Virtual worlds for teaching statistics

Abstract

Data analysis and statistical reasoning are an integral component of modern scientific and medical research. However, students typically take a statistics course in the first or second year of their programs, at which stage it is not at all clear from their other courses why statistics might be useful to them. This can make teaching such a statistics course a challenge. It is not helped by the fact that most students have seen statistics before as part of their mathematics at secondary school, and take from this an obsession with mathematical detail that is not required at the tertiary level (Higgins 1999). The ideal approach to get students aware of the holistic nature of statistical investigation is to have them carry out studies from beginning to end. They should design experiments or surveys, perform them, explore and summarise the results, and then make appropriate statistical conclusions. This approach is feasible for inclass work in small classes or for project work in larger classes. Mackisack (1994) gives an overview of the other benefits of experimental work, including the appreciation of the practical issues involved in carrying out experiments and collecting data. However, it is difficult to supervise, guide, and discuss these activities in large classes, primarily because of resource and time limitations. To address this issue, a collection of ‘virtual worlds’ has been developed to efficiently give students a scientific context for their statistic work, with some control over the experiment itself. The focus here is on the life sciences but the principles are equally useful in other applications of statistics, as well as in applications of general mathematical modeling.