Join Kristin V. Stephens-Martinez, Assistant Professor of the Practice of Computer Science, as she leads a discussion on how can data from large classrooms improve learning. Many classes are growing in enrollment, and one way instructors handle this growth is by increasing their use of technology. A concern about these two factors is how they affect the quality of learning. However, they are also an opportunity to collect and analyze large, high-dimensional data sets and conduct in situ experiments at scale. This talk will include both past and new work taking advantage of this scale. First, the discussion will cover the results of applying mixed methods to a corpus of 332,829 wrong answers from students answering code-tracing, constructed-response assessments. Second, the discussion will address the findings of using those results in an in situ experiment with over 900 students. This experiment automatically delivered hints while the students answered assessments. Third, this talk will present new work that collects data from multiple class tools and integrates them into a single data analysis pipeline, as well as potential research directions using this data source.

Light refreshments will be provided. Please feel free to bring a lunch.