Manipulation of Discrete Random Variables with discreteRV

A prominent issue in statistics education is the sometimes large disparity between the theoretical and the computational coursework. discreteRV is an R package for manipulation of discrete random variables which uses clean and familiar syntax similar to the mathematical notation in introductory probability courses. The package offers functions that are simple enough for users with little experience with statistical programming, but has more advanced features which are suitable for a large number of more complex applications. In this paper, we introduce and motivate discreteRV, describe its functionality, and provide reproducible examples illustrating its use. Introduction One of the primary hurdles in teaching probability courses in an undergraduate setting is to bridge the gap between theoretical notation from textbooks and lectures, and the statements used in statistical software required in more and more classes. Depending on the background of the student, this missing link can manifest itself differently: some students master theoretical concepts and notation, but struggle with the computing environment, while others are very comfortable with statistical programming, but find it difficult to translate their knowledge back to the theoretical setting of the classroom. discreteRV (Buja et al., 2015) is an approach to help with bringing software commands closer to the theoretical notation. The package provides a comprehensive set of functions to create, manipulate, and simulate from discrete random variables. It is designed for introductory probability courses. discreteRV uses syntax that closely matches the notation of standard probability textbooks to allow for a more seamless connection between a probability classroom setting and the use of statistical software. discreteRV is available for download on the Comprehensive R Archive Network (CRAN). discreteRV was derived from a script written by Dr. Andreas Buja for an introductory statistics class (Buja). The package rv2 (Buja and Wickham, 2014), available on GitHub, provides a useful example of using devtools (Wickham and Chang, 2015) to begin basic package development, and also uses Dr. Buja’s code as a starting point. The goal of rv2 seems more focused on learning package development, while the goal of discreteRV is to be a useful statistics education and probability learning tool. The functions of discreteRV are organized into two logical areas, termed probabilities and simulations. This document will illustrate the use of both sets of functions. All code used in this document is available in a vignette, accessible by loading discreteRV and calling vignette("discreteRV").