Using Learning Analytics to Characterize Programming Practices in a Code-First Environment for Learning About Evolution

• In Event: 16.062 - Poster Session 2
  In Poster Session: 16.062-4 - Learning Sciences SIG Poster Session

Abstract

This paper reports on a preliminary analysis on log data generated from students’ interactions with a code-first learning environment. In such environments students learn about the mechanisms of target phenomena by programming the behavior of computational agents with code blocks. We characterize students’ programing practices by four measures we developed from our analysis. Although aggregated changes in the four measures at task level is not statistically significant, we found distinct patterns as students proceed from basic to more advanced tasks. Our analysis shows higher resolution pictures of students’ programing practices in a code-first environment. Such analysis can be triangulated with other types of data sources to provide a fuller picture of students’ learning pathways in these types of learning environments.

Authors

• Yu Guo, Northwestern University

• Aditi Wagh, Tufts University

• Sharona T. Levy, University of Haifa

• Corey Brady, Vanderbilt University - Peabody College

• Michael S. Horn, Northwestern University

• Uri J. Wilensky, Northwestern University