EJMT Abstract

Research has shown that the use of computer simulation methods as an alternative to traditional methods enhances the understanding of the statistical concepts. The increasing availability of technology allows instructors and students to use computationally intensive methods such as simulation. This paper presents the use of real data and a simulation approach to help students understand the confidence intervals for population mean and variance. Use of real data makes the concepts more real for students and enhances their ability to ground the new concepts in their existing knowledge. We use the R programming environment for simulating repeated sampling from a fairly large dataset and compute the approximate sampling distributions of sample mean and variance. We notice that the confidence intervals for population variance work poorly if the normality assumption is violated. Our preliminary assessment shows that students gain better understanding of confidence intervals using this approach.