Abstract

A key statistical modelling approach is to use the data you have to predict outcomes/data you don’t (yet) have. The data that we can learn from is rapidly changing (including data sourced from a variety of digital technologies) and can provide exciting opportunities to broaden student awareness and support their personal connection to data. In this workshop, we’ll engage in a range of activities that demonstrate how exploring and investigating relationships between variables can support student engagement with informal approaches to prediction and provide a foundation for predictive modelling at the higher curriculum levels.

The team for workshop was: Anna Fergusson (a.fergusson@auckland.ac.nz), Chris Wild, Jim Davis, Clare Nelson, Lisa Mulvey, Jacqui Hammond, Ash Rambhai, Amy Hooper, Hanna Reid, Marion Steel

Overview

This workshop has a teaching focus rather than assessment. There’s lots of things that are uncertain about what we may be teaching in the near future, due to the NCEA change programme (RAS) and the curriculum refresh project. You might be looking for something definitive, like a set of steps that explain exactly how to assess one of the new proposed NCEA Level One Mathematics and Statistics achievement standards. This is totally understandable, but you won’t find that in our materials :-)

What we have focused on for our workshop are what we believe are core ideas related to prediction that could both benefit our teaching right now but also inform how we could be teaching in the future. We hope that through our examples and notes that you do obtain a clear understanding of what would be important to teach, and so assess, and gain some new ideas for data contexts and ways to engage ākonga with their learning from data.

There’s also an unashamedly data science influence! We’ll explore foundations for predictive modelling at Y12/13 (e.g. image recognition), dynamic sources of data, some ideas for how to access, explore, and use data about ourselves, and in general build awareness of digital technologies & related data technologies and create awesome things from data!

The workshops were structured using the following six key ideas.

  1. A key statistical modelling approach is to use the data you have to predict an outcome you don’t (yet) have
  2. Prediction is a broad core idea and exploration provides an accessible, flexible, creative way of supporting predictive modelling ideas
  3. Time series data is a great introduction for weaving storytelling and informal predictive modelling (& for working more closely/personally with digital data + CT)
  4. There are different purposes/goals for prediction and different ways of designing data (e.g. experiments)
  5. We can learn from features of scatter plots, using informal methods for prediction, and evaluating models in terms of accuracy and precision
  6. Putting all these ideas together with ePPDAC

The static slides from the workshop can be downloaded here but you’ll get more out of watching the videos from each of the six sections of the workshop below :-)

Time series data

At the start of the second hour, teachers did a quick reflection in groups about what they thought were key ideas, and these are shared below.

Designing data

Informal prediction models

License You are free to use these materials for non-commercial purposes (i.e. for your own teaching) but need to take care to make references to the work of others and any diagrams or frameworks used, as we have done so throughout these materials. If you have any questions about the materials, or use of the materials, please contact Anna Fergusson (University of Auckland, a.fergusson@auckland.ac.nz).