This project establishes a theoretical framework for the reproducibility of science. Within this framework the researchers will create protocols for increasing the use of reproductions and replications in human-environment and geographical science (HEGS) research. Reproductions repeat a study while holding the input data and parameters of the study constant, thereby establishing credibility and generalizability of prior research. This project is responsive to increasing governmental requirements for transparent and reproducible research. It will conduct reproduction and replication studies of recent high-impact HEGS research to establish credibility and generalizability necessary to influence public policy. The investigators will integrate this framework for reproducible science into undergraduate and graduate methods courses resulting in a curriculum for teaching replication studies and reproducible research practices.
HEGS has been slower than other disciplines to adopt more reproducible research practices, limited by a lack of well-developed exemplar replication studies, curricula for teaching reproducibility, and the challenge that variable geographic contexts cause for replication. What is the state of research practices in HEGS with regard to reproducibility and barriers to reproducibility? Can the credibility and generalizability of recent high-impact HEGS research be established through reproductions and replications? Can teaching reproductions and replications achieve the same benefits as project-based learning pedagogies and develop researchers with the capabilities for doing reproducible research in HEGS? The proposed work is anticipated to establish the foundational survey research, reproduction and replication studies and pedagogies necessary to transform research practices in HEGS by building knowledge through the scientific process of reproducing and replicating studies.
Many of our licenses list “HEGSRR Contributors” as the copyright holder in the tradition of other open source projects with numerous individual contributions.
The contributors to the HEGSRR project include: Peter Kedron, Kufre Udoh, Drew An-Pham, Derrick Burt, Sarah Bardin, Tyler Hoffman, the Open Source GIScience students at Middlebury College, and the GIS Capstone and Bayesian Statistical Modeling students at Arizona State University.