Introducing an Informal Approach to Predictive Modelling within a Large Introductory Statistics Course: Connecting Research with Practice

Abstract: Teaching recommendations for implementing data science at the introductory level include placing greater emphasis on predictive modelling. In contrast to how simple linear regression is introduced, suggested approaches include training and testing models with different data sets, validation through residual analysis and predictive accuracy, and targeting concepts such as overfitting, underfitting, and generalisability. As part of a design-based research study, I developed a web-based task to introduce learners to predictive modelling and online databases (APIs), which was then implemented with six high school statistics teachers. The findings of this research were then used to inform the design and implementation of an informal approach for introducing predictive modelling to students in a very large introductory statistics course (n ≈ 2200 per semester) at the University of Auckland | Waipapa Taumata Rau. In this talk, I will describe and illustrate the development of the learning progression for introducing predictive modelling, demonstrate some of the tasks and data technologies we use, and discuss the opportunities and challenges of establishing and sustaining connections between research and practice.