Improving Statistical Literacy for Physician Scientists: Sampling Distributions

This is the first in a multi-part series on teaching statistical inference to physician-scientists training to work as members of interdisciplinary scientific teams. This unique student audience has greater scientific sophistication than a typical statistics student but less background in mathematics and computer programming, which presents challenges for traditional approaches to teaching statistics. Here, we illustrate an innovative approach to teaching sampling distributions to physician-scientists. Sampling distributions are a fundamental element of statistical inference; they are a building block of confidence intervals and hypothesis tests which are vital tools for performing clinical research. As such, it is essential that physician-scientists have a strong foundation in sampling distributions. Key elements of our innovative approach include the use of a running example, delivery of content in small pieces to reduce cognitive burden, preceding formulae with pictures, combining static and dynamic content using an R Shiny app, and use of self-graded quizzes to provide immediate feedback. The resulting course module can be reused in multiple contexts, including as part of self-directed, asynchronous learning or by incorporation into a traditional or flipped classroom setting.

Duke Scholars

Author Gregory P. Samsa Biostatistics & Bioinformatics, Division of Biostatistics

Author Caroline Falvey  

Author Jesse David Troy Biostatistics & Bioinformatics, Division of Biostatistics

Author Gina-Maria Pomann Biostatistics & Bioinformatics, Division of Biostatistics

Author Steven C. Grambow Biostatistics & Bioinformatics, Division of Biostatistics

Author Megan Lee Neely Biostatistics & Bioinformatics, Division of Biostatistics