NormalityAssessment: An Interactive Classroom Tool for Testing Normality Visually

Many inferential and predictive statistical procedures possess underlying theoretical assumptions that should be met in order for the results of those procedures to be considered reliable. One assumption associated with methods for population means, including linear regression coefficients, is that of normality of a population(s). When assessing normality, two graphical tools that are often utilized are normal quantile-quantile (QQ) plots and histograms. However, while these tools are popular, they still present challenges for many who use them due to the subjectivity oftentimes involved when examining them. In this article, we describe a free, interactive Shiny application, downloadable as an R package, which implements two procedures recently developed for graphical inference with a specific emphasis on assessing normality. The application was created and designed with a focus on statistics education.