science: London, GB

Assistant Professor in Social Research Methods at Heriot-Watt University and Visiting Fellow at the London School of Economics and Political Science. He has published several influential empirical studies on the prevalence, causes and impact of bias, misconduct, and other sources of irreproducibility, and has contributed to several international initiatives to advance research Integrity and reproducibility including the EU Mutual Learning Exercise on Research Integrity, and two national research integrity committees, for the Luxembourg Agency of Research Integrity and the Italian National Research Council. He is leading the theoretical components of WP1, aimed at modelling and predicting reproducibility in different contexts, and will be contributing to WPs 3,6, and 7.