Collaborative Research: Community-Building and Infrastructure Design for Data-Intensive Research in Computer Science Education

Carnegie Mellon University, Pittsburgh PA

Investigators

Abstract

The Building Community and Capacity in Data Intensive Research in Education program seeks to enable research communities to develop visions, teams, and capabilities dedicated to creating new, large-scale, next-generation data resources and relevant analytic techniques to advance fundamental research for areas of research covered by the Education and Human Resources Directorate. Successful proposals will outline activities that will have significant impacts across multiple fields by enabling new types of data-intensive research. Online educational systems, and the large-scale data streams that they generate, have the potential to transform education as well as our scientific understanding of learning. Computer Science Education (CSE) researchers are increasingly making use of large collections of data generated by the click streams coming from eTextbooks, interactive programming environments, and other smart content. However, CSE research faces barriers that slow progress: 1) Collection of computer science learning process and outcome data generated by one system is not compatible with that from other systems. 2) Computer science problem solving and learning (e.g., open-ended coding solutions to complex problems) is quite different from the type of data (e.g., discrete answers to questions or verbal responses) that current educational data mining focuses on. This project will build community and capacity among CSE researchers, data scientists, and learning scientists toward reducing these barriers and facilitating the full potential of data-intensive research on learning and improving computer science education. The project will bring together CSE tool building communities with learning science and technology researchers towards developing a software infrastructure that supports scaled and sustainable data-intensive research in CSE that contributes to basic science of human learning of complex problem solving. The project will support community-building and infrastructure capacity-building whose ultimate goal is to develop and disseminate infrastructure that facilitates three aspects of CSE research: (1) development and broader re-use of innovative learning content that is instrumented for rich data collection, (2) formats and tools for analysis of learner data, and (3) best practices to make large collections of learner data and associated analytics available to researchers in CSE, data science, or learning science. To achieve these goals, a large community of researchers will be engaged to define, develop, and use critical elements of this infrastructure toward addressing specific data-intensive research questions.The project will host workshops, meetings, and online forums leveraging existing communities and building new capacities toward significant research outcomes and lasting infrastructure support. This project will provide an infrastructure that can support various kinds of research in CSE domain as a one-stop-shop, and will be the first to focus on full-cycle educational research infrastructure in any domain. CSE tool developers and educators will become more productive at creating and integrating advanced technologies and novel analytics. Learning researchers will have better tools for analyzing the huge amounts of learner data that modern digital education software produces. Data scientists will have rich new datasets in which to explore new machine learning and statistical techniques. Collectively, these efforts will reduce barriers to educational innovation and support scientific discoveries about the nature of complex learning and how best to enhance it. The project will support scientific investigations through community meetings and mini-grants to others addressing questions such as: What is the optimal ratio of solution examples and problem-solving practice? How do computational thinking skills emerge? In what quanta are programming skills acquired? Can automated tutoring of programming be effective at scale in enhancing student learning?. Many of the innovations developed under this project will directly impact learning in any discipline. Educational software will more quickly be developed in the future, that more easily generates meaningful learner data, which in turn can be more easily analyzed.