Working Group leader: Berna Devezer

Group members: Erkan Buzbas, Gustavo Nardin, Bert Baumgaertner

Originated: Fall 2015


Our team aims to study the current problem of nonreproducibility of scientific findings across multiple disciplines by advancing theory. We have developed statistical theory to demonstrate how nonreproducibility may emerge from a variety of factors. We adopt a stochastic modeling framework to identify the types of factors that affect reproducibility and to quantify their effects by modeling the scientific discovery process collectively pursued by a scientific community. We model the scientific community in a network structure and within this structure there are as many models as scientific projects, creating a universe of complex interactions among scientific models. We are also working on an interactive web application for educational purposes that will demonstrate the dynamics of scientific discovery and reproducibility in a user friendly fashion, as well as a website to compile resources and tools that help enhance reproducibility of research.

Berna Devezer and the work she is doing with this group was highlighted in the Lewiston Tribune on October 21, 2018.