This paper extends work on the construction of instructional modules that use graphical and simulation techniques for teaching statistical concepts (Marasinghe, et al. 1996; Iversen and Marasinghe 2001). These modules consist of two components: a software part and a lesson part. A computer program written in LISP-STAT with a highly interactive user interface that the instructor and the students can use for exploring various ideas and concepts comprises the software part. The lesson part is a prototype document providing guidance for instructors for creating their own lessons using the software module. This includes a description of concepts to be covered, instructions on how to use the module and some exercises. The regression modules described here are designed to illustrate various concepts associated with regression model fitting such as the use of residuals and other case diagnostics to check for model adequacy, the assessment of the effects of transforming the response variable on the regression fit using well-known diagnostic plots and the use of statistics to measure effects of collinearity on model selection.
The CAUSE Research Group is supported in part by a member initiative grant from the American Statistical Association’s Section on Statistics and Data Science Education