This project explores ways to support educators as they introduce interactive data-driven activities to help learners develop intuitions about data and acquire specific analytic skills. While educators agree that the availability of data offers new educational opportunities in many disciplines, there is little consensus about how data-driven educational activities should be created. Often current solutions mandate teaching both instructors and students to program; these solutions are ambitious in scope, but can easily fall short of their aspirations, discouraging those who have difficulties learning to program and distracting the class from its primary educational goals. The prediction simulations framework enables an alternative type of activity, but how teachers perceive the opportunities and challenges offered by this type of activity, and how prediction simulations relate to their students, subject area, and educational context all pose new questions. Data analysis skills are a fundamental element of many disciplines, motivated by the widespread availability of datasets, ubiquitous sensing capabilities, and ready access to processing power. The ability to create prediction simulations has a potentially transformative effect on data science pedagogy. Concepts can be introduced to learners at different levels and disciplines to motivate learners and keep them engaged. Besides allowing learners to "dig into" data to develop their own understanding and construct better mental models of the underlying phenomena, prediction simulations can deepen a learner's understanding and intuition for the use of data in other domains.
This project will provide an understanding of how teachers view the increasing availability of data in their domains, the tools and curricular activities currently available to them, and the overhead of creating and tailoring activities to their environment. Towards that goal, two interacting lines of research will be pursued: a multi-method study of the non-programmers who might create prediction simulations for use in pedagogical settings; and the design and evaluation of techniques to support authoring of prediction simulations. These two intertwined activities will help answer the following research questions: (1) What are the characteristics of domains, educators and students, and pedagogical situations that center on data analysis skills? Which would benefit from and be amenable to the introduction of prediction simulations? (2) Can generalized capabilities be developed to support the creation of novel situation-specific prediction simulations based on new combinations of datasets, visualizations, and analytic tools? (3) What are the most effective ways to support the authoring of new prediction simulations by non-programmers? Methods to be investigated include template-based and specification-based authoring.
This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.
