Software Testing Practices for Reproducible Open Science

Over the last two decades, the scientific community has been raising concerns about the “reproducibility crisis in science”. According to the Baker, M. 2016 study, in almost every area of science, more than 50% of scientists were unable to reproduce their own work, let alone another group's. This crisis gained national attention, during which the US Congress tasked the National Academies of Sciences, Engineering, and Medicine to investigate this issue. They defined reproducibility, i.e. computational reproducibility, as “obtaining consistent results using the same input data; computational steps, methods, and code; and conditions of analysis.” Almost all of today's research requires writing code, and as scientists, likely alone. In this poster, we extend the intent of the Koufos, A. TESS 2024 poster, focusing on testing practices you can start following that will help improve the likelihood of reproducibility in your research.