Ensuring Reproducible Research through Rigorous Study Design, Statistical Analysis Planning, and Data Presentation within a Team Science Framework

The rigor and reproducibility of research relies on the appropriate use of statistical methods, in addition to the careful planning and designing of studies to answer research questions and make meaningful inferences based on data. While statistical methodology is widely used and disseminated in all areas of research, it is often misused due to lack of knowledge and training in statistical concepts and inadequate planning strategies. The field of neuroscience in particular has been criticized for overall poor study planning and misuse of statistical methodology, specifically related to inadequate sample size and power to detect clinically meaningful effects, as well as inattention to fundamental statistical assumptions. Although true for all disciplines, it is particularly important to encourage collaboration and team science in the field of neuroscience to not only offer a larger scope of perspectives and expertise, but also to result in improvements in the rigor and reproducibility of the resulting research. In partnership with CENTER, the team of collaborative biostatisticians, epidemiologists, and data scientists at Virginia Tech’s Center for Biostatistics and Health Data Science (CBHDS) will develop, evaluate, and disseminate a series of educational units in rigorous and reproducible approaches to study planning and design within a team science framework, including: 1) creating a formal plan for experimental design, execution and analysis; 2) exercising a sound process for choosing appropriate descriptive statistics; and 3) identifying assumptions and sources of error. The proposed educational units will be adaptable across various disciplines, formats, levels of learners, learning environments, and diverse backgrounds and perspectives.