This is a session of Topic 4: Statistics education at the post-secondary level

Randomisation and bootstrapping: the quick way to inference

Organizer

• Michael Forster (New Zealand)

Abstract

In introductory statistics courses we establish the connected ideas of sampling variation and randomisation of group labels early on and then embark on what is typically a long and convoluted road to classical statistical inference. In navigating our way along this path we often confuse and even lose many of our students due to the multitude of ideas and concepts we have to teach them. We need to be able to establish the core concepts of statistical inference much faster and connect better with their prior experience and intuition, not only to maintain our students’ interest but also to prepare them for when we do move to the more formal ideas of mathematically based inference. If we can establish the core concepts of inference early on, our students will have a much better idea of where they are going and why they need to go there. Modern computer technology now allows us to use interactive visualisations of statistical techniques such as bootstrapping and randomisation tests so that we can have beginners doing statistical inference in a matter of hours rather than days or weeks. The question then arises as to how best to us these tools so that students really do extract from them the lessons we want them to. In this session we explore both the concept and various methods that have been used in an attempt to achieve this lofty but very desirable goal.

Papers